You are on page 1of 14

INSTITUTE OF TECHNOLOGY

DEPARTMENT OF ELECTRICAL AND COMPUTER


ENGINEERING

DIGITAL AND SIGNAL PROCESSING PROJECT


DOCUMENTATION
DEFECT DETECTION OF CERAMIC TILES

PREPARED BY ID
 Ermias Amare 4402/08
 Surafel Kebede 4901/08
 Melkamu Feye 5705/08
 Mekonen Ayanaw 4688/08
Submitted to: MR.Belayneh

Submission date:22/09/2010 E.C


Acknowledgment
Firstly, we would like to thank our lord for everything. Secondly, we would like to thank our
respected teacher Mr.Belayneh for giving us this project because this project initiates us for
the new invention (development) and connected us with the real word application. And also
our teacher was with us while working this project; this makes us to change our mind for a
better project. Lastly, we would like to thank our friends for sharing us their ideas.
Abstract
In ceramic tile factory, the production process has now performed automatically by industrial
automation system, except the examination process for ceramic quality classification which is
still conducted manually. Classification process is performed using the human visual
assessment to find and to classify defect, where human judgment is dependent entirely on
experience and knowledge. Therefore, this process requires an automatic mechanism which
can provide an assessment of the ceramic quality accurately and consistently.
Table of Contents
Chapter 1 ................................................................................................................................................. 1
Introduction ......................................................................................................................................... 1
Objectives........................................................................................................................................ 1
Motivation……………………………………………………………………………………………………………………………….1

Project scope………………………………………………………………………………………………………………………….1

Chapter 2................................................................................................................................................. 2
Related works...................................................................................................................................... 2
Chapter 3………………………………………………………………………………………………………………………………………3

Project Methodology .......................................................................................................................... 3


Chapter 4................................................................................................................................................. 3
System Design ..................................................................................................................................... 3
Chapter 5................................................................................................................................................. 5
Result Analysis and Discussion ......................................................................................................... 2
Chapter 6................................................................................................................................................. 7
Conclusion ........................................................................................................................................... 7
Sample code..................................................................................................................................... 7
Reference...........................................................................................................8
List of figures

Figure 1:System design


Figure 2:User interface
Figure 3:picking image
Figure 4:Testing image
Figure 5:defected image

Figure 6: Undefected image


Chapter One
Introduction
Computer technology has been growing and expanding its use in helping to resolve the
problems on various aspects of human life.
This project introduces about defect detection of ceramic tiles in the factory.Currently,the
production process at ceramic tiles factory have been done automatically by machine
through industrial automation system, except the examination process for ceramic equality
classification which is still performed manually by the intervention of human operator.
Sincethese human operations have its own drawbacks such as time efficiency, accuracy and
endurance, which is very expensive, since it requires workers to work in shifts and the other
drawback is subjectivity since it is fully influenced by the experience and knowledge of the
workers. These disadvantages can lead to errors in the phase of ceramic quality identification,
so we proposed system defect detection in ceramic tiles to improve this detect damaging of
tiles. This improvement can effectively replace human operation (manual ceramic tiles) with
better efficiency.

Objectives
This project has its own objectives:
 The main objective is to detect the defects such as cracks in ceramic tiles.
 To save time ,endurance and do efficient works.

Motivation

Because of our teacher, we were motivated to do this project since he gave us the project to
do as assignment. After that, when we go to started the project, we become mostly
interested to improve the problem of defect detection in ceramic tiles for our country.

Project scope
The scope of the project will be wide as much as the ceramic factories exist. Since
construction is a basic thing in our world use of ceramics is also very significant in
construction sites. So the manufacture of defect less ceramics is recommended for good
construction. Means ceramics has a lot of uses in many areas of construction so its scope is
very wide.

1
Beneficiary body from the project
The beneficiary bodies from the project are ceramic tile fabrics or companies I which they
can identify the defects on the ceramic tiles easily and in more systematic way.

Indirectly the users who need ceramic tiles in their work will be beneficiary since all the
defected materials or ceramics are identified and will not be out for sale so the construction
will be better.

Chapter 2

Literature Review/Related works

There are a lot of related works like wall crack detection which is used for identification of
cracked walls, and also bone fracture detection used in medical purposes to find out if a bone
is broken or not.

There are also a lot of ceramic crack, defect detection projects in another methods such as by
morphological methods.

Chapter 3

Project Methodology

We used a matlab programing language to perform defect detection process of ceramic tiles.
As we are asked to detect the defection of ceramic tiles, first we have used defect detection
by black pixel counting method. We insertedreference image and test image. This is because
we detect the test image in reference to our reference image. Then we convert both images
means the reference and the test images to grey scale. After that we calculated the black
pixels for both images. As we have mentioned above, we had taken two images which both of
them had white and black pixels. Thus we represented them by a1 and a2 for reference and
then compare the black pixel count of the images.
 If a1=a2, then we will have ‘No defect’ since the black pixel value is equal means
2
 the two ceramics are perfectly identical.
 If the black pixel count value is different there must be some defect in the test
image with reference to the reference image, so defect is found
Finally the resulting message will pop up in a matlab command window.

Chapter 4

SYSTEM DESIGN
system specifications
the specifications of this system is it must have a reference image in order to operate. since
companies have their own quality measurement value our program must have a reference
image as an input and it evaluates the all the test images according to the reference that is
given.

The system is designed using matlab software using black pixel counting comparison
method.

We calculated the black pixel of the reference image and the test image and compare if the
counts are equal. If they are equal the test image has no defect in relation to the reference
image if not the test image has defects on its surface.

We constructed a process flow chart to illustrate how our system is design

3
• Load reference
and test images

• convert both
images to gray
scale
• Calculate
black pixel of
both images

Check if the black


pixels are equal

Defect No
found defect
Figure 1 flow chart of system design found

4
Chapter 5

Result Analysis and Discussion

Figure 2 user interface

Figure 2 shows the user interface in which the user can insert the reference image by clicking
“LOAD REFERENCE IMAGE” button and test image by clicking “LOAD TEST IMAGE”
button.

Fig 3:picking of an image

5
Fig 4:Testing image

Fig 5 : defected image

Finally the results will come up with the calculated black pixel count and with message in command
window

6
If the images are perfectly the same

Fig 6:undefected image

Chapter 6

Conclusion

Generally our project is used for detecting ceramic tiles whether they are defected or notIt
uses black pixel comparison method to detect the defections the exist in the ceramic. The
system will input two images of the ceramic as one the test and one as a reference image.
Then turn the images to grey scale and calculate the black pixels and decide whether the test
image has defect or not.

Its strength is since the calculation is very sensitive even a little defect will not be passed. In
order to not have a defect the tiles must be perfectly the same. This also can be regarded as a
weak side of the project since nothing is perfectly the same.

Sample code
imO=imread('refer.jpg');
bO=rgb2gray(imO);
bvO=im2bw(bO);
white0=sum(imO(:));
black0=numel(imO)-white0;
imT=imread('test.jpg');
bT=rgb2gray(imT);
bvT=im2bw(bT);
whiteT=sum(imT(:));
7
blackT=numel(imT)-whiteT;
subplot(2,2,1)
imshow(imO);
title('reference image');

subplot(2,2,2)
imshow(bvO);
title(['black pixel of reference image is =',num2str(black0)]);

subplot(2,2,3)
imshow(imT);
title('test image');

subplot(2,2,4)
imshow(bvO);
title(['black pixel of test image is =',num2str(blackT)]);
a1=blackT;
a2=black0;
if a1>a2
disp('defect is found');
else
disp('defect is not found')'
end

References

 Youtube
 Google
 Digital signal processing books

You might also like