Research And Implementation Of Pavement Crack Detection System Based On FPGA

More than 30 years of Chinese Economic Reform and Opening-up, rapid development of the economic construction of our country,with the national attention to infrastructure, road construction is also rapid developing in our country, however,the continuous improvement of the highway network, accompanied by the workload is huge road maintenance management. Pavement diseases are developed from the crack, and pavement diseases may cause traffic accidents, endangering national safety of life and property losses, so the scientific and periodic pavement crack detection can avoid the road more serious damage, as much as possible to reduce the probability of the occurrence of traffic accidents. However, traditional artificial pavement crack detection has been unable to meet to detect the larger diseases of the roads, so the demand of the development of the intelligent pavement crack detection system is very urgent. In this paper, the digital image processing technology, FPGA technology and BeiDou satellite positioning technology our country has independent intellectual property rights are studied to combine into the intelligent pavement crack detection system.In this paper, the design of pavement crack detection system which is based on FPGA uses CMOS image sensor to acquire video and image, uses the SDRAM to cache video and image, uses the VGA to display video and image, BeiDou positioning location module acquire the location information, which can display the identify and classification results and geographical location information on LCD12684. The module of the whole system is conducted using Verilog HDL to complete. System will first preprocess the collected image , convert the RGB565 image from CMOS for a gray image, then the converted gray image using median filtering noise reduction processing, after the segmentation of the image, in order to improve the detection accuracy and anti-interference ability, puts variable the denominator of the different operation into the Sobel edge detection algorithm to extract effective goals, and uses the morphology for restoration of the image, then uses the projection algorithm for feature extraction and recognition classification of the target image, finally generates an enabling signal to receive location information for the extraction of effective information. Building and debugging the system, then through the outdoor experiment results show that the intelligent pavement crack detection system is designed in this paper can implement the normal pavement crack detection.