基于数字图像处理技术的路面裂缝检测算法分析-analysis of pavement crack detection algorithm based on digital image processing technology.docxVIP

南京邮电大学硕士研究生学位论文 ABSTRACTABSTRACTIn the work of road maintenance, road crack is the most important indicator of road quality. In China, we mostly detect the cracks on the highway manually, but it is no longer able to meet the testing requirements, the road crack detection technology which is based on digital image processing technology becomes the best choice of road crack detection, it can not only save human labor, remove the subjective human factors, but also evaluate the quality of the road quickly and accurately .This thesis is devoted to the studies of the key technologies of pavement crack detection, and has covered the following areas:this thesis illustrates the pavement crack image enhancement algorithms: given the fact that there are many shortcomings regarding the classical enhancement algorithms this thesis presents a pavement crack enhancement algorithm based on the guided filter, this algorithm can not only reduce the noise effectively, but also improve the contrast of the cracks and background, and has a good timeliness;this thesis has studied the pavement crack image segmentation algorithms: traditional segmentation algorithms have defects, this thesis suggested the mean shift segmentation algorithm which is based on the nearest neighbor to the pavement crack image , this type of algorithm has been proved to be able to obtain good results;this thesis has also covered the studies of the classification of road cracks and parameter extraction algorithm, basically, the thesis identified the miscellaneous cracks and linear cracks based the threshold value , then classified transverse cracks and longitudinal cracks based on projection algorithm, and combined skeletonization algorithm and the geometric characteristics of cracks to calculate the parameters of the cracks.After a large number of experiments in asphalt, the results indicate that this type of road detection algorithm is able to achieve very good results, and is very useful in classifying cracks