Deep Learning Based Concrete Beam Crack Detection and Classification

1
DEEP LEARNING BASED CONCRETE BEAM CRACK

DETECTION AND CLASSIFICATION
by
MD.SAMIN YEASER FAHIM
(STUDENT ID: 201304160)
In partial fulfillment of the requirement for the degree of

BACHELOR OF SCIENCE IN CIVIL ENGINEERING
DEPARTMENT OF CIVIL ENGINEERING
BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA
October, 2018 
2
ACKNOWLEDGEMENT
All praise to the Almighty Allah for blessing me with the means, courage and
guidance in performing this research study, despite all odds.
First of all, I express my profound gratitude to Prof. Dr. Ahsanul Kabir, my
supervisor and revered mentor, for his constant direction and inspiration in pursuing
the whole investigation of this research project and for correcting the mistakes
whenever necessary. I thank him from the bottom of my heart for his sincere
cooperation, precious advice and consistent encouragement and motivation. Had he
not been there, the research could not have been completed.
I convey my heartiest regards to Dr. Md. Hadiuzzaman, Associate Professor,

Department of Civil Engineering, BUET and his research assistant Nazmul Haque
for their help, co-operation and valuable advice during this research project that has
helped me immensely. I am extremely grateful to Nazmul Haque for sharing his
wisdom and for generously examining my research works despite his busy schedule.
Finally, I would like to thank my family and friends for their unflagging love and
unconditional support throughout my life and studies which worked as a stimulus
for me in making a successful and productive completion of the thesis.
ii
3
ABSTRACT
Cracks in buildings are of common occurrence which develops whenever stress in
the component exceed its strength. Most of the building components are designed as
crack section, except for some special structures such as water retaining structures,
ETP etc., so crack is inevitable in RCC structure. However cracks in structure may
pose threat and reduce durability and sustainability of the structure if it happens due
to reduction in load bearing capacity of a structural component.
A number of image processing techniques (IPTs) have been implemented for
detecting civil infrastructure defects to partially replace human conducted onsite
inspections. These IPTs are primarily used to manipulate images to extract defect
features, such as cracks in concrete and steel surfaces. However, the extensively
varying real-world situations (e.g., lighting and shadow changes) can lead to
challenges to the wide adoption of IPTs. To overcome these challenges, use of deep
architecture of convolutional neural networks (CNNs) are useful. As CNNs are
capable of learning image features automatically without calculating defect feature.
In this study two slightly different CNN image classifiers are built for the purpose
of detecting and classifying cracks from concrete beams. Training and validation
was conducted using two different datasets are created for the different classifiers,
containing 40051 labelled images of 32x32 pixels and 123 labelled images of 64x64
pixels.
The performance and adaptability of the proposed approaches are tested on 10

images taken randomly from internet and onsite inspections and which are not used
for training and validation processes. The results show that the proposed methods
show quite a good performances and can indeed classify concrete cracks in realistic
situations.  
iii
4
Table of Contents
Page No.
ACKNOWLEDGEMENT ii
ABSTRACT iii
Table of Contents iv
List of Tables vii
List of Figures viii
Chapter 1
1
INTRODUCTION
1
1.1 Background
1
1.2 Objectives
2
1.3 Scope of study
2
Chapter 2
3
LITERATURE REVIEW
3
2.1 Concrete
3
2.2 Concrete cracks
3
2.3 Reasons for occurring cracks
3
2.3.1 Plastic shrinkage
4
2.3.2 Expansion and contraction
4
2.3.3 Heaving and settling
4
2.3.4 Overload
4
2.3.5 Improper drying
4
2.3.6 Natural disaster
5
2.4 Types of cracks in concrete
5
2.4.1 Shrinkage crack
5
2.4.2 Hairline crack
6
2.4.3 Settlement crack
6
2.4.4 Temperature and shrinkage crack
7
2.4.5 Shear crack
8
2.4.6 Flexural crack
8
2.5 Neural network
9
2.6 Types of neural network
9
iv
5
2.6.1 Feedforward neural network – artificial neuron

9
2.6.2 Radial basis function neural network
10
2.6.3 Kohonen self organizing neural network
11
2.6.4 Recurrent neural network(RNN)
12
2.6.5 Convolutional neural network
13
2.6.6 Modular neural network
14
2.7 Architecture of CNN
15
2.7.1 Convolution layer
15
2.7.2 Pooling layer
16
2.7.3 Activation layer
17
2.7.4 Auxiliary layers
18
2.7.5 Softmax layer
18
2.7.6 Softmax loss and stochastic gradient descent
19
2.8 CNN for real world problems
20
2.9 Recent works on crack detection using CNN
21
Chapter 3
24
COMPUTATIONAL PROCEDURE
24
3.1 Introduction
24
3.2 Computational platform
24
3.3 Crack test methodology
24
3.4 Data Collection
25
3.4.1 Standardized Data
26
3.4.2 Non-Standardized Data
27
3.5 Data Resizing and Augmentation
28
3.5.1 Image Resizing
28
3.5.2 Image Augmentation
28
3.5.3 Data Organization
28
3.6 Architecture of The CNN
30
3.6.1 Libraries Used
30
3.6.2 Keras Libraries and Packages
31
3.6.3 Convolution and Pooling Layers
31
3.6.4 Flattening and Full Connection Layer
33
3.6.5 Compiling the CNN
33
3.7 Prediction Algorithm
34
3.8 Testing accuracy of the CNNs
35
Chapter 4
36
RESULTS AND DISCUSSIONS
36
4.1 Introduction
36
4.2 Volume of layers
36
4.2 Training time vs epoch
37
v
6
4.3 Accuracy vs epoch

38
4.4 Loss vs epoch
40
4.5 Final accuracy
42
4.6 Test results
42
4.7 Summary
46
4.8 Performance study
47
Chapter 5
48
CONCLUSIONS AND SUGGESTIONS
48
5.1 Conclusions
48
5.2 Suggestions
49
REFERENCES
50
APPENDIX A (PYTHON CODES)
53
CNN_1 python codes
53
CNN_2 python code
56
APPENDIX B (PREDICTION CODE)
59
For Dataset_1
59
For Dataset_2
60
APPENDIX C (SUMMARY OF THE CNN)
61
CNN_1
61
CNN_2
62
APPENDIX D (TRAINING AND VALIDATION)
63
For Dataset 1
63
For Dataset 2
64
APPENDIX E (TIME VS EPOCH TABLES)
65
For Dataset 1
65
For Dataset 2
65
vi
7
List of Tables
Table No. Title of the Table Page No
Table 4.1: Test results for CNN_1. 44
Table 4.1: Test results for CNN_2. 46
Table 4.1: Summary of the test images. 48
vii
8
List of Figures
Figure No. Title of the Figure Page No
Figure 2.1: Shrinkage crack.

5
Figure 2.2: Hairline crack.
6
Figure 2.3: Settlement crack.
7
Figure 2.4: Temperature and shrinkage crack.
7
Figure 2.5: Shear crack.
8
Figure 2.6: Flexural crack.
9
Figure 2.7: Simple diagram of feedforward neural network.
10
Figure 2.8: Simple diagram of radial basis function neural network.
11
Figure 2.9: Simple diagram of Kohonen self organizing neural network.
12
Figure 2.10: Simple diagram of recurrent neural network.
13
Figure 2.11: Simple diagram of convolutional neural network
14
Figure 2.12: Simple diagram of modular neural network.
15
Figure 2.13: Convolution Example.
16
Figure 2.14: Pooling layer.
17
Figure 2.15: Non-linear activation functions.
17
Figure 3.1: Flowchart for detecting concrete cracks for dataset_1.
25
25
Figure 3.3: Sample Standardized crack and non crack images.
26
Figure 3.4: Sample Non-Standardized crack images.
27
Figure 3.5: Folder hierarchy for dataset_1.
29
29
Figure 3.7: General steps in constructing CNN.
30
Figure 3.8: ReLu activation function.
32
Figure 3.9: Process happening in the flattening layer.
33
Figure 3.10: Generalized image showing connection between layers.
34
viii
9
Figure 4.1: Layer diagram of CNN_1.

36
37
Figure 4.3: Training time vs Epoch (for dataset_1).
38
38
Figure 4.5: Dataset_1 accuracy vs epoch curves.
39
39
Figure 4.7: Dataset_1 loss vs epoch curves.
41
41
Figure 4.9: Test Crack images.
43
ix
10
List of Abbreviations
SHM Structural Health Monitoring
NN Neural Network
CNN Convolutional Neural Network

ANN Artificial Neural Network
RNN Recurrent Neural Network
IPT Image Processing Technique
GUI Graphical User Interface

ReLU Rectified Linear Unit
AL Active Learning
ReNet Residual Network
RGB Red, Green and Blue

API Application Programming Interface
relu Rectifier Activation Function
DB Data Base
SGD Stochastic gradient descent

ML Machine Learning
x
1
Chapter 1
INTRODUCTION
1.1 Background
Structural cracks are of major concern for structural safety. Many old structures in
the world are now in the process of decaying and is losing structural strength.
Cracks in beams and columns, piers and equivalents are like forecast of a structural
failure. That is why structural cracks are dealt with great importance and detecting
cracks is the first step of retrofitting these structures.
In developing and least developed countries many cases are found which shows that
residential, commercial and institutional buildings are not designed and constructed
as per code. As a result the constructed building does not gain targeted strength and
design life putting peoples’ lives at risk.
Failure in beam results in localized failure of a structure. So identifying and

classifying the type of cracks generated in a beam will help to take immediate
action to retrofitting it thus saving further damage to the structure and loss of lives.
People living in low income countries are bound to live in risk prone building and
there are lack of building inspectors. So giving them a tool or technology which is
low cost and easy to use will help to determine which cracks to ignore and which
cracks to take seriously and then contact appropriate authorities to take action and
thus preventing future catastrophes.
Several SHM systems have been developed to detect cracks from a concrete
structure using X-ray, vibrations, sensors etc. Due to low price and availably of
smartphones with camera it is now a possibility to help mass people to check

2
structural health of their own occupying building by technologies developed using
image processing techniques.
1.2 Objectives
The main objective of this study is to develop a model using Convolutional Neural
Network (CNN) to detect and classify cracks in concrete beams.
Specific objectives of the study includes:
i. Constructing two image classifier using CNN. One will be called CNN_1
and other CNN_2. Both of them are categorical classifier.
ii. Creating two dataset of images for two slightly different image classifiers.
One will be called dataset_1 and the other will be called dataset_2.
iii. Developing an algorithm for the dataset_1 for CNN_1 classifier, which
utilizes the input of crack location and outputs crack type.
iv. Comparing results and test accuracy between this two classifiers.
1.3 Scope of study
In this study a detailed literature review was conducted concerning concrete cracks
and various types of CNN. Then two suitable CNN models was architected and
constructed to work as a concrete beam crack image classifier. Two different
datasets was created to be training and validated by the classifiers. After that
comparisons was made between the classifiers with regard to training time,
performance, accuracy and real world compatibility. And finally, future suggestions
was made for enriching the datasets and increasing the performances of the
classifiers.
3
Chapter 2
LITERATURE REVIEW
2.1 Concrete
Concrete is inert mass which grows from a cementing medium. Concrete is a

product of two major components, one is the cement paste and another is the inert
mass. In order to form the cementing medium, cement would mix with water.
Coarse aggregates and fine aggregates are the part of inert mass. In properly mixed
concrete, these materials are completely surrounded and coated by cement paste
filling all the void space between the particles [1]. With time, the setting process of
the concrete starts and it starts to gain its strength.
2.2 Concrete cracks
Some concrete professionals believe that reinforced concrete structures should not
crack. With that belief, when cracking does occur, they often claim that the concrete
contractor caused the cracks and should pay for repair. Cracks in reinforced
concrete, however, are not a defect but are specifically included as part of the
design process. Reinforced concrete structures normally crack when carrying

service loads [2].
2.3 Reasons for occurring cracks
Whether it’s roads, driveways, sidewalks, or parking slabs, eventually, almost every
kind of concrete cracks. It’s an inevitable element of the wear and tear suffered by a
property – the same as paint peeling, garden plants dying, or walls getting dirty.
Some main reasons for occurring concrete cracks are given below [3].
4
2.3.1 Plastic shrinkage
Concrete in its non-solid state is basically just water and cement. As it hardens and
the water evaporates, it shrinks. Too much water means too much shrinkage – and
too much shrinkage leads to cracks in the finished project. This is usually a
cosmetic flaw rather than a structural one.
2.3.2 Expansion and contraction
Like every solid, concrete tends to change subtly in shape and size based on
weather conditions. In hot weather concrete expands, pushing against anything in its
path as it does so. If it meets an object that doesn’t flex or give during that
expansion, it will likely crack.
2.3.3 Heaving and settling
Grounds movement is one of the most frequent causes of cracking, especially with
sidewalks, driveways, and roads. The growth of a tree’s roots or an excessive freeze
and thaw cycle can cause the ground to push upwards on the concrete, causing it to
crack and break – a process known as heaving. Consequently, shifting in soil due to
water or the rotting away of a tree’s roots cause the foundation of concrete to
collapse – this is known as settling. As with expansion and contraction, heaving and
settling is likely to occur over time.
2.3.4 Overload
If too much weight is placed on concrete, it can cause cracks or in extreme cases
crush the concrete.
2.3.5 Improper drying
Crazing and crusting occur when laid concrete is improperly dried. The former
results from concrete not having enough moisture or losing moisture too quickly,
5
while the latter results from a pattern being pressed into concrete before it is
properly dried.
2.3.6 Natural disaster
Natural disaster like earthquake or heavy wind load might cause cracks to occur if
not properly designed following building codes. It may lead to reduced load bearing
capacity of the concrete component and in extreme cases, structural failure.
2.4 Types of cracks in concrete
Tremendous forces can build up inside the wall due to any causes of cracks. When
the forces exceed the strength of the concrete, cracks will develop. Each of these
causes leaves a "signature" typically in the type of crack it creates. Some of the
major and most common cracks are discussed here [4, 5].
2.4.1 Shrinkage crack
Shrinkage cracks in a poured concrete foundation can be diagonal or vertical and
are usually uniform in width. Sometimes these cracks have a V-shape (less
frequent), with the top of the crack looking larger and the crack getting smaller as it
travels towards to floor and diminishing or stopping before reaching the bottom of
the foundation wall. If the crack reaches the bottom, the crack might damage the
building's footings, and the crack might have a significant impact on the foundation
structure. Figure 2.1 shows a sample of shrinkage crack on concrete surface.
Figure 2.1: Shrinkage crack.  

6
Shrinkage cracks should not be ongoing nor of structural significance, though they
may invite water entry through the wall. In poured concrete foundations, shrinkage
cracks are usually due to conditions at original construction: poor concrete mix,
rapid curing or possibly other states. In any case, concrete shrinkage causes the
concrete to develop internal stresses. Unless control joints were included in the wall
or floor slab design, these stresses would cause the wall or floor is likely to crack in
a classic "concrete shrinkage pattern" as the concrete cures.
2.4.2 Hairline crack
Hairline cracks may develop in concrete foundations as the concrete cures. Hairline
cracks do not cause problems with the stability of the foundation but do cause
leakage problems. If the cracks appear shortly after pouring the concrete
foundation, concrete may have been mixed poorly or poured too quickly. In poured
concrete foundations, hairline crack frequently appear in the center of the walls
because the wall corners have greater stability.
Figure 2.2: Hairline crack.
2.4.3 Settlement crack
Settlement cracks may appear when the underlying ground has not been compacted
or appropriately prepared or if the subsoil was not of the proper consistency. A
settlement crack may also appear as a random crack above areas where the soil of
the subgrade was uneven after the concrete was poured. Settlement cracks are
7
Figure 2.3: Settlement crack.
usually more extensive at the top of the crack than the bottom as the foundation
"bends" over a single point, allowing differential settlement. This type of crack is
usually continuous and may occur multiple times in a wall.
2.4.4 Temperature and shrinkage crack
Horizontal cracks found in the center of the wall are most likely caused by an
applied load such as backfill around foundation compacted improperly or too soon,
earth compacting as it settles, hydrostatic pressure against foundation due to high
water table and poor drainage against the foundation wall, or heavy equipment
operated too soon or too close to the foundation wall. Horizontal cracks found high
up on the wall are most likely caused by frost damage.
Figure 2.4: Temperature and shrinkage crack.

8
2.4.5 Shear crack
Shear cracks form near the supports of members and are inclined at between about
30o and 45o to the axis of the beam, from the tension face of the member back
towards the support. As with flexural cracks, they will be widest at the tension face,
reducing in width with distance from the face. When the member has been correctly
designed, shear cracks will be controlled by the provision of shear reinforcement in
the form of links.
Figure 2.5: Shear crack.
2.4.6 Flexural crack
Flexural cracks on the sides of a beam start at the tension face and will extend, at
most, up to the neutral axis. Crack widths will be greatest at the tension face and
will reduce with distance from that face. In general, the cracks will be uniformly
spaced along the most heavily loaded portion of the beam, i.e. near the mid-span in
sagging or over the supports in hogging.Flexural cracks on the soffit of a slab will
run at right-angles to the span, again roughly uniformly spaced in the region of
maximum moment. In beams and slabs in buildings that have been correctly
designed, average crack widths should not exceed 0.3 mm. In bridges, the cracks
should generally be narrower. 

9
Figure 2.6: Flexural crack.
2.5 Neural network
Artificial intelligence technique that mimics the operation of the human brain
(nerves and neurons), and comprises of densely interconnected computer processors

working simultaneously (in parallel). A key feature of neural networks is that they
are programmed to 'learn' by sifting data repeatedly, looking for relationships to
build mathematical models, and automatically correcting these models to refine
them continuously. Also called neural net [6].
2.6 Types of neural network
Artificial neural networks are computational models which work similar to the
functioning of a human nervous system. There are several kinds of artificial neural
networks. These type of networks are implemented based on the mathematical
operations and a set of parameters required to determine the output. Some of the
types that are most used in today’s machine learning (ML) applications are given
below [7].
2.6.1 Feedforward neural network – artificial neuron
This neural network is one of the simplest form of ANN, where the data or the input
travels in one direction. The data passes through the input nodes and exit on the
output nodes. This neural network may or may not have the hidden layers. In simple
words, it has a front propagated wave and no back propagation by using a
classifying activation function usually. Figure 2.7 shows a Single layer feed forward
10
Figure 2.7: Simple diagram of feedforward neural network.
network. Here, the sum of the products of inputs and weights are calculated and fed
to the output. The output is considered if it is above a certain value i.e
threshold(usually 0) and the neuron fires with an activated output (usually 1) and if
it does not fire, the deactivated value is emitted (usually -1). Application of Feed
forward neural networks are found in computer vision and speech recognition where
classifying the target classes are complicated. These kind of Neural Networks are
responsive to noisy data and easy to maintain.
2.6.2 Radial basis function neural network
Radial basic functions consider the distance of a point with respect to the center.
RBF functions have two layers, first where the features are combined with the
Radial Basis Function in the inner layer and then the output of these features are
taken into consideration while computing the same output in the next time-step
which is basically a memory. Figure 2.8 is a diagram which represents the distance
calculating from the center to a point in the plane similar to a radius of the circle.
Here, the distance measure used in euclidean, other distance measures can also be
used. The model depends on the maximum reach or the radius of the circle in
classifying the points into different categories. If the point is in or around the
radius, the likelihood of the new point begin classified into that class is high. There
can be a transition while changing from one region to another and this can be
11
controlled by the beta function. This neural network has been applied in Power
Restoration Systems. Power systems have increased in size and complexity. Both
factors increase the risk of major power outages. After a blackout, power needs to
be restored as quickly and reliably as possible.
Figure 2.8: Simple diagram of radial basis function neural network.
2.6.3 Kohonen self organizing neural network
The objective of a Kohonen map is to input vectors of arbitrary dimension to

discrete map comprised of neurons. The map needs to me trained to create its own
organization of the training data. It comprises of either one or two dimensions.
When training the map the location of the neuron remains constant but the weights
differ depending on the value. This self organization process has different parts, in
the first phase every neuron value is initialized with a small weight and the input
vector. In the second phase, the neuron closest to the point is the ‘winning neuron’
and the neurons connected to the winning neuron will also move towards the point
like in the graphic below (Fig 2.9). The distance between the point and the neurons
is calculated by the euclidean distance, the neuron with the least distance wins.
Through the iterations, all the points are clustered and each neuron represents each
kind of cluster. This is the gist behind the organization of Kohonen Neural Network.
12
Figure 2.9: Simple diagram of Kohonen self organizing neural network.
Kohonen Neural Network is used to recognize patterns in the data. Its application
can be found in medical analysis to cluster data into different categories. Kohonen
map was able to classify patients having glomerular or tubular with a high accuracy.
2.6.4 Recurrent neural network(RNN)
The Recurrent Neural Network works on the principle of saving the output of a
layer and feeding this back to the input to help in predicting the outcome of the
layer. Here, the first layer (Figure 2.10) is formed similar to the feed forward neural
network with the product of the sum of the weights and the features. The recurrent
neural network process starts once this is computed, this means that from one time
step to the next each neuron will remember some information it had in the previous
time-step. This makes each neuron act like a memory cell in performing
computations. In this process, it is needed to let the neural network to work on the
front propagation and remember what information it needs for later use. Here, if the
prediction is wrong the learning rate or error correction is used to make small
changes so that it will gradually work towards making the right prediction during
the back propagation.

13
Figure 2.10: Simple diagram of recurrent neural network.
The application of Recurrent Neural Networks can be found in text to speech(TTS)
conversion models.
2.6.5 Convolutional neural network
Convolutional neural networks are similar to feed forward neural networks , where
the neurons have learn-able weights and biases. Its application have been in signal
and image processing which takes over OpenCV in field of computer vision. Below
is a representation of a ConvNet, in this neural network, the input features are taken
in batch wise like a filter. This will help the network to remember the images in
parts and can compute the operations. These computations involve conversion of
the image from RGB or HSI scale to Gray-scale. Once we have this, the changes in
the pixel value will help detecting the edges and images can be classified into
different categories. ConvNet are applied in techniques like signal processing and
image classification techniques. Computer vision techniques are dominated by
convolutional neural networks because of their accuracy in image classification.

14
The technique of image analysis and recognition, where the agriculture and weather
features are extracted from the open source satellites like LSAT to predict the future
growth and yield of a particular land are being implemented.
Figure 2.11: Simple diagram of convolutional neural network
2.6.6 Modular neural network
Modular Neural Networks have a collection of different networks working
independently and contributing towards the output. Each neural network has a set of
inputs which are unique compared to other networks constructing and performing
sub-tasks. These networks do not interact or signal each other in accomplishing the
tasks. The advantage of a modular neural network is that it breakdowns a large
computational process into smaller components decreasing the complexity. This

breakdown will help in decreasing the number of connections and negates the
interaction of these network with each other, which in turn will increase the
computation speed. However, the processing time will depend on the number of
neurons and their involvement in computing the results. Modular Neural Networks
(MNNs) is a rapidly growing field in artificial Neural Networks research. It is
getting popular in solving biological, psychological, hardware, and computational
problems.  
15
Figure 2.12: Simple diagram of modular neural network.
2.7 Architecture of CNN
A CNN model has several layers which has a specific job. Most commonly used
layers are, convolution layer, pooling layer, auxiliary layers, softmax layer.
2.7.1 Convolution layer
A convolution layer performs the following three operations throughout an input

array as shown in Figure 2.13. First, it performs element-by-element multiplications
(i.e., dot product) between a subarray of an input array and a receptive field. The
receptive field is also often called the filter, or kernel. The initial weight values of a
receptive field are typically randomly generated. Those of bias can be set in many
ways in accordance with networks’ configurations [8]. Both values are tuned in
training using a stochastic gradient descent (SGD) algorithm. The size of a
subarray is always equal to a receptive field, but a receptive field is always smaller
than the input array. Second, the multiplied values are summed, and bias is added to
the summed values. Figure 2.13 shows the convolutions of the subarrays (solid and
dashed windows) with an input array and a receptive field. One of the advantages of
the convolution is that it reduces input data size, which reduces computational cost.
An additional hyper parameter of the layer is the stride. The stride defines how
16
many of the receptive field’s columns and rows (pixels) slide at a time across the
input array’s width and height. A larger stride size leads to fewer receptive field
applications and a smaller output size, which also reduces computational cost,
though it may also lose features of the input data. The output size of a convolution
layer is calculated by the equation shown in Figure 2.13.
Figure 2.13: Convolution Example.
2.7.2 Pooling layer
Another key aspect of the CNNs is a pooling layer, which reduces the spatial size of
an input array. This process is often defined as downsampling. There are two
different pooling options. Max pooling takes the max values from an input array’s
subarrays, whereas mean pooling takes the mean values. Figure 2.24 shows the
pooling method with a stride of two, where the pooling layer output size is
calculated by the equation in the figure. Owing to the stride size being larger than
the convolution example in Figure 2.14, the output size is further reduced to 3 × 3.
Max pooling performance in image data sets is better than that of mean pooling [9].
This article verified that the architecture with max pooling layers outperforms those
with mean pooling layers. Thus, all the pooling layers for this study are max
pooling layers.
17
2.7.3 Activation layer
The most typical way to give nonlinearity in the standard ANN is using sigmoidal
functions, such as y = tanh(x), but saturating nonlinearities slows computations
[10]. Recently, the ReLU was introduced [10] as a nonlinear activation function.
Figure 2.15 depicts the several examples of nonlinear functions. Briefly, while other
nonlinear functions are bounded to output values (e.g., positive and negative ones
and zeros), the ReLU has no bounded outputs except for its negative input values.
Intuitively, the gradients of the ReLU are always zeros and ones. These features
facilitate much faster computations than those using sigmoidal functions and
achieve better accuracies.
Figure 2.14: Pooling layer.
Figure 2.15: Non-linear activation functions.

18
2.7.4 Auxiliary layers
Overfitting has been a long-standing issue in the field of machine learning. This is a
phenomenon where a network classifies a training data set effectively but fails to
provide satisfactory validation and testing results. To address this issue, dropout
layers are used [11]. Training a network with a large amount of neurons often
results in overfitting due to complex coadaptations. The main idea of dropout is to
randomly disconnect the connections between neurons of connected layers with a
certain dropout rate. Accordingly, a network can generalize training examples much
more efficiently by reducing these coadaptations. A well-known trick, taking the

average values of a training data set (i.e., whitening), has often been used to shorten
network training time [12]. However, the distribution of layer’s input shifts by
passing through layers, which is defined as internal covariate shift, and this has
been pointed out as being the major culprit of slow training speed. A BN was
proposed to adapt the similar effect of whitening on layers [13]. As a result, this
technique facilitates high-learning rate and leads to much faster network
convergence.
2.7.5 Softmax layer
To classify input data, it is necessary to have a layer for predicting classes, which is
usually located at the last layer of the CNN architecture. The most prominent
method to date is using the softmax function given by the equation given below,
which is expressed as the probabilistic expression p(y( i ) = n | x( i ) ;W) for the ith
training example out of m number of training examples, the jth class out of n
number of classes, and weights W, where Wn T x(i) are inputs of the softmax layer.
The sum of the right-hand side for the ith input always returns as 1, as the function
always normalizes the distribution. In other words, this equation returns
probabilities of each input’s individual classes.

19
2.7.6 Softmax loss and stochastic gradient descent
As the initial values of W are randomly assigned during training, the predicted and
actual classes do not usually coincide. To calculate the amount of deviations
between the predicted and actual classes, the softmax loss function is defined by a
equation.
The new index L is introduced to indicate that ∑nl=1 exp(WTx(i)) is independent of
∑nj=11{.}. The term 1{y(i)=j} is the logical expression that always returns either
zero or ones. In other words, if a predicted class of the ith input is true for j class,
the term returns ones, returning zeros otherwise. The last hyper-parameter λ in the
equation is a regularization (i.e., weight decay) parameter to penalize large weights,
which is also a well-known trick for preventing overfitting [14, 15].

20
To narrow the deviations, an algorithm that updates receptive field weights is
necessary for obtaining the expected results (i.e., predicting true classes). This
process is considered for CNN training. There are several known methods, but SGD
using back-propagation is considered the most efficient and simplest way to
minimize the deviations [12]. The standard gradient descent algorithm performs
updating W on an entire training data set, but the SGD algorithm performs it on
single or several training samples. To accelerate the training speed, the momentum
algorithm [14] is also often used in SGD. The overall updating process is as
follows. First, the gradient ∇W of a loss function is calculated with respect to W.
Second, the hyper-parameters of momentum ε and learning rate α are introduced to
update (←) velocity υ, where momentum is defined as mass times velocity in

physics, but with unit mass being what is considered in SGD. Finally, the weights
are updated. A network can be tuned by repeating the explained process several
times until Wj←Wj+υ converges. The superscript (i) indicates the ith training
sample, where the range of i is dependent on a mini-batch size, which defines how
many training samples out of the whole data set are used. For example, if 100
images are given as the training data set and 10 images are assigned as the mini-
batch size, this network updates weights 10 times; each complete update out of the
whole data is called an epoch.
2.8 CNN for real world problems
Many types of artificial neural networks (ANNs), including the probabilistic neural
network (NN) [16], have been developed and adapted to research and industrial
fields, but convolutional neural networks (CNNs) have been highlighted in image
21
recognition, which are inspired by the visual cortex of animals [17]. CNNs can
effectively capture the grid-like topology of images, unlike the standard NNs, and
they require fewer computations due to the sparsely connected neurons and the
pooling process. Moreover, CNNs are capable of differentiating a large number of
classes [18]. These aspects make CNNs an efficient image recognition method [19,
20]. The previous issue of CNNs was the need for a vast amount of labeled data,
which came with a high-computational cost, but this issue was overcome through
the use of well-annotated databases (ImageNet, no date; CIFAR-10 and CIFAR-100
data set, no date; MNIST Database, no date) and parallel computations using
graphic processing units [21]. Owing to this excellent performance, a study for
detecting railway defects using a CNN was later proposed [22]. Rail surfaces are
homogenous, and the images are collected under controlled conditions. This cannot
be considered the same as detecting concrete surface defects due to non-

homogenous surface. Therefore, a carefully configured deep architecture and
abundant data set, taken under extensively varying conditions, is essential for
dealing with the true variety of real-world problems.
2.9 Recent works on crack detection using CNN
There has been a number of research in detecting cracks automatically on concrete

surfaces using deep neural networks to reduce human conducted on site inspections.
Some are in basic levels others focusing on increasing accuracy.
In 2017 Cha et al. [23] proposed a vision based method using a deep architecture of
Convolutional Neural Networks (CNNs) for detecting concrete cracks without
calculating the defect features. Which allowed the architecture to detect cracks in
concrete surfaces in the extensively varying real-world situations (e.g., lighting and
shadow changes). The CNN has three convolutional layers, 3 pooling layers, uses
Rectified Linear Unit (ReLU) at the end of the hidden layers and uses max pooling.
To classify input data into binary output the CNN architecture uses softmax
function. The overall architecture ends up detecting cracks with 98% accuracy.
22
In 2017 Feng et al. [24] proposed a deep Active Learning (AL) framework.
Addressing the efficient training and deployment of an automatic defect detection
system. A deep residual network (ResNet) is designed as the classifier for detection
and classification of defects in an input image patch. In their paper, they prepared a
labelled dataset, then fed it to an existing deep neural architecture, designed a
suitable cost function for optimization and tuned hyper parameters for training to
achieve a best performance. The model is able to detect cracks, deposit and water
leakage on concrete surface with an accuracy of 87.5%.
In 2013 Salman et al. [25] proposed an approach to automatically distinguish cracks

in digital images based on the Gabor filtering. Multi- directional crack detection can
be achieved by high potential Gabor filter. The Gabor filter is a highly potential
technique for multidirectional crack detection. The image analysis of the Gabor
filter function was directly related to the manual visual perception. Once filtering
was completed, the cracks aligned to different directions are detected. They have a
detection precision of 95% for their proposed methodology.
In 2005 Iyer et al. [26] have designed a three-step method for the crack detection
from the high contrast images. The proposed method detects the crack like pattern
in the noisy environment using curvature evaluation and mathematical morphology
technique. It was based on mathematical morphology and curvature evaluation that
detects crack-like patterns in a noisy environment. In their study, segmentation is

done defining the crack like pattern with respect to a precise geometric model.
Linear filtering was performed after cross curvature evaluation to distinguish them
from analogous background pattern. They have identified the irregularity
sequentially by the Geometry-based recognition crack features.
In 2014 Nguyen et al. [27] have proposed a method based on the edge detection of
concrete cracks from noisy 2D images of concrete surfaces. They have observed the
23
cracks as tree-like topology. Then based on the PSCEF non-crack objects were
removed. After the separation, thresholding filter, and morphological thinning

algorithm have been used to binarize the image for the crack centre line estimation.
Then the centre line was fitted by cubic splines. They have linked the edge points to
form the desired continuous crack edge. From the crack edge, the surface of the
crack was attained.
In 2015 Yang et al. [28] have proposed an image analysis method to capture thin
cracks and minimize the requirement for pen marking in reinforced concrete
structural tests. They have used the studies like crack depth prediction, change in
detection without image registration, crack pattern recognition based on artificial
neural networks, applications to micro-cracks of rocks , and efficient sub-pixel
width measurement. Stereo triangulation method was the adopted technique based
on cylinder formula approximation and image rectification. Once they have the
rectified output, the surface of the observed regions can be unfolded and presented
in a plane image for following displacement and deformation analysis. From which
the crack detection was analyzed.
In 2010 Wang et al. [29] have proposed a system for the image based crack
detection and to characterize the crack based upon their effectiveness. They have
categorized the present image based crack detection into four categories. They are
an integrated algorithm, morphological approach, percolation approach and
practical technique. A shading correction was done using integrated algorithm. The
unclear crack prediction was detected using percolation method. The crack
detection was done using morphological approach for the micro crack detection
with the practical method providing high- performance feature extraction with more
than 95% of accuracy. 

24
Chapter 3
COMPUTATIONAL PROCEDURE
3.1 Introduction
The main purpose of this study is to make CNN architectures and datasets which in
combine will be able to determine and classify beam cracks into shear cracks,
flexural cracks, shrinkage cracks and spalling of concrete with precision and high
accuracy. For this reason, firstly, a CNN model will be introduced and then a dataset
containing images with different crack labels will be created. Secondly, an
alternative dataset containing only cracks, non-cracks and spalling images will be
created to be trained and validated by a slightly different CNN model and a

relationship between length of the beam and crack location from support will be
developed which will help to predict the classification of the cracks.
3.2 Computational platform
For this study following platform was used

Operating System: mac osx 10.14
CPU: 2.5 GHz Intel Core i5
GPU: none
RAM: 8 GB 1600 MHz DDR3

Programming Language: Python 3.6
Software: Spyder
3.3 Crack test methodology
Two slightly different test methodologies are followed for two different type of
crack image classifiers. Dataset_1 is the dataset containing standardized crack and
non-crack images and non-standardized spalling images to be trained and validated
by CNN_1.
25
Dataset_2 contains non-standardized shear, flexural cracks and spalling on concrete
surfaces images to be trained and validated by CNN_2.
Output Result:
Convolutional
Shear cracks,
Input Image Preprocessing Neural
Flexural cracks
Network
or Spalling
3.4 Data Collection
Both standardized and non-standardized images were collected. For creating

labelled dataset of shear cracks, flexural cracks and spalling of concrete random
26
non-standardized datas from internet was collected. And for creating the dataset
containing cracks and non crack images, standardized cracks and non-cracks images
was collected [30]. All images used in the study are in .jpg format.
3.4.1 Standardized Data
For classifying beam cracks into binary crack or non-crack, a standardized dataset
was used. The dataset contains concrete images having cracks. The data is collected
from various METU (Middle East Technical University) Campus Buildings. The
dataset is divided into two as negative and positive crack images for image
classification. Each class has 20000 images with a total of 40000 images with 227 x
227 pixels with RGB channels (Fig 3.3). The dataset is generated from 458 high-
resolution images (4032x3024 pixel) with the method proposed by Zhang et al [31].
High-resolution images have variance in terms of surface finish and illumination
conditions. No data augmentation in terms of random rotation or flipping is applied.
a) Crack images
b) Non crack images
Figure 3.3: Sample Standardized crack and non crack images.

27
3.4.2 Non-Standardized Data
A total of 123 images was collected randomly from the internet without following
any specific procedure (samples shown in Figure 3.4). Then they were cropped
around the specific type of cracks the image was carrying. From which 40 are of
shear cracks images on concrete beam surface, 32 are of flexural cracks on concrete
beam surfaces and 51 are of spalling of concrete on concrete beam surfaces. These
images are of RGB channels with varying pixel sizes. No Data augmentation in
terms of random rotation or flipping is applied.
Figure 3.4: Sample Non-Standardized crack images.

28
3.5 Data Resizing and Augmentation
Due to memory, time and computational capacity constraints, the images from the
dataset had to be resized (both standardized and non-standardized datas). And due
to small number of images in both dataset (CNN requires large number of images)
images had to be augmented to increase their numbers.
3.5.1 Image Resizing
Images from standardized dataset are of 227x227 pixels. Images from non-
standardized dataset varies from 199x71 pixels to 2816x2112 pixels before
cropping. This varieties of pixels can not be used for testing purpose, rather a single
size is needed for all the images. Here in this study, all the input images are resized
into 32x32 pixels in CNN_1 and 64x64 pixels in CNN_2 considering time and
resource constraints.
3.5.2 Image Augmentation
In this study image datas were augmented to increase image numbers in order to
increase accuracy. Every images from the datasets(except in dataset_2, rotation was
avoided to prevent confusion between shear and flexural cracks) were randomly
rescaled, zoomed, rotated, sheared and flipped horizontally-vertically using keras
‘ImageDataGenerator’ class [32].
3.5.3 Data Organization
Keras library is being used for training and validation purpose.Keras is an open
source neural network library written in Python. It is capable of running on top of
TensorFlow, Microsoft Cognitive Toolkit or Theano. Designed to enable fast

experimentation with deep neural networks [33]. The total images are organized in a
special way by which Keras library is able to recognize which type of cracks it is
and continue the computational process.

29
For the standardized dataset, the images are put in two folders naming ‘training_set’
and ‘val_set’ for testing and validation purpose. In each folder, there are two
folders, ‘Crack’ folders contains crack images and ‘noncrack’ folder contains non-
crack images. All the images in the folder are labeled with the type and a number to
give them a serial which will help keras library to read the images. A third subfolder
was added in both ‘training_set’ and ‘val_set’ folders named ‘Spalling’ for detecting
spalling on concrete surface which contains non-standardized spalling images and
the images were labelled and numbered as previously stated (Figure 3.5).

30
For the non-standardized dataset, similar procedure was done to divide the images
into ‘training_set’ and ‘val_set’. ‘Flexural’, ‘Shear’ and ‘Spalling’ subfolders were
created to accommodate flexural, shear and spalling cracks and they were labelled
and numbered as was done with standardized dataset (Figure 3.6).
3.6 Architecture of The CNN
The proposed CNN contains two convolutional layers, two pooling layers, one
flattening layer and finally one full connection layer. For constructing the CNN
model Keras library was used and the packages that were used are: Sequential,
Conv2D, MaxPooling2D, Flatten and Dense. Building the CNN involved using
various library and packages. For faster computation, Tensorflow

backend was used. Layers constituting the CNNs used in this study is shown in
Figure 3.7.
Figure 3.7: General steps in constructing CNN.
3.6.1 Libraries Used
In this study, three python libraries were used named: Theano library, Keras library
and Tensorflow library. So they had to be installed on the computer. Keras library
runs on Theano and Tensorflow library.
31
3.6.2 Keras Libraries and Packages
Sequential, Conv2D, MaxPooling2D, Flatten, Dense packages and matplotlib
library were used for this study.The CNN is a sequence of layers, and so sequential
package was used to initialize the CNN.
Conv2D is used to build convolutional layer. Since the study deals with 2D images,
2D convolutional layers has been used. In this study max pooling has been used.
That is why MaxPooling2D package was used. For flattening purpose, Flatten was
used. Flattening is the layer in which all the max pooled feature maps are converted
into a large feature vector which then works as an input for the fully connected
layer. Dense package was used to create a fully connected layer and a classic ANN.
Matplotlib library was used to plot graphs.
3.6.3 Convolution and Pooling Layers
The CNN architecture contains two 2D convolution layers each followed by a 2D
max pooling layer. The CNN is named ‘classifier’ as the purpose of the NN to
classify concrete beam cracks.
For the first convolution layer, 32 feature detectors of 3x3 matrix was used as it is
common practice to start the first convolution layer with 32 feature detector. And as
mentioned earlier, input image was of the size of 32x32 pixels(for CNN_1 and for
CNN_2 64x64) and the activation function used was ‘relu’ (Figure 3.8) to discard
negative pixel values in the feature map. Discarding negative pixels will introduce
non-linearity in the CNN as classifying object is a non-linear problem.
Pooling layer helps to reduce the size of the feature maps. In the CNN model a 2x2
max pooling layer was used with a stride of 2. No padding was used. The reason
behind this is so that no information is lost from the input image.

Output image shape can be determined by the following formula :
32
Figure 3.8: ReLu activation function.
m−f
Height of padded feature map, nH = +1
s
n− f
Width of the padded feature map, nw = +1
s
Here,
m = number of pixels in the height of the image
n = number of pixels in the width of the image

f = size of the pooling
s = number of stride
After each layer, size of the input image become, nH x nW x number of feature
detectors.
For the second layer, the same convolution and pooling layer was used except that
here 32x2 or 64 feature detectors were used. As it is a common practice to increase

the number of feature detector 100% in each convolution layer.
33
3.6.4 Flattening and Full Connection Layer
Flattening layers takes in all the pooled feature maps and put them in one single
column vector. A sample procedure is shown in Figure 3.9. And this column vector
will be the input layer for the full connection layer. Dense function was used to add
a fully connected layer.
Number of output nodes in hidden layer is chosen to be 128 as it is a common
practice to chose a number which is a power of 2. Larger nodes was not selected
due to computational capacity constraints. Relu activation function was used.
For both standardized and non-standardized dataset three final output is expected.
That is why in the final output layer, three output node is selected for both cases
and softmax function was used for this purpose. Here softmax function was used as
more than two outcome is expected.
Figure 3.9: Process happening in the flattening layer.
3.6.5 Compiling the CNN
The whole CNN model is compiled using Stochastic gradient descent algorithm, a
loss function and a performance matrices.

34
For adding Stochastic gradient descent algorithm, optimizer parameter was used
and the algorithm used was ‘adam’. As the loss function, ‘categorical cross entropy’
was chosen as the CNN deals with categorical problems. For performance matrices,
the most common ‘accuracy’ was chosen. A generalized diagram is shown in Figure
3.10 to understand the connection between layers after the flattening layer.
Figure 3.10: Generalized image showing connection between layers.
3.7 Prediction Algorithm
For prediction purpose, two different types of approaches are selected for dataset_1(
Standardized crack and non-crack images and non standardized spalling images)
and dataset_2( Non-standardized shear, flexural and spalling images).
For dataset_2, direct output is shown which results from the final output layer of the
CNN. Output categories shown here are ‘Flexural’, ‘Shear’ and ‘ Spalling’. A
random probability of 0.4 is selected as threshold value for these outputs.
For dataset_1, a more elaborated approach is selected. At first the CNN gives result
whether an input test image contains cracks, noncracks or spalling. If the output
shows ‘cracks’, then a ‘length of beam’, ‘depth of crack’ and ‘edge of crack from
35
nearest support’ is developed to classify whether the crack is ‘Shear’, ‘Flexural’ or
‘Shrinkage’ crack.
If the location of crack and length of beam is less than 30%, the input image will be
shown as ‘Shear Crack’. If the crack resides within 40% to 50% of the beam length
and the crack is in full depth, the output will show ‘Both Flexural and Shrinkage
Crack’ and if the crack is not in full depth of the beam, ‘Flexural Crack’ will be
shown. For ‘Shrinkage Crack’ output, the crack have be within 30% to 40% of the
beam length and full depth [34]. But for detecting spalling of concrete, only the test
image is sufficient.
3.8 Testing accuracy of the CNNs
Finally the accuracy of the two CNN was measured using some parameters.
Accuracy measures overall how often the model is correct. It should be as high as
possible.
TP + T N
Accuracy =
Total nu m ber of test im ag es
Here,
TP = True positive, when the predicted positive value equals the actual positive
value.
TN = True negative, when the predicted negative value equals the actual negative
value.
36
Chapter 4
RESULTS AND DISCUSSIONS
4.1 Introduction
Performance of CNNs are compared in this chapter with regard to training time,
accuracy and loss. Test images are used which are not in the training or validation
datasets to test and compare the compatibility of the CNNs in real world situations.
4.2 Volume of layers
Layer diagrams for CNN_1 is shown in figure 4.1. Here it can be seen that the input
image has a volume of 32x32x3. After applying convolution layer-1 which is of
size 3x3, with 32 filters and stride of 1, the height and width becomes,
n− f 32 − 3
n= + 1= + 1= 30
s 1
Output volume in each layers can be calculated in this procedure. And finally it
gives three outputs, ‘Crack’, ‘Noncrack’ and ‘Spalling’.

37
Similarly, output volume of each layers for CNN_2 can be calculated. Layer
diagrams for CNN_2 is shown in Figure 4.2.
4.2 Training time vs epoch
In this study 10 epoch was performed for each dataset. In Deep Learning, an epoch
is a hyper-parameter which is defined before training a model. One epoch is when
an entire dataset is passed both forward and backward through the neural network
only once [35]. The code needed for training and validation purpose is slightly for
both dataset. Batch size, which was chosen to be 32 in case for dataset_1(because
dataset_1 has large number of images and it is feasible considering computational
capacity constraints) and 3 in the case for dataset_2(because dataset_2 has a small
number of dataset). So time needed to perform an epoch gives us an idea of the
volume of calculations necessary for each dataset.
From the graphs(Fig 4.3, Fig 4.4) it is visible that both the graphs are almost linear.
So the relationship between training time and epoch is a liner one. Here it can be
seen that training time needed for dataset_1 is much higher than dataset_2, it is
because dataset 1 contains much higher number of images. Also, training time
might have been effected due to overclocking of the CPU and other side tasks.
38
4.3 Accuracy vs epoch
There are two types of accuracy involved in NN (Neural Network) ‘training
accuracy’ and ‘validation accuracy’. After 10 epochs, training and validation
accuracy for both dataset are noted.

39
In case of dataset 1, it can be seen that training accuracy gradually increases (Fig
4.5) with epoch and then reach a pick point and then becomes a straight line. Test
accuracy reach a pick point and then fluctuates between a range as number of epoch
increases.
Figure 4.6: Dataset_2 accuracy vs epoch curves. 

40
In case of dataset_2, test accuracy curve (Fig 4.6) has the same behavior as the
accuracy curve of dataset_1. But in case of training accuracy, it is still trending, It

will probably take more epochs to reach to a saturation point.
4.4 Loss vs epoch
Loss function used for this study is ‘categorical cross entropy’. The equation for
this loss function is:
where, p=true distribution and q=coding distribution.

Loss involved in NN is of two types, ‘training loss’ and ‘validation loss’. After 10
epochs, training and validation loss for both dataset are shown below.
In case of dataset_1, it can be seen that training loss gradually decreases with epoch
and eventually becomes a straight line (Fig 4.7). Test loss reach a lowest point and
then fluctuates between a range as number of epoch increases.
In case of dataset_2, training loss curve has the same behavior as the training loss
curve of dataset_1. But in case of test loss, it is still trending, It will probably take
more epochs to reach to a saturation point and fluctuate between two points (Fig
4.8).
41
Figure 4.8: Dataset_2 loss vs epoch curves. 

42
4.5 Final accuracy
For both the CNN and dataset, a pick accuracy is reached and then the accuracy
fluctuates between a range. At the 10th epoch for both the dataset, the accuracy was
found to be 99.7% for dataset_1 and 75.0% for dataset_2.
4.6 Test results
To examine the performance of the CNNs, some random images collected from the
internet and on-site inspection are be put through the CNN models and prediction
algorithm will be used to predict classes.
Here 10 test images are used (Figure 4.8), which covers various lighting, distance,
position, location and camera angle conditions. So, testing performance of the
constructed CNNs using these images will give a good idea about their
performance.
Test images constitutes of 3 shear crack images, 3 spalling of concrete images, 2

shrinkage crack images and 2 flexural crack images. CNN_1 is designed to address
all the four category of cracks in a concrete beam. But CNN_2 can not distinguish
shrinkage cracks from flexural cracks. So, for CNN_2, flexural and shrinkage
cracks will be addressed together as ‘Flexural crack’.
Informations regarding beam and crack did to come along the images as they most
of them were collected randomly from the internet. So, length of beams, location of
cracks from nearest support and depth of beam affected by cracks are assumed
arbitrarily but logically from what can be seen from respected pictures.
Final predictions after running all the test results through CNN_1 and prediction
algorithm for dataset_1 are shown in Table 4.1. 

43
(a) Shear Cracks
(b) Spalling of Concrete.
(c) Shrinkage Cracks.
(d) Flexural Cracks.
Figure 4.9: Test Crack images.  

44
Table
Table4.1:
4.1: Test results for CNN_1.
Location
of Crack Depth of
Length
Test Image Crack No Crack Preliminary from the beam
Sl. of Beam Final Result
Crack Probability Probability Prediction Nearest covering
(m)
Support crack
(m)
Shear Full Shear

1 1 0 Crack 10 0.01
Crack Depth Crack
Shear Not Full Shear

2 1 0 Crack 10 0.01
Crack Depth Crack
Shear Full Shear

3 1 0 Crack 25 0.01
Crack Depth Crack
Spalling
4 of 0 0 Spalling - - - Spalling
concrete
Spalling
5 of 1 0 Crack - - - Crack
concrete
Spalling
6 of 1 0 Crack - - - Crack
concrete
Flexural
Shrinkage Full and
7 1 0 Crack 10 4
Crack Depth Shrinkage
Crack
Shrinkage Full Shrinkage

8 1 0 Crack 10 3.5
Crack Depth Crack
Flexural Not Full Flexural

9 1 0 Crack 10 5
Crack Depth Crack
Flexural Full Flexural

10 1 0 Crack 10 6
Crack Depth Crack
45
Final predictions after running all the test results through CNN_2 and prediction
algorithm for dataset_2 are shown in Table 4.2.
Table
Table4.2:
4.2: Test results for CNN_2.
Flexural
Serial Shear Crack Spalling Final
Test Image Crack Crack
No. Probability Probability Prediction
Probability
1 Shear Crack 1 0 0 Shear Crack
2 Shear Crack 1 0 0 Shear Crack
3 Shear Crack 0 0 1 Spalling
Spalling of
4 0 0 1 Spalling
concrete
Spalling of
5 0 0 1 Spalling
concrete
Spalling of
6 1 0 0 Shear Crack
concrete
Flexural
7 Shrinkage Crack 0 1 0
Crack
Flexural
8 Shrinkage Crack 0 1 0
Crack
9 Flexural Crack 0 0 1 Spalling
10 Flexural Crack 0 0 1 Spalling

46
4.7 Summary
Summary of the test results are shown in Table 4.3.
Table
Table4.3:
4.3: Summary of the test images.
Actual CNN_1 CNN_2

Serial No. Result Result
Crack Prediction Prediction
Shear
1 Shear Crack OK Shear Crack OK
Crack
Shear
2 Shear Crack OK Shear Crack OK
Crack
Shear
3 Shear Crack OK Spalling NOT OK
Crack
Spalling of
4 Spalling OK Spalling OK
concrete
Spalling of
5 Crack NOT OK Spalling OK
concrete
Spalling of
6 Crack NOT OK Shear Crack NOT OK
concrete
Flexural and
Shrinkage Flexural
7 Shrinkage OK OK
Crack Crack
Crack
Shrinkage Shrinkage Flexural
8 OK OK
Crack Crack Crack
Flexural Flexural
9 OK Spalling NOT OK
Crack Crack
Flexural Flexural
10 OK Spalling NOT OK
Crack Crack
47
4.8 Performance study
It can be seen form Table 4.3 that both the CNNs has poor performance identifying
spalling of concrete. Otherwise both the CNNs has shown reliability for using in
real world scenario.
TP + T N
Accuracy for CNN_1 =
8
= = 0.80=80%
10
So, the accuracy of CNN_1 is 80%, which is far below than validation accuracy,
which is 99.7%.
TP + T N
Accuracy for CNN_2 =
6
= = 0.60=60%
10
Accuracy of CNN_@ is 60%, which is far below than validation accuracy, which is
75%.
So, for real world scenarios CNN_1 is more compatible and reliable than CNN_2.  
48
Chapter 5
CONCLUSIONS AND SUGGESTIONS
5.1 Conclusions
The two CNN models developed for this study is able to detect and classify
concrete beam cracks. The goal is successfully achieved.
The research reveals the following:
i. Dataset_1 performs better than dataset_2 in real world scenario with much
better accuracy and reliability.
ii. Increasing or decreasing convolution and pooling layers have little effect
on validation accuracy.
iii. Increasing and decreasing batch number has little effect on validation
accuracy
iv. Standardized images produces better prediction.
v. For classifying cracks, one have to carefully observe crack location and
depth of cracks. Although a slight miscalculation does not affect the
results.
vi. Training on GPU with a higher volume of RAM would have reduced the
training and validation time. Although it would had no effect on training
and validation accuracy.
vii. Using large number of image increase accuracy of the model.
viii.After a certain epoch number, increasing the epoch number does not
increase accuracy.
49
5.2 Suggestions
The following suggestions are made for future study.
i. Studying the relationship between input image size and training time and
accuracy.
ii. Studying the relationship between number of output nodes in hidden layer
and training time and accuracy.
iii. A study to compare test results using different Stochastic gradient descent
algorithms.
iv. A study to compare training time and accuracy using a larger number of
image augmentation.
v. A study to prepare distinguishable ‘Shear crack’ and ‘Flexural crack’

image databases.  
50
REFERENCES
[1] Raju N.K., Basavarajaiah B.S., Ramakrishna N, 1979. “A comparative study

of concrete mix design procedures”, Indian Concrete
Journal, pp. 13-16.
[2] James K. Wight and James G. MacGregor, “Reinforced Concrete (Prentice

Hall, 2009)”.
[3] https://baldeagleconstruction.com/why-concrete-cracks [Last access on 13

Oct. 2018].
[4] https://www.flexomeric.com/Types-of-Cracks-in-Concrete_ep_50.html [Last

access on 13 Oct. 2018].
[5] http://www.concrete.org.uk/fingertips-nuggets.asp?cmd=display&id=186
[Last access on 13 Oct. 2018].
[6] http://www.businessdictionary.com/definition/neural-network.html [Last

access on 13 Oct. 2018].
[7] https://www.analyticsindiamag.com/6-types-of-artificial-neural-networks-
currently-being-used-in-todays-technology [Last access on 13 Oct. 2018].
[8] Krizhevsky, A., Sutskever, I. & Hinton, G. E. (2012), Imagenet classification

with deep convolutional neural networks, Advances in Neural Information
Processing Systems, 1097–105.
[9] Scherer, D., Mu ̈ ller, A. & Behnke, S. (2010), Evaluation of pooling

operations in convolutional architectures for object recognition, in K.
Diamantaras, W. Duch, and L. S. Iliadis (eds.), Artificial Neural Networks–
ICANN 2010, Springer Verlag, Berlin Heidelberg, 92–101.
[10] Nair, V. & Hinton, G. E. (2010), Rectified linear units improve restricted
Boltzmann machines, in Proceedings of the 27th International Conference
on Machine Learning (ICML-10), Haifa, Israel, June 21–24, 807–14.
[11] Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I. & Salakhutdinov, R.
(2014), Dropout: a simple way to prevent neural networks from overfitting,
Journal of Machine Learning Research, 15(1), 1929–58.
51
[12] LeCun, Y. A., Bottou, L., Orr, G. B. & Mu ̈ ller, K.-R. (2012),
Efficient backprop, in G. Montavon, G. B. Orr, and K.-R. Müller (eds.), Neural
Networks: Tricks of the Trade, 2nd edn., Springer, Berlin Heidelberg, pp. 9–48.
[13] Ioffe, S. & Szegedy, C. (2015), Batch normalization: accelerating deep network
training by reducing internal covariate shift, arXiv preprint arXiv:1502.03167.
[14] Bengio, Y., Goodfellow, I. J. & Courville, A. (2016), Deep Learning, An MIT
Press book. Online version is available at: http://www.deeplearningbook.org,
accessed July 2016.
[15] Bottou, L. (2012), Stochastic gradient descent tricks, in G. Montavon, G. B. Orr,

and K.-R. Mü ller (eds.), Neural Net- works: Tricks of the Trade, 2nd edn.,
Springer, Berlin Hei- delberg, pp. 421–36.
[16] Ahmadlou, M. & Adeli, H. (2010), Enhanced probabilistic neural network with
local decision circles: a robust classifier, Integrated Computer-Aided
Engineering, 17(3), 197– 210.
[17] Ciresan, D. C., Meier, U., Masci, J., Maria Gambardella, L. & Schmidhuber, J.
(2011), Flexible, high performance convolutional neural networks for image
classification, in Proceedings of International Joint Conference on Artificial
Intelligence, Barcelona, Spain, pp. 1234–42.
[18] Krizhevsky, A., Sutskever, I. & Hinton, G. E. (2012), Imagenet classification

with deep convolutional neural networks, Advances in Neural Information
Processing Systems, 1097–105.
[19] Simard, P. Y., Steinkraus, D. & Platt, J. C. (2003), Best practices for
convolutional neural networks applied to visual document analysis, in
Proceedings of the 7th International Conference on Document Analysis and
Recognition, Edinburgh, UK, August 3–6, 958–62.
[20] LeCun, Y., Bengio, Y. & Hinton, G. (2015), Deep learning, Nature, 521(7553),
436–44.
[21] Steinkrau, D., Simard, P. Y. & Buck, I. (2005), Using GPUs for machine
learning algorithms, in Proceedings of 8th International Conference on
Document Analysis and Recognition, Seoul, Korea, August 29–September 1,
1115–19.
[22] Soukup, D. & Huber-Mö rk, R. (2014), Convolutional neural networks for steel
surface defect detection from photometric stereo images, in Proceedings of 10th
International Symposium on Visual Computing, Las Vegas, NV, 668–77.
52
[23] Cha, Young-Jin and Choi, Wooram and Buyukozturk, Oral. (2017), Deep
Learning-Based Crack Damage Detection Using Convolutional Neural
Networks, Computer-Aided Civil and Infrastructure Engineering. 32. 361-378.
10.1111/mice.12263.
[24] Feng, Chen and Liu, Ming-Yu and Kao, Chieh-Chi and Lee, Teng-Yok. (2017),
Deep Active Learning for Civil Infrastructure Defect Detection and
Classification. 298-306. 10.1061/9780784480823.036.
[25] M. Salman, S. Mathavan, K. Kamal, M. Rahman, Pavement crack detection

using the gabor filter, in: Proceedings of 16th International IEEE Annual
Conference on Intelligent Transportation Systems, 2013, pp. 2039–2044.
[26] Shivprakash Iyer, Sunil K. Sinha, A robust approach for automatic detection and
segmentation of cracks in underground pipeline images, Image Vis. Comput. 23
(10) (2005) 931–933.
[27] Hoang-Nam Nguyen, Tai-Yan Kam, Pi-Ying Cheng, An automatic approach for
accurate edge detection of concrete crack utilizing 2D geometric features of
crack, J. Signal Process. Syst. 77 (3) (2014) 221–240.
[28] Yuan-Sen Yang, Chung-Ming Yang, Chang-Wei Huang, Thin crack observation
in a reinforced concrete bridge pier test using image processing and analysis,
Adv. Eng. Softw. 83 (2015) 99– 108.
[29] Pingrang Wang, Hongwei Huang, Comparison analysis on present image-based

crack detection methods in concrete structures, in: Proceedings of 2010 3rd
International Congress on Image and Signal Processing (CISP2010), vol. 5,
2010, pp. 2530–2533.
[30] https://data.mendeley.com/datasets/5y9wdsg2zt/1 [Last access 14 Oct. 2018].
[31] Lei Zhang, Fan Yang, Yimin Daniel Zhang, and Ying Julie Zhu, Road crack
detection using deep convolutional neural network, in: Proceedings of 2016
IEEE International Conference on Image Processing(ICIP2016), DOI: 10.1109/
ICIP.2016.7533052.
[32] https://keras.io/preprocessing/image [Last access 14 Oct. 2018].
[33] https://en.wikipedia.org/wiki/Keras [Last access 14 Oct. 2018].
[34] Debora B.(Editor) 2016, “Design of Concrete Structures”, Fourteen Edition, The
McGraw-Hill Companies, Inc.
[35] https://www.quora.com/What-is-an-epoch-in-deep-learning [Last access

14 Oct. 2018]. 
53
APPENDIX A (PYTHON CODES)
CNN_1 python codes
1. # Convolutional Neural Network

2.
3. # Installing Theano
4. # pip install theano
5.
6. # Installing Tensorflow
7. # pip install tensorflow
8.
9. # Installing Keras
10. # pip install keras
11.
12. # Part 1 - Building the CNN
13.
14. # Importing the Keras libraries and packages
15. from keras.models import Sequential
16. from keras.layers import Conv2D
17. from keras.layers import MaxPooling2D
18. from keras.layers import Flatten
19. from keras.layers import Dense
20. import matplotlib.pyplot as plt
21.
22.
23. # Initialising the CNN
24. classifier = Sequential()
25.
26. # Step 1 - Convolution
27. classifier.add(Conv2D(32, (3, 3), input_shape = (32, 3
2, 3), activation = 'relu'))
28.
29. # Step 2 - Pooling
30. classifier.add(MaxPooling2D(pool_size = (2, 2)))
31.
32. # Adding a second convolutional layer
33. classifier.add(Conv2D(64, (3, 3), activation = 'relu')
)
35.
36. # Step 3 - Flattening
37. classifier.add(Flatten())
54
38.
39. # Step 4 - Full connection
40. classifier.add(Dense(units = 128, activation = 'relu')
)
41. classifier.add(Dense(units = 3, activation = 'softmax'
))
42.
43. # Compiling the CNN
44. classifier.compile(optimizer = 'adam', loss = 'categor
ical_crossentropy', metrics = ['accuracy'])
45.
46. # Part 2 - Fitting the CNN to the images
47. from keras.preprocessing.image import ImageDataGenerat
or
48.
49. train_datagen = ImageDataGenerator(rescale = 1./255,
50. shear_range = 0.2,
51. zoom_range = 0.2,

52. horizontal_flip = T
rue,
53. vertical_flip = Tru
e,
54.
55. rotation_range = 36
0
56. )
57.
58.
59. val_datagen = ImageDataGenerator(rescale = 1./255,
60. shear_range = 0.2,
61. zoom_range = 0.2,
62. horizontal_flip = Tr
ue,
63. vertical_flip = True
,
64.
65. rotation_range = 360
66. )
67.
68.
69.
70.
71. training_set = train_datagen.flow_from_directory('/
Users/saminyeaser/Desktop/
Concrete Crack Images for Classification/
training_set',
72. targe
t_size = (32, 32),
55
73. batch
_size = 32,
74. class
_mode = 'categorical')
75.
76. val_set = val_datagen.flow_from_directory('/Users/
saminyeaser/Desktop/
Concrete Crack Images for Classification/val_set',
77. target_siz
e = (32, 32),
78. batch_size
= 32,
79. class_mode
= 'categorical')
80.
81. history = classifier.fit_generator(training_set,
82. steps_per_epoch = 32041,
83. epochs = 10,
84. validation_data = val_set,
85. validation_steps = 8010)
86.
87.
88.
89. # list all data in history
90. print(history.history.keys())
91.
92. # summarize history for accuracy
93. plt.plot(history.history['acc'])
94. plt.plot(history.history['val_acc'])
95. plt.title('model accuracy')
96. plt.ylabel('accuracy')
97. plt.xlabel('epoch')
98. plt.legend(['train', 'test'], loc='upper left')
99. plt.show()
100.
101.# summarize history for loss
102.plt.plot(history.history['loss'])
103.plt.plot(history.history['val_loss'])
104.plt.title('model loss')
105.plt.ylabel('loss')
106.plt.xlabel('epoch')
107.plt.legend(['train', 'test'], loc='upper left')
108.plt.show()
56
CNN_2 python code
1. # Convolutional Neural Network

2.
3. # Installing Theano
4. # pip install --upgrade --no-deps git+git://
github.com/Theano/Theano.git
5.
6. # Installing Tensorflow
7. # pip install tensorflow
8.
9. # Installing Keras
10. # pip install --upgrade keras
11.
12. # Part 1 - Building the CNN
13.
14. # Importing the Keras libraries and packages
15.
16. from keras.models import Sequential
17. from keras.layers import Conv2D
18. from keras.layers import MaxPooling2D
19. from keras.layers import Flatten
20. from keras.layers import Dense
21. import matplotlib.pyplot as plt
22.
23.
24.
25. # Initialising the CNN
26. classifier = Sequential()
27.
28. # Step 1 - Convolution
29. classifier.add(Conv2D(32, (3, 3), input_shape = (64, 6
4, 3), activation = 'relu'))
30.
31. # Step 2 - Pooling
33.
34. # Adding a second convolutional layer
35. classifier.add(Conv2D(64, (3, 3), activation = 'relu')
)
37.
38.
39.
40.
57
41.
42.
43.
44.
45.
46. # Step 3 - Flattening
47. classifier.add(Flatten())
48.
49. # Step 4 - Full connection
50. classifier.add(Dense(units = 128, activation = 'relu')
)
51. classifier.add(Dense(units = 3, activation = 'softmax'
))
52.
53. # Compiling the CNN
54. classifier.compile(optimizer = 'adam', loss = 'categor
ical_crossentropy', metrics = ['accuracy'])
55.
56. # Part 2 - Fitting the CNN to the images
57. from keras.preprocessing.image import ImageDataGenerat
or
58.
59. train_datagen = ImageDataGenerator( rescale=1./255,
60. shear_range=0.2,
61. zoom_range=0.2,
62. horizontal_flip=True,
64.
65. )
66.
67.
68.
69. test_datagen = ImageDataGenerator(rescale = 1./255,
70. shear_range=0.2,
71. zoom_range=0.2,
72. horizontal_flip=True
,
74.
75. )
76.
77.
78.
79. training_set = train_datagen.flow_from_directory('/
Users/saminyeaser/Desktop/dataset/training_set',
80. targe
t_size = (64, 64),
81. batch
_size = 3,
58
82. class
_mode = 'categorical')
83.
84. val_set = test_datagen.flow_from_directory('/Users/
saminyeaser/Desktop/dataset/val_set',
85. target_siz
e = (64, 64),
86. batch_size
= 3,
87. class_mode
= 'categorical')
88.
89. history = classifier.fit_generator(training_set,
90. steps_per_epoch = 99,
91. epochs = 10,
92. validation_data = val_set,
93. validation_steps = 24)
94.
95. # list all data in history
96. print(history.history.keys())
97.
98. ['acc', 'loss', 'val_acc', 'val_loss']
99.
100.# summarize history for accuracy
101.plt.plot(history.history['acc'])
102.plt.plot(history.history['val_acc'])
103.plt.title('model accuracy')
104.plt.ylabel('accuracy')
107.plt.show()
108.
109.# summarize history for loss
110.plt.plot(history.history['loss'])
111.plt.plot(history.history['val_loss'])
112.plt.title('model loss')
113.plt.ylabel('loss')
116.plt.show()
117.
118.
119.classifier.save('dataset_02.h5')
120.
121.
122.from keras.models import load_model
123.new_model= load_model('dataset_02.h5')
124.new_model.summary()  
59
APPENDIX B (PREDICTION CODE)
For Dataset_1
1. # Part 3 - Making new predictions

2. import numpy as np
3. from keras.preprocessing import image
4. test_image = image.load_img('/Users/saminyeaser/
Desktop/dataset/single_prediction/test.
10.jpg', target_size = (32, 32))
5. test_image = image.img_to_array(test_image)
6. test_image = np.expand_dims(test_image, axis = 0)
7. result = classifier.predict(test_image)
8. training_set.class_indices
9. if result[0][0] >= .4:
10. prediction = 'Spalling'
11. elif result[0][1]>=.4:
12. prediction = 'crack'
13. lb= float(input("Length of the beam:"))
14. lc= float(input("Location of crack from nearest su
pport:"))
15. d = int(input("Full Depth: 1 and Not Full Depth: 0
"))
16. ratio= lc/lb
17. if ratio<0.3:
18. ultimate="Shear Crack"
19. print(ultimate)
20. elif (0.4<= ratio <=0.5 and d==0):
21. ultimate="Flexural Crack"
22. print(ultimate)
23. elif (0.4<= ratio <=0.5 and d==1):
24. ultimate="Both Flexural and Shrinkage Crack"
25. print(ultimate)
26.
27. elif (0.3<= ratio <0.4 and d==1):
28. ultimate="Shrinkage Crack"
29. print(ultimate)
30.
31. else:
32. ultimate="Flexural Crack"
33. print(ultimate)
34.
35. elif result[0][2]>=.4:
36. prediction = 'no crack'  
60
For Dataset_2
1. #Prediction
2. import numpy as np
3. from keras.preprocessing import image
4. test_image = image.load_img('/Users/saminyeaser/
Desktop/dataset/single_prediction/yes.
1.jpg',target_size = (32, 32))
5. test_image = image.img_to_array(test_image)
6. test_image = np.expand_dims(test_image, axis=0)
7. classifier.predict(test_image)
8. result = classifier.predict(test_image)
9. training_set.class_indices
10. if result [0][0]>=.4:
11. prediction='Flexural'
12. elif result [0][1]>=.4:
13. prediction='Shear'
14. elif result [0][2]>=.4:
15. prediction='Spalling'  
61
APPENDIX C (SUMMARY OF THE CNN)
CNN_1
1. ______________________________________________________
___________
2. Layer (type) Output Shape
Param #
3. ======================================================
===========
4. conv2d_3 (Conv2D) (None, 30, 30, 32)
896
5. ______________________________________________________
___________
6. max_pooling2d_3 (MaxPooling2 (None, 15, 15, 32)
0
7. ______________________________________________________
___________
8. conv2d_4 (Conv2D) (None, 13, 13, 64)
18496
9. ______________________________________________________
___________
0
11. ______________________________________________________
___________
12. flatten_2 (Flatten) (None, 2304)
0
13. ______________________________________________________
___________
14. dense_3 (Dense) (None, 128)
295040
15. ______________________________________________________
___________
387
17. ======================================================
===========
18. Total params: 314,819
19. Trainable params: 314,819
20. Non-trainable params: 0
21. ______________________________________________________
___
62
CNN_2
1. ______________________________________________________
___________
2. Layer (type) Output Shape
Param #
3. ======================================================
===========
4. conv2d_5 (Conv2D) (None, 62, 62, 32)
896
5. ______________________________________________________
___________
0
7. ______________________________________________________
___________
8. conv2d_6 (Conv2D) (None, 29, 29, 64)
18496
9. ______________________________________________________
___________
0
11. ______________________________________________________
___________
12. flatten_3 (Flatten) (None, 12544)
0
13. ______________________________________________________
___________
1605760
15. ______________________________________________________
___________
387
17. ======================================================
===========
18. Total params: 1,625,539
19. Trainable params: 1,625,539
20. Non-trainable params: 0
21. ______________________________________________________
___
63
APPENDIX D (TRAINING AND VALIDATION)
For Dataset 1
1. Found 32041 images belonging to 3 classes.

3. Epoch 1/10
4. 32041/32041 [==============================] -
3332s 104ms/step - loss: 0.0343 - acc: 0.9892 -
val_loss: 0.0158 - val_acc: 0.9953
5. Epoch 2/10
6. 32041/32041 [==============================] -
3228s 101ms/step - loss: 0.0165 - acc: 0.9953 -
val_loss: 0.0206 - val_acc: 0.9947
7. Epoch 3/10
8. 32041/32041 [==============================] -
3569s 111ms/step - loss: 0.0142 - acc: 0.9960 -
val_loss: 0.0221 - val_acc: 0.9949
9. Epoch 4/10
10. 32041/32041 [==============================] -
3362s 105ms/step - loss: 0.0136 - acc: 0.9961 -
val_loss: 0.0132 - val_acc: 0.9963
11. Epoch 5/10
12. 32041/32041 [==============================] -
3303s 103ms/step - loss: 0.0130 - acc: 0.9964 -
val_loss: 0.0160 - val_acc: 0.9957
13. Epoch 6/10
14. 32041/32041 [==============================] -
3301s 103ms/step - loss: 0.0123 - acc: 0.9966 -
val_loss: 0.0140 - val_acc: 0.9965
15. Epoch 7/10
16. 32041/32041 [==============================] -
3482s 109ms/step - loss: 0.0127 - acc: 0.9965 -
val_loss: 0.0156 - val_acc: 0.9959
17. Epoch 8/10
18. 32041/32041 [==============================] -
3363s 105ms/step - loss: 0.0124 - acc: 0.9967 -
val_loss: 0.0134 - val_acc: 0.9965
19. Epoch 9/10
20. 32041/32041 [==============================] -
3379s 105ms/step - loss: 0.0123 - acc: 0.9967 -
val_loss: 0.0194 - val_acc: 0.9951
21. Epoch 10/10
22. 32041/32041 [==============================] -
4547s 142ms/step - loss: 0.0124 - acc: 0.9967 -
val_loss: 0.0143 - val_acc: 0.9964
64
For Dataset 2

3. Epoch 1/10
4. 99/99 [==============================] - 7s 67ms/
step - loss: 0.9766 - acc: 0.5118 - val_loss: 0.8089 -
val_acc: 0.7222
5. Epoch 2/10
6. 99/99 [==============================] - 6s 64ms/
val_acc: 0.7917
7. Epoch 3/10
8. 99/99 [==============================] - 6s 62ms/
val_acc: 0.8056
9. Epoch 4/10
10. 99/99 [==============================] - 6s 62ms/
val_acc: 0.7778
11. Epoch 5/10
12. 99/99 [==============================] - 6s 65ms/
val_acc: 0.8194
13. Epoch 6/10
14. 99/99 [==============================] - 6s 64ms/
val_acc: 0.8056
15. Epoch 7/10
16. 99/99 [==============================] - 6s 61ms/
val_acc: 0.8750
17. Epoch 8/10
18. 99/99 [==============================] - 7s 67ms/
val_acc: 0.7778
19. Epoch 9/10
20. 99/99 [==============================] - 7s 75ms/
val_acc: 0.7639
21. Epoch 10/10
22. 99/99 [==============================] - 6s 63ms/
val_acc: 0.7500 
65
APPENDIX E (TIME VS EPOCH TABLES)
For Dataset 1
Time Required per Time Required per Cumulative Time

Epoch
Epoch (s) Epoch (min) (min)
1 3332 5.55E+01 5.55E+01
2 3228 5.38E+01 1.09E+02
3 3569 5.95E+01 1.69E+02
4 3362 5.60E+01 2.25E+02
5 3303 5.51E+01 2.80E+02
6 3301 5.50E+01 3.35E+02
7 3482 5.80E+01 3.93E+02
8 3363 5.61E+01 4.49E+02
9 3379 5.63E+01 5.05E+02
10 4547 7.58E+01 5.81E+02
 
For Dataset 2
Time Required Per Time Required per Epoch Cumulative Time

Epoch
Epoch (s) (min) (min)
1 7 1.17E-01 1.17E-01
2 6 1.00E-01 2.17E-01
3 6 1.00E-01 3.17E-01
4 6 1.00E-01 4.17E-01
5 6 1.00E-01 5.17E-01
6 6 1.00E-01 6.17E-01
7 6 1.00E-01 7.17E-01
8 6 1.00E-01 8.17E-01
9 7 1.17E-01 9.33E-01
10 6 1.00E-01 1.03E+00

Deep Learning Based Concrete Beam Crack Detection and Classification

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Deep Learning Based Concrete Beam Crack Detection and Classification

Caricato da

Copyright:

Formati disponibili

1

DEEP LEARNING BASED CONCRETE BEAM CRACK

(STUDENT ID: 201304160)

In partial fulfillment of the requirement for the degree of

DEPARTMENT OF CIVIL ENGINEERING

BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA

guidance in performing this research study, despite all odds.

First of all, I express my profound gratitude to Prof. Dr. Ahsanul Kabir, my

I convey my heartiest regards to Dr. Md. Hadiuzzaman, Associate Professor,

helped me immensely. I am extremely grateful to Nazmul Haque for sharing his

unconditional support throughout my life and studies which worked as a stimulus

for me in making a successful and productive completion of the thesis.

Cracks in buildings are of common occurrence which develops whenever stress in

to reduction in load bearing capacity of a structural component.

A number of image processing techniques (IPTs) have been implemented for

architecture of convolutional neural networks (CNNs) are useful. As CNNs are

capable of learning image features automatically without calculating defect feature.

The performance and adaptability of the proposed approaches are tested on 10

List of Figures viii

2.6.1 Feedforward neural network – artificial neuron

4.3 Accuracy vs epoch

Table No. Title of the Table Page No

Table 4.1: Test results for CNN_1. 44

Table 4.1: Test results for CNN_2. 46

Table 4.1: Summary of the test images. 48

Figure No. Title of the Figure Page No

Figure 2.1: Shrinkage crack.

Figure 4.1: Layer diagram of CNN_1.

SHM Structural Health Monitoring

CNN Convolutional Neural Network

RNN Recurrent Neural Network

IPT Image Processing Technique

GUI Graphical User Interface

ReNet Residual Network

RGB Red, Green and Blue

relu Rectifier Activation Function

SGD Stochastic gradient descent

cracks is the first step of retrofitting these structures.

design life putting peoples’ lives at risk.

Failure in beam results in localized failure of a structure. So identifying and

thus preventing future catastrophes.

smartphones with camera it is now a possibility to help mass people to check

structural health of their own occupying building by technologies developed using

image processing techniques.

Network (CNN) to detect and classify cracks in concrete beams.

Specific objectives of the study includes:

and other CNN_2. Both of them are categorical classifier.

1.3 Scope of study

constructed to work as a concrete beam crack image classifier. Two different

Concrete is inert mass which grows from a cementing medium. Concrete is a

the concrete starts and it starts to gain its strength.

2.2 Concrete cracks

design process. Reinforced concrete structures normally crack when carrying

2.3 Reasons for occurring cracks

2.3.1 Plastic shrinkage

cosmetic flaw rather than a structural one.

2.3.2 Expansion and contraction

2.3.3 Heaving and settling

crush the concrete.

2.3.5 Improper drying

2.3.6 Natural disaster

2.4 Types of cracks in concrete

2.4.1 Shrinkage crack

Figure 2.1: Shrinkage crack.  

should generally be narrower. 