In the electronics industry, the an lèt detache sikwi tablo (PCB) is the main component of various electronic products. The soldering quality of the components on the PCB directly affects the performance of the product. Therefore, the quality inspection and testing of PCB boards is the quality control of PCB application manufacturers. An indispensable link. At present, most of the PCB soldering quality inspection work is done through manual visual inspection. The influence of human factors is easy to miss and misdetect.

Therefore, the PCB industry urgently needs online automated visual inspection, and foreign products are too expensive. Based on this situation, the country began to develop this. Detection Systems. This paper mainly studies the identification of PCB board welding defects: identification of color ring resistance, identification of component leakage welding and identification of capacitor polarity.

The processing method in this paper is to combine the reference comparison method and the non-reference comparison method to obtain the PCB board image from the digital camera, and use the methods of image positioning, image preprocessing and image recognition, feature extraction to realize the automatic detection function. Through the experiment of multiple PCB images, the positioning method of PCB image features is improved to obtain accurate image positioning.

The standardized part of the breakdown is an important part. This is the circuit board and the standard board. Perform the first step of an exact match. In the image preprocessing part, a new geometric correction method is used to correct the image to obtain accurate PCB images and precise pixel coordinates of each component, and perform image binarization, median filtering, edge detection and other methods to obtain the best Recognition. In the next image recognition of the effect image, features are extracted from the image after preprocessing, and different recognition methods are adopted for different welding defects.

Applying statistical methods to extract relatively standard color energy to accurately identify the color ring resistance, and solve the identification of the color ring resistance from color segmentation to saturated filling. Regarding the geometric characteristics of the polar capacitor, the geometric identification method is applied to the application of component leakage welding. The probabilistic recognition method has achieved good recognition results. Therefore, this method has a good reference value for the automatic identification of PCB defect detection in China.