Research On Automatic Detection Method For Patching Image On Asphalt Pavement

The automatic detection method for distress patching has been a research hot in the area of road traffic. After roads are built, kinds of distress will appear in various degree,because of payload,environment and so on. The distress must be repaired,and the patching should be detected. It is the most urgent problem for the development of highway. Because of the low contrast between distress patching and pavement breakage, and much noise in pavement images, the common detection method is difficult to extract patching information. Compared with traditional manual measuring, automatic detection method for distress patching is not restricted to time or environment, with high expense and slow speed. A automatic method of patching information extraction is proposed in this paper,the primary responsibilities are:1. Gray-level regulation is adopted to enhance the image with poor quality. For the problem of noise caused by oil stain and stone. Median filter based on priori knowledge is proposed into the process of image smoothness.2. For the problem of the poor segmentation results by traditional segmentation algorithms,a segmentation algorithm based on window contrast is proposed,and various kinds of methods for different pavement are proposed.3. An algorithm based on the intensity factor of zero-crossing is proposed to detect contour in patching images. Fast LoG is used to improve operation speed, intensity factor of zero-crossing is adopted to detected zero-crossing as judgement criterion. It can remove sham zero-crossing, and the position of patching contour is calculated accurately.4. Detection and selection of patching feature is used to finish patching classification and preliminarily measurement data of patching is obtained. Finally, the software realization and module of the algorithms are designed based on algorithm research in this paper.