Automatic edge Detection Method of Power Chip Packaging Defect Image Based on Improved Canny algorithm

The current proposed image edge automatic detection method is susceptible to external interference and the detection results are fuzzy. To address this issue, a new image edge automatic detection method for power chip packaging defects is proposed based on the improved Canny algorithm. Using a dual filtering algorithm to remove noise from images of power chip packaging defects and improve image quality. The traditional Canny algorithm has problems of insufficient accuracy and threshold selection relying on human experience in edge detection. Therefore, the Scharr operator is used instead of the Sobel operator for image gradient calculation to improve the accuracy of edge detection; At the same time, the maximum class variance method (Qtsu algorithm) is used to adaptively select the optimal threshold of the image, thereby avoiding errors caused by human experience. The Laplace operator is introduced to sharpen the image and fuse the sharpened image with the original image, thereby preserving the information of the original image and enhancing the effect of edge detection. The experimental results show that the image edge automatic detection method for power chip packaging defects based on the improved Canny algorithm studied has a high noise reduction ratio and strong anti-interference ability. The boundary error for defect image detection is less than 2mm, and the results are relatively accurate, in line with the inspection standards for chip packaging, and have high stability.