Three-dimensional Pavement Disease Detection based on Three components Decomposition and TUFF Algorithm

While automated road surface information scanning has become very popular, however, the method of extracting available disease information from the collected pavement data is not so efficient. This paper provides a new crack extracting method using 3D pavement components decomposition model and tubular flow field. The method decompose the 3D pavement data into 3 parts: low-frequency component, sparse component and texture component. Then, tubular flow field is used to extract crack information, and indicate the distribution of fractures in the original data map. This method can extract crack as well as connect the broken pieces. And the accuracy of extracted crack information is over 95.75%.