Regarding the obvious disadvantages of small cracks in the metal structure of the helicopter anti-twist joint by regular manual inspection,an automatic image monitoring method for the metal structure crack is designed.The image acquisition module is constructed to collect the metal structure image of the anti-twist joint of the helicopter at the appropriate position,the collected image is preprocessed,enhanced by gray stretching method,and the image noise is filtered by median filtering method to upgrade the image definition.he improved OTSU algorithm and Canny multi-scale spatial edge detection algorithm are used respectively to determine the maximum and minimum thresholds for crack judgment,and the candidate crack segment features are obtained.Tensor voting is applied to remove the discrete points in the extracted candidate crack segment features and connect the crack segments,so as to realize automatic image monitoring of the cracks.The experimental results show that the image preprocessing effect of the proposed method is more ideal with smaller extracted crack error and higher frame rate per second
关键词
直升机防扭接头/金属结构裂纹/灰度拉伸/双阈值Canny算子/张量投票
Key words
helicopter anti twist joint/cracks in metal structures/gray scale stretching/double threshold Canny operator/tensioning voting
维普
万方数据