You are on page 1of 25

FABRIC DEFECT DETECTION USING

DIGITAL IMAGE PROCESSING


Under the supervision of
Ms. Pragya Gupta
Presented By :
Shubham Bhargava
Shirin Dewan
Abhinav Arora

131027
131075
131085

CONTENTS

Introduction

Objective

Practical Implementation

Literature Review

RGB image

HSV image

Noise removal

Masking

Flow diagram

Code explanation

Results

Conclusion

References
2

INTRODUCTION
Defects in fabrics are generated by the manufacturing machines which are not 100% efficient.
A fabric is a flat structure consisting of fibrous materials, either natural or man- made. Fabric defect could be simply
defined as an abnormality in or on the fabric construction.
There are different types of defects

Holes

Stains

Colour spills

Weaving defects

CONTD.
Presently, in this project we will be working on two defects. They are as follows
HOLES

STAINS

OBJECTIVE
The objective of our project is to detect the presence of various types of defects in plain fabrics which occur during
the manufacturing process. Different techniques in spatial and frequency domains are used to implement an
effective system for detecting defects such as holes and stains.

PRACTICAL IMPLEMENTATION
In the manufacturing process, the image of the fabric is taken by a high quality camera. The fabric sheet is exposed
to proper light such that the light is evenly spread all over the surface. The acquired image is then downloaded in
the system for fault detection. The image passes through a series of blocks which include image enhancement,
image segmentation, etc. to produce a final output which highlights the defect in the rgb image. A control signal is
then sent to the machinery to stop and asks users what action to take. In the current model weve implemented the
detection of two defects namely holes and stains using MATLAB.

LITERATURE REVIEW

WHAT IS AN RGB IMAGE?

WHAT IS AN HSV IMAGE?

NOISE REMOVAL

10

MASKING
Write about masking in image segmentation

11

FLOW DIAGRAM

12

MATLAB CODE
clc;

% Extract out the H, S, and V images individually

clear all;

hImage = hsvImage(:,:,1);

close all;

sImage = hsvImage(:,:,2);

% Read image LOAD AN IMAGE

vImage = hsvImage(:,:,3);

[filename,pathname]=uigetfile('fabricTest1.jpg','C:\Users\abhi
nav arora\Documents\matlab files\');

v2Image = im2uint8(vImage);
image

scan_img1=imread(fullfile(pathname, filename));

%Noise reduction using mean filter

scan_img=imresize(scan_img1,[300 300]);

z=fspecial('average',3);

%Convert the RGB Image to HSV Image

filtImage=imfilter(v2Image,z);

hsvImage = rgb2hsv(scan_img);

%Display

%Conversion into 8-bit

13

CONTD..
% Setting the Threshold Value For Hue,Saturation and
Value

%Display the Mask Image


figure(3);

valueThresholdLow=0;
subplot(1,2,1);
valueThresholdHigh=0.2;
imshow(totalMask);
% Now apply each color band's particular thresholds to
the color band
totalMask= and((vImage >= valueThresholdLow),(vImage <=
valueThresholdHigh));
%Dilate the Output Image to Smoothen the Image%
w5=strel('disk',10);
totalMask_dilated =imdilate(totalMask,w5);

title('Mask');
subplot(1,2,2);
imshow(totalMask_dilated);
title('Dilated Mask');

14

CONTD..
%Applying mask to H, S and V

figure(4);

hue1=totalMask_dilated.*hImage;

imshow(rgbimage_dd);

sat1=totalMask_dilated.*sImage;

title('Fabric Defect RGB Image');

value1=totalMask_dilated.*vImage;
%Again Converting the HSV image to RGB image
hsv1(:,:,1)=hue1;
hsv1(:,:,2)=sat1;
hsv1(:,:,3)=value1;
rgbimage_dd = hsv2rgb(hsv1);

15

RESULTS - TEST IMAGE 1

16

CONTD.

17

TEST IMAGE 2

18

CONTD..

19

TEST IMAGE 3

20

CONTD..

21

TEST IMAGE 4

22

CONTD..

23

CONCLUSION
Using the HSV filtering and thresholding methods the the defect in the fabric was clearly detected. However, the
efficiency of this method was 85% because defects of very small size were undetectable. This method proves to be
successful for images which have dark color defects as in holes and stains. But the weaving and yarn defects still
remain undetected.
For our next term, wed be using spectral approaches such as Gabor filter etc to detect colour spills, weaving
and yarn defects and improve the efficiency of current system.

24

REFERENCES

Sheetal Thorave Prof. M.S.Biradar | Fabric Defect Detection Based on Fuzzy C-Mean Algorithms |
( International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 2 Issue: 2 367 371) IJRITCC | February 2014.

S.S. Bedi1, Rati Khandelwal |Various Image Enhancement Techniques- A Critical Review(International
Journal of Advanced Research in Computer and Communication Engineering Vol. 2, Issue 3) | March 2013.

Abdel Salam Malek | Online fabric inspection by Image Processing Technology | University of Haute
Alsace | 2012.

25

You might also like