科研之友
微信
新浪微博
Facebook
分享链接
摘要
With the support of modern dam facilities, many dams are equipped with multiple cameras. A large number of labeled images can be provided by the crack extraction process. But the most existed dam crack detection algorithms are based on one image. A new dam crack detection algorithm based on combined cloud model (CCM) is proposed in this paper. The algorithm can effectively utilize the detected image. The randomness and fuzziness of cracks are expressed by unified mathematical model. The algorithm extracts the pixel distribution rulers of crack and background through training samples. The expectation, entropy and hyper entropy of concept clouds are calculated based on the inverse cloud model. The concepts of the foreground crack and background are extracted by CMM. The membership degree of each pixel is calculated for crack detection. The results of experiment show that CMM algorithm has a good universality. The CMM method can extract the background and foreground concepts. Segmentation results show that our method performs well in eliminating noise points and detecting cracks.
