Research On Automatic Inspection Technology Of Circuit Board Based On Machine Vision

Machine vision systems are widely used in industrial production lines,especially in areas where operational hazards exists and cannot provide continuous human visual supervision.Considering the current situation and problems of printed circuit board quality inspection,we developed an automatic inspection technology based on machine vision to realize automatic and pipelining inspection.The main contributions are as follows:(1)Considering the existing problems and production requirements,a machine vision-based circuit board inspection system that can realize automatic detection was designed in this paper.Then we analyzed and selected the hardware system,including lighting mode,light source selection,camera selection as well as image acquisition card matched to the former.We further designed the corresponding software part according to the functions to be implemented by the system.(2)Taking into account the system processing requirements,the acquired images needed to be preprocessed.We selected a series of approaches which were more suitable for circuit boards to enhance the quality of images,including denoising,segmentation,and edge detection.(3)We proposed algorithms for identifying common defects of the circuit board in the production process,including missing parts detection based on template matching with gray value;component welding anti-identification based on polarity circle recognition;mis-welding detection using neural network via digital component recognition;poor soldering identification based on color matching.(4)The system simulation software system was designed by MATLAB GUI.We employ the developed system to detect multiple sets of boards with known defect conditions,and got the experimental results.Furthermore,we analyzed the possible shortcomings of the system and the potential solutions.