SMT Components Appearance Defect Detection System Based On Machine Vision

Electronic components are in great demand in China, and would be a vast market. And electronic components are also gradually to the trend of development of intelligence, integration, miniaturization. This would increase more difficulties in the testing for the products. Because the manual inspection is great labor intensity, low efficiency, Real-time performance, and is vulnerable to the influence of subjective factors. These would greatly restrict the promotion enterprise’s production efficiency and product quality. With the manual inspection having so much shortcomings, So we need more advanced testing methods to test the product appearance defects. With the development of the modern imaging technology, computer technology and image analysis and processing technology, there is a measured and detected technology by machines instead of human eyes. We also call it machine vision. Machine vision is a kind of automated testing technology, which has no contact, no damage to the detected objects. It has been applied in many areas and play a more and more important role. In this paper, I combined with the actual project and develop of SMT components appearance defect detection system based on machine vision. It will be used to accomplish detecting the SMT components appearance defects.In this paper, considering the capacitance of SMT components as the research object. First, particularly introduced the types of defects of the SMT components and classified for them. Classified the defects on the middle ceramic body parts or the both ends of the electrodes. Then introduces the appearance defects by using image processing aspects of theoretical knowledge, including image segmentation, morphology, Blob analysis, etc. In this paper from the overall to the part of the test methods. First segmented the SMT components from the background. And then checked whether its overall size meets the requirements. Then used the Blob analysis detect the presence of dark spots and scratches. After the overall test, if didn’t detect the defects, entering the next test. Using the threshold segmentation on the SMT components, segmented the ceramic body and the electrodes from the SMT component. Calculated the area of these three parts. Testing each area whether meets the requirement. If didn’t meet, the electrode or ceramic body has the problem of lack of angle, otherwise proceed to next step, detecting of holes. Only each part of the test meets the requirements, the final result to judge as qualified. If one part does not meet the requirements, do not need to enter the next step of detection, judged to be unqualified SMT components immediately. The last part introduces the hardware and software of the system. It introduces the system’s framework and its working principle. Especially the visual part, the selection of the camera, camera lens, light source, light controller has carried on the detailed introduction. Software part is based on the vs2010 development platform, using it to write the system’s software.