Research On Crack Image Detection System Of Small And Medium-sized Concrete Bridge

Crack is one of the most common diseases of the small and medium-sized concrete bridges.To realize the non-destructive cracks detection and improve operational safety,this paper studies the image detection system of crack for small and medium-sized concrete bridge.Through the requirement analysis of the crack detection for small and medium-sized concrete bridges,the main technical indexes of the crack image detection system are defined.The overall scheme of image detection of bridge crack is designed.The key technologies such as image acquisition system,image splicing and crack image detection algorithm are analyzed.In order to meet the performance requirements of the image acquisition hardware system for bridge crack detection,the hardware selection design of remote control shooting module,motion control module,lighting module and ranging module are implemented.The parameter matrix and distortion coefficient of the camera are obtained through camera calibration,which realizes the image calibration of bridge crack.The overlap area of adjacent image is estimated by phase correlation algorithm,then the SIFT feature points are detected and selected in the overlapped area.The improved RANSAC algorithm enhances the accuracy and efficiency of feature registration for sequence image.The research of multiple image splicing algorithm for small and medium-sized concrete bridge is completed through the weighted fusion of the projective image,which realizes the fast fidelity splicing of large area crack image.The initial location of bridge crack is confirmed by dividing mesh region of the large area image.The crack morphological features are extracted accurately by improving algorithms of the image denoising enhancement,the crack edge segmentation,and the morphological processing.The accurate detection of bridge crack feature data is realized through the identification of target cracks and the calculation of characteristic parameters.According to the process of crack image detection of small and medium-sized concrete bridges,the data processing software system is developed.The system experiment of image acquisition,image splicing and crack detection are implemented.The results show that the crack image detection system not only improves the efficiency of crack image detection,but also enhance the detection accuracy of crack width to 0.1mm,which can meet the detection requirements of engineering.The fast fidelity stitching algorithm of large area bridge images and the accurate detection algorithm of crack feature data are studied in this paper.The research results establish foundation for further engineering research of crack image detection system,and provide some technical reference for image detection in other fields.