Scribd
Upload
133 views

An Algorithm For Concrete Crack Extraction and Identification

This document presents an algorithm for concrete crack extraction and identification based on machine vision. It discusses collecting images to build a library of different crack conditions. It then describes preprocessing images using grayscaling, smoothing, and morphological operations to identify cracks. Finally, it discusses developing crack recognition software using machine vision that classifies cracks based on a support vector machine model.

Uploaded by

Sandhya
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
133 views

An Algorithm For Concrete Crack Extraction and Identification

This document presents an algorithm for concrete crack extraction and identification based on machine vision. It discusses collecting images to build a library of different crack conditions. It then describes preprocessing images using grayscaling, smoothing, and morphological operations to identify cracks. Finally, it discusses developing crack recognition software using machine vision that classifies cracks based on a support vector machine model.

Uploaded by

Sandhya
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 26

AN ALGORITHM FOR CONCRETE CRACK

EXTRACTION AND IDENTIFICATION


BASED ON MACHINE VISION
GUIDED BY:- Dr.D.REGAN

presented by
M.sandhya yamini
P.lakshmi devi
K.venkata rani
CONTENTS
 INTRODUCTION
 ABSTRACT
 PROBLEM STATEMENT
 EXISTING SYSTEM
 PROPOSED WORK
 SOFTWARE TOOLS
 PROPOSED SYSTEM
 BLOCK DIAGRAM
 PROBLEMS FACED
 MEASURES TAKEN
 REFERENCES
Introduction

Firstly, we collect a large number of road surface images of 3


different conditions including transverse crack, longitudinal
crack and turtle crack separately to construct road surface
conditions library.

Secondly, deal the road damaged image with gray, gray


transform and image smoothing. Then, use mathematical
morphology method to deal with crack image and projection to
identify crack category.

Finally, develop the pavement crack recognition software


based Machine Vision
ABSTRACT
 The proposed algorithm will provide a solutions for the
difficulty of identification and existing in crack processing.

 The first solution entails that it will provide difference


between crack and back ground by using the OTSU
algorithm.

 The second solution is to improve the crack image by


applying crack characterstics using the support vector
machine (svm) for extracting and identification of cracks.
MOTIVATION TOWARDS THE WORK

 Now a days increase of more number of vehicles and


speed driving system accidents are increasing
because of unfavourable road conditions using of low
cost materials used in the construction of roads.
 In order to prevent the accidents this project is very
useful to identified the cracks by using this machine
vision technology.
LITEATURE SURVEY

 We have searched in internet about how we can


predict the concrete cracks by machine vision
technology.
 This article conducts in depth research on concrete
cracks features as input, enabling better identification
and classification of crack images.
EXISTING SYSTEM

Pavement crack damage is one of the most common diseases in the


road destruction phenomena[1], the way to detect road conditions
mainly relies on people and instruments at present.

The Earth technology companies in the United States developed a


pavement condition evaluation system (PCES) that uses threshold
segmentation to extract damaged information of the road surface.

Japan's consortium developed a Komatsu system which


implemented the data of various pavement disease detection, such as
rut, cracks and section.
Disadvantages Of
EXISTING SYSTEM
1.) Not accurate.

2.)Highly complex.

3.) Efficiency is very less.

4.)Poor Performance under noisy conditions.

5.)Requires huge hardware.

6.) High processing time.

7.)Crack or pavement detection is not reliable.

8.)Consumes huge power .

9.) High Operational and maintenance cost.


Proposed METHOD:

Concrete crack images are often characterized as having low color


saturation, low contrast, and a complex background composition.

Therefore, ideal results cannot be easily obtained with traditional


algorithms, such as threshold segmentation.

The OTSU algorithm, i.e., the maximum interclass variance method,


can separate the image into two classes (background and object)
according to the grayscale feature of the image
SOFTWARE REQUIRMENTS

 Matlab R2013a
 Image Processing Toolbox
PROPOSED SYSTEM
 The high frequency domain of the image focuses on the details,
false contour and noise, but the most effective image
information is mainly concentrated in the low frequency
domain.

Global threshold method is a method that chooses a threshold


to divide the whole image into two areas, and the choice of
segmentation threshold is based on the histogram of the image.

These two areas are color-coded, and the typical color is black
and white, color images can also be divided into red, green or
any other color that has quite different hue.
BLOCK DIAGRAM
Edge
Crack PPre Binarization Morphological
detection
imagerc processing rc analysisrc

Extract crack
Classification Morphological
Resultsrc boundaries and
(SVM)c filling
features

Databasec

Fig(1):- Schematic Block Overview of the proposed system.


Proposed System
The schematic block overview of the proposed crack recognition
and identification system is shown in fig(1).

The proposed system first creates a data base with all features of
different crack types and patterns .

 The created database is deployed as a reference for classifying the


concrete crack patterns.

Now after deploying the system, during operation the system first
considers the crack image and performs preprocessing operation on
it to enhance its visual clarity and quality.

Next an Edge Detection operation is performed to extract all edge


patterns of the test crack image.
Proposed system
After extracting the edge and structural features from the test
crack image, all these edge patterns are binarized first to bring them
into two levels.

Next to isolate the unwanted road portions and to highlight the


required crack patterns we perform the morphological erosion and
dilation operations.

After highlighting the crack patterns with morphological filling


operations, we draw rectangular bounding boxes around the crack
and extract them to compare and classify with data base features
through SVM classifier.
Proposed System

Based on the classification results, the type of the crack, size


of the crack and its impact will be estimated and predicted.

All these crack recognition and classification results will be


displayed on the screen.
Advantages of the Proposed System
1.) Highly accurate.

2.)Computational complexity is less.

3.) Efficiency is very high.

4.) Better Performance even under noisy conditions.

5.)Consumes less hardware.

6.)Speed of processing is very high.

7.)Crack or pavement detection is robust

8.)Consumes less power .

9.)Low Operational and maintenance cost.


Problems Faced

During implementation, segmentation and extraction of


crack patterns among all structural features became a bit
difficult.

Morphological filling and marking of crack boundaries


became a bit tedious task.
Measures Taken

 Segmentation , extraction ,recognition and identification


of crack patterns from all the remaining structures are
performed efficiently with the help of canny edge detector
and morphological operations.

An extreme care was taken to recognize crack boundaries


and to fill them.
INPUT
OUTPUT
OUTPUT
CONCLUSION
Based on the video image, we develop road crack recognition
software with Matlab.

The experiment result shows: the software can identify cracks


accurately, and classify cracks detailedly according to its
characteristics.
REFERENCES
 J.W.Long, X.J.Shen, H.P.Chen. “ Interactive Document
Images Thresholding Segmentation Algorithm Based on
Image Regions,”Journal of Computer Research and
Development.
 A.K.Miyamotoetc.“Nonlinear dynamic analysis of reinforced
concrete slabs under impulsive loads,”ACI Structural Journal.
 J. K. Oh,J. Giho, S. Oh, et a1. “Bridge inspection robot
system with machine vision,” Automation in Construction.
 Y. Han, X. H. Zhang, L. Cheng, et al. “Image analaysis
method for road disease morphology characteristic,”
Journal of Wuhan Institute of Technology.
 R. Wang, T. Y. Qi. “Study on crack characteristics based on
machine vision detection,” China Civil Engineering Journal.
REFERENCES
 X. H. Tong, J. Guo, Y. Ling, et al. “A new image-based method for
concrete bridge bottom crack detection,” in Image Analysis and
Signal Processing.
 S. Peleg, A. Zomet, A. Levin, et al. “Seamless image stitching by
minimizing false edges,” IEEE Transactions on Image Processing.
 N. A. Otsu. “Threshold selection method from gray 2 level
Histogram,” IEEE trans.
 S. German, Z. H. Zhu, I. Brilakis. “Visual retrieval of concrete crack
properties for automated post-earthquake structural safety
evaluation,” Automation in Construction.
 H. L.Jeong, M. L.Jong, J. K.Hyung,et al. “Machine Vision System for
Automatic Inspection of Bridges,”in 2008 Congress on Image and
Signal Processing, SanYa, HaiNan,China.
REFERENCES
 T. Yamaguchi, S. Hashimoto. “Fast crack detection method for large-size
concrete surface images using percolation-based image processing,”
Machine Vision and Applications.
 R. Wang, T. Y. Qi, B. Lei, et al. “Characteristics extraction of cracks of
tunnel lining,” Chinese Journal of Rock Mechanics and Engineering.
 J. F. Canny. “A computational approach to Edge Detection,”IEEE
Transactions on PAMI.
 S. S. Keerthi, C. J. Lin. “Asymptotic behaviors of support vector
machines with Gaussian kernel,” Neural Computation.
 A. Albert, A. Nii “Evaluating pavement cracks with bidimensional
empirical mode decomposition,” Journal of Advances in Signal Processing.
 Y. Hawwar, A. Reza. “Spatially Adapative Multiplicative Noise Image
Denoising Technique,” in IEEE Transaction on Image Processing.
Than
k you

You might also like