| The significance of the electronic industry for a country is very important,it plays an important role in industries such as national defense,manufacturing,information transmission,and software services.PCB is a critical component in the entire electronics industry chain,so conducting quality inspections on PCB is an important step in improving product quality.The bonding pads and vias in the PCB serve to fix components,connect components and circuits,and fix the substrate.The bonding pads and vias are the critical elements for a PCB to achieve interconnection.With the development of technology,the complexity of PCB is continually increasing,and the size is continually shrinking.Traditional manual inspection can no longer meet current stage requirements,which poses new challenges for PCB defect detection.Therefore,researching a system for detecting defects in PCB hole rings is of great significance.Thesis is based on machine vision technology,combining processing algorithms,developed an automated defect detection system from image acquisition to image defect detection.The system consists of two parts: hardware system and software system.The hardware system includes lighting system,motion control system,and image acquisition system.The motion control system is responsible for transporting the PCB,and the image acquisition system is responsible for shoot images of the PCB.The software system uses Python to write image processing program to achieve image acquisition,image preprocessing,image registration,defect detection and other functions,and uses Py Qt to make man-machine interface to achieve man-machine interaction.In the image preprocessing section,in order to solve the problem of inaccurate segmentation of the target area caused by dispersion at the edge of the bonding pad and solder unevenness.In the thesis,Laplace algorithm is applied to enhance the target area,analyze the RGB characteristics of the background and target area,combines growth algorithms to design segmentation criteria and growth criteria for accurate image segmentation.Finally,uses morphological operations to eliminate small defects.To solve the problem of time-consuming in image matching with template image and sample image.In the thesis,combines the characteristics of images and the system’s requirements for real-time performance,optimize SURF algorithm,based on the invariance of feature descriptors under scaling transformation conditions,by utilizing the relationship between scaling factors and transformation matrices,the computational complexity of the program is reduced and the registration time is shortened.The gradient Hoff algorithm has some shortcomings in detection efficiency and accuracy when locating and measuring holes and rings.In the thesis,the parameters of the algorithm were optimized by predicting the relevant parameters of the target region,so as to improve the accuracy and precision of detection and shorten the detection time.Through analysis the characteristics of the defects such as convex,concave,hollow,burr in the hole,hole deformation,too large and too small of the welding bonding pad,the relevant defect detection algorithm and quantitative evaluation criteria were designed to realize the rapid identification and intelligent judgment of these defects.The experiment shows that the system can realize the rapid identification and intelligent judgment of the hole and ring defects on the PCB board,and the accuracy rate is 98%.The detection time of a piece of PCB is within 8s,which basically meets the needs of industrial detection. |
