A Method of Welding Defects Detection in X-ray Image Based on BP Neural Networks

摘要：

Firstly,welding image is processed get binary image by the binarization segmentation using the Method of Maximum Class Square Error.Then the binary images are filtered using the morphological structure element.The region growing method is used to search and extract the welding defect image region in the binary images.The moment invariants are applied to compute the eigenvector of the defect region.Based on back propagation (BP) neural networks,fuzzy factor is introduced to identify and classify for the welding defects.Experimental results show that the proposed method is effective.

关键词： welding defect detection t image binarization segmentation morphological structure element moment invariants BP neural networks

作者: Han Xiaowei Yin Xiaohui Zhang Wenqi Qi Xiaoxuan

作者单位: School of Information Engineering,Shenyang University,Shenyang,China Shenyang Foreign Enterprises Service Corporation,Shenyang,China

会议类型: 国际会议

会议名称: 2010 4th International Conference on Intelligent Information Techonlogy Application(第四届智能信息技术应用国际学术研讨会 IITA 2010)

会议地点: 秦皇岛

会议语种:英文

页码: 185-188

在线出版日期: 2010-11-05（万方平台首次上网日期，不代表论文的发表时间）