Design And Implementation Of Intelligent Camera Defect Detection System Based On Heterogeneous Multi-core Platform

The development and application of smart camera defect detection technology has solved the limitations of today's manual detection,such as low detection rate,slow detection speed and easy to operate errors.Based on the wide range of vision detection applications of smart cameras,specific detection devices and image processing algorithms can be set for different detection environments and detection contents.The smart camera integrates the functions of video image acquisition,processing,analysis,and transmission.Its features such as high performance,low power consumption,and miniaturization make it widely used in various industries.The main processor of the smart camera selects a Zynq chip with an integrated ARM + FPGA heterogeneous multi-core architecture,and completes the image processing system platform design based on the idea of hard-soft co-design.This article mainly relies on Zynq-7000 AP SoC as the main processor for intelligent camera defect detection system design and implementation,based on visual inspection image processing technology,designed to meet the actual application of defect detection algorithm,optimize the application layer image processing pipeline,and finally verify system performance And analyze the results.The main work of the dissertation includes the following aspects:(1)Introduce the development status of surface defect detection technology and the advantages and disadvantages of various types of algorithms and systems.According to the requirements of defect detection system,design the system based on the development platform.(2)Analyze some commonly used image preprocessing techniques,perform algorithm principle exploration and combine the actual processing effects of the system,and select an image preprocessing technique suitable for defect detection schemes.(3)To study the processing flow of defect detection algorithm,focusing on the principle of contour extraction and extraction process,the principle of detection and labeling of defective connectivity,the calculation process and method of defect parameters,and the principle of Hough transform.(4)design image acquisition and processing pipeline system framework,based on the original Xilinx reference design to achieve the application layer image processing pipeline optimization structure optimization,defect image processing function frame design,display mode optimization,video data acquisition based on the V4L2 framework With processing.Finally,the application layer processing algorithm design of the detection system is completed.The test platform is set up to carry out joint hardware and software testing of the algorithm,and the defect processing results are displayed.The system can detect defects and obtain related defect parameters to meet the design requirements.