Design And Research Of Automatic Fabric Defect Detection System Based On TMS320DM6446

Fabric exists defect will reduce the quality of the product, to reduce product sales and sales prices fall, profits decrease, so the textile enterprises to fabric defect detection very seriously. However, the modern textile enterprises using artificial detecting fabric defects still exists the problem of low efficiency and low precision and accuracy. In order to solve these problems, and puts forward the fabric defect detection automation research topic, the realization of the automatic detection of fabric defect can improve the production efficiency and product quality, so the design of the automatic fabric defect detection system and the research is of great significance. This topic is design and research of automatic fabric defect detection system based on TMS320DM6446, the main research work is as follows:(1) Design and implementation of fabric defect detection algorithm. This topic research object is the defect of the grey fabric, the main detecting grey fabric defects. According to fabric defect detection method, this paper puts forward an improved Otsu image segmentation of fabric defect detection algorithm. Due to the classical Otsu algorithm efficiency is low, cannot satisfy the real-time demand automatic fabric defect detection system, as a result, the simulated annealing optimization(Otsu_sa) threshold image segmentation algorithm, simulated annealing algorithm(SA) is excellent global optimization ability, improve the efficiency of the Otsu algorithm and meet the requirements of real-time automatic fabric defect detection.(2) Design and implementation of fabric defect detection of image processing. Firstly, the fabric image preprocessing, including uneven illumination correction, median filtering and histogram stretching, fabric image after preprocessing, uniform illumination, the image noise filtering, strengthen the fabric defect image information, and then uses the Otsu_sa fabric defect detection segmentation algorithm, finally the division after image processing, morphology to get clear defect detection results.(3) Design and implementation of fabric defect detection of image processing software and hardware. This topic uses a SEED-DVS6446 embedded system developing platform, industrial CCD camera and LCD liquid crystal display of fabric defect detection image processing hardware platform, the design of fabric defect detection algorithm on DSP and encapsulation, fabric defect detection on the ARM the design of the software process, finally uses the DVSDK software platform as a medium, realize the fabric defect detection software design.(4) Automatic fabric defect detection system of experimental verification and optimization software. Finally, set up a set of automatic fabric defect detection system hardware platform for defect detection algorithm for experimental verification and optimize the system software. The experimental results show that the system has a good real-time performance and high detection efficiency, accuracy and precision, has the very good application to textile enterprises, to fabric defect detection real time online.