System Design Of Fabric Defect Detection Based On Image Processing Technology Analysis

The quality improvement of fabric product is the key of survive to the textile enterprises. And the fabric defect detection is a key link to ensure product quality in the process of textile enterprises. At present, the fabric defect detection in the domestic textile enterprises is mainly rely on the inspector looking after the inspecting machine.Because of low speed,unreliability and easily affected by subjective on human test, as well as the injury of inspectors, eyes and body, it makes less and less able to meet the needs of modern production. Now, as a result of the computer science increasingly mature, the textile industry is greatly promoted for better development, and makes the automatic detection of fabric a hot spot at home and abroad research. Based on the related theory research on wavelet transform and threshold segmentation method, a kind of fabric defect detection system based on image processing technology was designed.First, the image was processed for illumination compensation on the pre-treatment.Second, the processed image was processed for noise reduction by using wavelet analysis and the combination of median filtering mixed noise reduction method. Third, the fabric defection was enhanced through the image enhancement processing. The fabric image after pretreating become a dominated by defect information, background noise and texture for minor digital image. Then the improved Otsu threshold segmentation method was used to test the treated fabric defect image segmentation, comparing the segmentation results with the unimproved Otsu threshold segmentation method, verifying the effectiveness of the algorithm in this paper. The article shows that the improved algorithm is accurate and reliable based on an analysis of a large number of experimental data, and can ensure the accuracy in detection, which has solved the incomplete segmentation and inaccuracy detection in the traditional detection process.Finally the composition and function of all parts was analyzed and the experiment was carried on the hardware platform. Then the experiment results of fabric defect were analyzed. System experimental results show that the improved algorithm proposed related on the design of fabric defect detection system can realize the accuracy of fabric defection.