Research Of Fabric Breakage Online Detection System Based On Machine Vision

With the innovation of science and technology in the world, the moment, the growing popularity of machine vision detection once the application has been overwhelming majority of researchers in all walks of life. In the food production process, typically uses cloth for food production food was filtered to remove impurities, high-quality food to retain liquid, and then follow-up treatment; however, due to the fabric used in the process is repeated often require folding and extrusion, easy damage caused by cloth, then re-use if damaged cloth production will cause the waste of food raw materials.Therefore, this paper studied the cloth that runs on a platform DaVinci DM6467 broken-line detection system. The system uses the industrial machine vision, image damage detection technology and network control technology is designed to meet industrial processes distributed online detection, control, and management requirements. The main work is as follows:The first introduces the status of machine vision research status of domestic and international research and fabric detection.Second, the introduction of the general framework cloth broken line detection system, which includes cloth detect hardware and software systems. Finally, according to the actual requirements of customers, we present key performance indicators of the detection system.Third, the introduction of contrast, the principle of choice and hardware cloth fabric collection breakage detection light source, and built by the infrared light source, optical systems, industrial high-definition cameras, TMS320DM6467 development board, an image processing unit, network communication systems, synchronous controller, the system controller, power supply component hardware test platform.Fourth, build a Linux development environment and a variety of server configurations, and tells the DaVinci Codec Engine algorithm engine technology and API functions, Codec Engine algorithm development process, and analyzes the Linux multi-threaded design (basic operation thread, synchronization and mutual exclusion), and finally introduces cloth image breakage detection application software design.Fifth, with cloth breakage testing requirements for image filtering, enhancement, segmentation, and other classic algorithms, experimental analysis of comparative advantages and disadvantages of various algorithms, select the appropriate filter value of this project, based on two-dimensional category variance between multi-threshold segmentation algorithm and Leonardo da Vinci SEED-DVS6467T development board to verify the feasibility of the whole algorithm.Finally, with Linux as the software platform, we developed a cloth broken line detection system software, real-time automatic detection of the industrial automation production lines damaged in the process of common cloth; capable of displaying images and damaged damaged location, size and other information, and is able to damaged image data storage management, and alarm information.