The System Design For BGA Void Rate Detection Based On X-Ray Digital Image Technology

BGA(ball grid array)is widely used in the industry as a new chip package with good characteristics,especially in the core components of electronic products.Therefore,the quality inspection of BGA solder joints becomes more important.In this paper,we have developed an automated void rate detection and analysis system of BGA solder joints based on the X-ray digital image processing technology.In this paper,we study the extraction of BGA solder joints in complex background and the calculation of the void rate.At the same time,the function modules such as image acquisition module,motion control module,micro-focus X-ray source control module,X-ray image processing module and database module are designed.We first study the denoising of X-ray digital image,and select the frame integral denoising method which is suitable for real-time detection.Then we design an automatic navigation algorithm to locate the detection target accurately and synchronize the motion and denoising by using multi-thread technology.At the same time,the micro-focus X-ray source control module is designed as an ActiveX component,so that it can be embedded into the application.After the image acquisition is completed,this paper detects the solder joint contour by the adaptive template matching method based on the edge contour feature.Compared with the threshold segmentation method,the interference of the complex background is reduced,and the void detection of abnormal solder joints such as short circuit and damage is avoided,while it have a strong ability to detect the solder joints which is covered by shadows.Then we use the image pyramid to speed up the matching speed,and achieving sub-pixel edge by using least squares to adjust the matching pose.Then,the numerical characteristic of poor contrast voids are highlighted by the difference between the collection of pixels in the neighborhood.And so a blob analysis method of poor contrast void is realized and the integral image is used to accelerate the detection process.We use database to storage and query the data of BGA solder joints.Finally,through the system test proved that the system is stable,efficient and accurate,it can meet the needs of industrial detecting.