Research On PCB Industrial On - Line Inspection System Based On Machine Vision

Development of technology driven increasingly miniaturized electronic products and digital, and more and more high accuracy and. density PCB board to bring greater pressure on manual inspection. Manual inspection, there are many drawbacks, such as prolonged high-intensity work, eye fatigue more easily, leading to missed cases have occurred, when the detection results are not satisfactory. With the advent of the industrial 4.0, the manufacturing sector will enter a new era and stage. The continuous development and progress as well as the rapid growth of machine vision and image processing PCB industry, PCB industry research online testing system based on machine vision is of great significance and value.This paper analysis the PCB industry line detection system based on machine vision from the PCB detection technology and image processing point of view. The main work is as follows:(1) From the present situation of PCB detection technology development on the importance of the PCB defect detection and analysis of the PCB industry online testing system structure and main influencing factors based on machine vision.(2) Analysis of the PCB images characteristics collected from the PCB defect detection system, the use of image pre-processing methods to improve the image contrast and reduce noise.(3) Considering there may exist translational or angular offset or other issues between test image and standard image, contrasts the template matching and image registration method based on Harris corner feature and implements the registration between test image and standard image.(4) There may be a short circuit, burrs, defects and other problems for PCB traces, through binary morphology algorithm can realize the detection of the defect; for measuring the line width of this paper, the gray columns of pixels cumulative method can effectively detect the line width from eligibility.(5) Considering the PCB may exist brightly colored chip capacitors missing, take color information based method can detect defects effectively; for chip, there may exist placement in the wrong direction, taking the polar circle Hough transform detection method can detect this defect.(6) For PCB circuit board, there may exist more than or less parts and other problems and taking subtraction method can effectively realize the detection of the defect.(7) Resistor for character recognition, it uses template matching method can achieve detection of resistance characters.