Research And Implementation Of Brige Crack Detection Technology Based On Tensor Voting

With the continuous development of social economy,China’s transportation network has been greatly improved.As an important part of road traffic network,the number of bridges has also been greatly increased.Most of the bridges need to be regularly and professionally inspected in order to find out the diseases in time for maintenance and eliminate the potential safety hazards.Crack is one of the main diseases of bridge components.At present,the detection of cracks is mainly by the way of human eye observation,and then use the relevant measuring instruments to obtain the length and width information of cracks.The disadvantages of this method are obvious: low efficiency,low security,large data error,so it is difficult to meet the growing needs of bridge detection.Crack detection based on image processing is a research hotspot at present.In this thesis,a bridge crack detection algorithm based on tensor voting is designed and implemented.The main work is as follows:1.A bridge crack location detection method based on tensor voting is designed.Firstly,based on the obvious contrast feature of the crack edge,the initial crack points are obtained by extracting the local maximum gradient value;secondly,the symmetrical feature of the crack edge is obtained by extracting the pixel points with the similar gradient value and the opposite direction to obtain the co-occurrence edge of the crack;finally,due to the linear geometric feature of the crack,the tensor voting algorithm with good performance in extracting the set feature of the image is introduced Method to extract the crack position;2.A crack width measurement algorithm based on gray histogram and co-occurrence edge of cracks is designed.The gray histogram equalization is used to highlight the crack boundary in the local crack area,which can effectively solve the problem that the edge of the crack appears virtual due to the exposure of light and camera,making the measurement result of the crack width more error,and improving the accuracy of the width measurement result;3.A fast image mosaic project optimization scheme is designed,which helps to realize the fusion transformation from local crack image to overall crack image,and can effectively improve the speed and reduce the memory consumption;4.A set of bridge crack detection demonstration software is implemented,which can complete image input,crack extraction,width measurement,image display,result output and other functions.In this thesis,the crack detection method based on image processing is studied in depth,and useful work is done from algorithm design and engineering implementation,which provides help for the improvement of automation,informatization and intelligence of bridge detection.