A PCB defect classification method based on an image multi-shape normal gradient difference value

• Main Authors: ZHANG MEIJIE, LI XIANWEI, WEI DENGMING, ZHANG LU, WANG HUALONG, HU XIAOQIANG, ZHANG PENGZHONG
• Format: Patent
• Language: Chinese; English
• Published: 25.06.2019
• Subjects: CALCULATING; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention provides a PCB defect classification method based on an image multi-shape normal gradient difference value. The contour of each feature shape is obtained through feature classification of an original template image, the normal vector of the corresponding contour shape is obtained, the defect position is accurately matched through a template, the defect type is judged by searching pixel point gray level gradients on the normal vector, defects are classified, and the phenomenon that classification is wrong due to the fact that classification is conducted only through features is avoided; The PCB defect detection and classification method improves the detection speed and precision of PCB defect detection and classification, effectively improves the detection rate and classification accuracy, reduces the occurrence of defect false detection or missed detection, achieves PCB defect detection and automatic classification, saves manpower, material resources and financial resources, and improves the pro