Study On A Handset Device For Measuring Building Crack

"Study on a Handset Device for Measuring Building Crack"is one project of Tianjin Society Development Fund. The device is handset and mainly used to measure building crack. The result of measuring is as one of impersonal and impartial gists to detect the failure and evaluate the quality of building. The measuring parameters are the length, the width and the tendency of crack. The measuring area is 25mm×25mm at one time and the measuring precision is able to reach 0.05mm. The tendency of crack can be shown clearly through the visual interface.The main tasks of this thesis are to design the instruction framework of the whole system in software and hardware, and the purpose is to realize the following func- tions: image-acquisition, image-processing, result-show, result-storation and the mutual communication with PC using USB.The major work of this thesis is as follows:1. The optic-imaging method is proposed to measure cracks and the image of crack has been acquired visually. The crack is projected on an optic-electronic sensor through optic system, and through that, the mathmatics model of the image of crack is built.2. The image is buffered using SDRAM and the high-speed high-resolution image-acquisition has been implemented. The digital image signal from the image sensor is buffered to SDRAM which is the data area of CPU through CPLD. CPU is capable to access this image data quickly in the image-compute process.3. The hardware structure for computing parameters of crack is built and the real-time, quick and high-precision operation of crack image has been implemented. The high-speed DSP has been used as processor and the image of crack has been processed real-time and the result has been shown with visual interface.4. The image segmentation arithmetic is designed to acquire the edge of crack. The crack image which is coded by RGB Bayer has been transformed to gray image.A suitable threshold is adopted to make the gray image into binary and the area of crack has been judged in blocks of binary image. In the area of crack, Laplacian operation has been implemented to acquire the edge of crack. The edge has been searched and flagged and the noise has been judged and abandoned according to the characteristics of crack. The parameters of crack, such as width, length and tendency,...