Design of circuit board defect detection system based on image processing
Aiming at the problems that the traditional circuit board defect detection features low speed and high cost due to manual detection is,this paper investigates and designs a precise and efficient circuit board defect detection sys-tem utilizing image processing techniques implemented on field programmable gate array(FPGA).Through using different interest intervals for different image information on the basis of traditional image enhancement algorithms,the method yields notable improvements in image enhancement.In order to reduce the impact of logo characters on the circuit board's computational efficiency for the matching algorithm,a silkscreen removal technology is used to eliminate the superfluous logo characters on the circuit board to decrease the computational load of image matching and enhance the detection speed.Additionally,adopting the de-averaging method based on the traditional sum of absolute differences(SAD)template matching algorithm to calculate image information to alleviate the impact of lighting variations on the overall performance.The traditional 2-operator Sobel edge detection is extended to 8-oper-ator edge detection,resulting in enhanced clarity and prominence of edge information.Using FPGA as the hardware plat-form,Verilog HDL hardware logic language is implemented in Vivado development environment and subse-quently deployed onto the FPGA.The experimental results demonstrate that the system attains an average detection accuracy of 98.53％,and its detection time of single circuit board is 8.204 s.Its system design provides substan-tial enhancements in both detection accuracy and speed,while maintaining a cost-effective research approach.
万方数据
维普
