Research On The Key Technologies Of Edge Detection Of Mechanical Parts Based On FPGA Image Processing

With the rapid development of digital image processing technology,the edge detection technology,as the basic content of digital image processing,has increasingly applied in the fields of machine vision,face recognition,target tracking,industrial detection and other fields.The core of edge detection technology is edge detection algorithm.The edge detection based on Sobel operator is generally applied in industrial detection practices,however,at present most edge detection technologies are based on PC software serial processing,which is extremely arduous to satisfy the real-time requirements for the system.The edge detection technology based on FPGA image processing makes full use of the advantages of combined processing of hardware and software,which can greatly improve the speed of edge detection.At the same time,the detection accuracy of the system can be guaranteed by improving the design of algorithm.This paper proposes an adaptive threshold multi-direction Sobel edge detection algorithm based on FPGA for actual mechanical parts detection scenarios.In order to improve the precision of edge detection algorithm,this paper extends the existing Sobel edge detection algorithm and uses adaptive threshold method to improve the image processing algorithm.Finally,through the FPGA image processing hardware platform,this paper chooses standard mechanical parts to perform edge detection experiments.And the experimental results reveal that the improved algorithm on the test accuracy and detection effectiveness has obvious advantages,therefore it is to verify the effectiveness of the improved algorithm.The main research contents of this paper are as follows:Through extensive reference to relevant literature in the domestic and overseas,analysis of image edge detection methods and the current situation of industrial parts detection,this paper researches and formulates an overall design scheme of parts edge detection system based on FPGA image processing.The mechanical structure is used to control the spatial three-dimensional movement of the image capture camera: a three-dimensional coordinate platform is selected as the mechanical control part of the image capture,and the movement control of the image capture camera is realized through three parts: a three-dimensional servo mechanism,a precision optical platform,and a part stage.An adaptive threshold eight-direction Sobel edge detection algorithm is proposed to solve the problem of insufficient detection accuracy of traditional first-order and secondorder differential edge detection algorithms.By extending the direction of the traditional Sobel edge detection operator,the accuracy of edge detection is improved.At the same time,the design of adaptive threshold algorithm makes designer need not to select the threshold only by subjective experience,improves the real-time detection and automation degree.Using the idea of parallel pipeline design,combined with the hardware description language Verilog and related design hardware and software,the improved algorithm is implemented on FPGA.The adaptive threshold algorithm and the extended multidirection edge detection algorithm are implemented using FPGA hardware in parallel.Finally,the standard part detection experiment results show that the improved algorithm can obviously enhance the detection accuracy of the system,which verifies the feasibility and real-time of the improved edge detection algorithm in this paper.