Research On Board The Machine Vision Inspection System And Method Based On

With the development of electronic technology，the integration rate and the output ofthe Printed Circuit Board are enhancing unceasingly, as a supporting body of electronicsproducts. In order to ensure the performance of electronic products, circuit boarddefecting technology has become very critical in the electronics industry technology.Compared with the traditional manual detection technology, machine vision detectionthat based on image processing technology, improves the efficiency and accuracy ofdefect detection. Therefore, it’s a very important practical significance to design anefficient and accurate detection circuit boards defective machine vision system.Firstly, we study the development of the circuit board and the circuit board detectiontechnology at home and abroad, and analyze the machine vision inspection technology’sresearch value and urgency. Hardware and software programs are designed for machinevision inspection system board. According to the detection needs, we study theperformance of the system parameters and selection on the basis of hardware, also, givethe software design process.Secondly, for a clear and accurate extraction circuit board defection featureinformation, The image is used the image gray to stretch image contrast enhancement,filtered through the weighted mean, adopted2D maximum threshold segmentation andSobel operator edge detection algorithm, in image pre-processing methods. The targetarea is separated quickly and accurately from the background area segmentation.Finally, it designs different recognition methods depending on the type of defectdetection. In this paper, it achieves the detection of chip capacitors sealing off in the useof HSI color model. Polar circle of chip through improved Hough transform is detectedquickly and accurately to determine whether the chip is flipped or not. For the nominalthe resistor value recognition, BP neural network is designed, and improved methodsare given. Meanwhile it describes the traditional template matching method, and sub-regional template matching method is proposed on the basis of that. Throughexperimental comparison, BP neural network detects nominal resistance values in agreater efficiency and more accurately. By identifying the nominal value of the resistor, the wrong equipment defects are achieved the detection.