Automatic Bridge Crack Detection System Based on UAV Image Technology and Support Vector Machine(SVM)
In order to automatically detect the bridge parts with cracks in a timely and effective manner and give maintenance suggestions,a more complete whole-process bridge crack detection system is proposed.This system uses LOG filter to preprocess the image collected by UAV,identifies the crack type based on SVM,calculates the crack length and area ratio according to the pixel,and calculates the maximum width of the crack by morphological corrosion.Finally,the cracks are divided into three categories:transverse cracks,longitudinal cracks and other cracks,and the three important parameters of crack length,crack maximum width and crack area ratio are obtained.Then the health status of the bridge is automatically evaluated and maintenance suggestions are given.The designed system is verified by using the crack pictures of Jiaoping high-speed interchange bridge section in Henan Province collected by UAV.The results show that the developed detection system can realize the automatic calculation of crack parameters,and the overall classification accuracy of 96％is obtained.The detection results can be exported as.xls type files,which can be used as historical data sets for subsequent use.
万方数据
