Research On Image Acquisition System Of Woven On-line Defect Detectiong Based On Machine Vision

Fabric defect detection is a very important quality control in the process of textile industry, a cloth weaving, first go through inspection, identification of its defects, assess its grade, as the main basis of trade settlement. Traditional manual testing, low efficiency, miss and false detection rate is higher, and offline testing equipment based on machine vision is expensive, companies often can’t afford it. In order to solve above two problems, this paper designs and realizes a kind of online machine fabric defect detection based on machine vision, image acquisition system adopted by the technical scheme overcomes the limitations of existing inspecting machine.First of all, this thesis topics on the background, as well as the fabric based on machine vision detection technology at home and abroad, this paper puts forward the main research targets,summarizes the main difficulties, namely online fabric problem such as transmission, image acquisition, positioning control, put forward the overall design scheme of the system.Secondly, from the aspects of system hardware structure, the system design scheme of mechanical transmission and mechanical control in detail. For offline testing results can not be timely feedback to the problem on the loom, adopt the method of the on-line detection equipment installed; In order to solve the on-line detection of loom vibration influence on image acquisition,adopt the method of the closed curly; In order to reduce the system development costs, enhance practicability, adopted a single array camera mobile photo image acquisition method. Around the three solutions, the hardware design of the key work focused on the mechanical structure design,hardware selection and the design of the mechanical control, etc.And then, in the aspect of system software structure, solves the problems of image transmission, storage, positioning control, communication and other problems. Mainly for image transmission and storage methods of setting, image transmission by cable of the TCP/IP protocol,the image location information for the Modbus ASCII protocol, a serial port communication program design in detail, and the writing of the specific software programs.Finally, the hardware and software of the whole system combined with the comprehensive debugging, each equipped with good hardware components installed, determined the related parameters and the final model machine, and the image acquisition and location information for experiments, through experimental verification, successfully captured for subsequent disposal of fabric defect image, and can accurately obtain the location of the defect image on the cloth, to achieve the goal of this study. The overall performance of the system reaches the expected.The machine online fabric defect detection based on machine vision image acquisition system, the success of research and development for further image processing, defect detection algorithm on the platform of the basis for further research and development testing, as well as all kinds of textile on-line detection provides a good platform.