Road Image Disease Detection Algorithm Automatically

This paper is focused on the study and implementation of auto-detecting of distresses in pavement image .Initially, the characteristics of pavement image are analyzed , then certain pre-processings are carried on the pavement image . Based on the definition and appearance of the distress image , the distresses are divided into two big classes : patch distress and crack distress .Then the paper reviewed the image segmetation and morphological methods . The attention is put on thresholding segmetation method which are famous for their high efficiency . Based on the above techniques , a hiberarchy model is presented which is flexible and effective in dealing with the complicated situation in the pavement image . An experiment on a large amount of data show the good performance of the model in finding the patch distress .For the crack distress , several classical methods in edge detecting are explored . Canny algorithm is chosen as the base of the detecting algorithm for crack distress . Several improvement are made . The tests on typical crack distress image prove the effectiveness of the algorithm.