Study On Detection Of Elevated Bridge Basal Crack Based On Fractal Theory

Elevated bridges are of an important part of highway transportations, whose qualitative safety attracts more and more attentions. Therefore, the crack detection of elevated bridge becomes an important content of bridge safety inspection. Because of the characteristics of non-contact, high precision and so on, the crack detection methods of bridge bottom based on digital image processing is attracting more and more scholars to carry out the related research. The crack detection method at the bottom of the elevated bridge is investigated based on the fractal theory in the thesis, and the main research achievements are as follows,(1) The cracks at bottom of the elevated bridges is reviewed and analyzed. Through the introduction of the statistical features of cracks, the distribution features and probability of the cracks at various part of bridge are analyzed. According to the properties and characteristics of the cracks at the bottom, the cracks are classified, the cracking reasons are analyzed, and the basic characteristics of the crack images are describes in this thesis.(2) The fractal theory is introduced and analyzed. The basic concepts including fractal dimension, multi-fractal and so on is introduced, the applications of fractal theory in image processing are analyzed, which are the foundation of the entire study.(3) The image pre-processing techniques are studied for the elevated bridge basal cracks. In consideration of the characteristics of the crack images, the elevated bridge basal crack image is preprocessed to remove the interference noise and enhance the objective in the crack image. The preprocessing includes gray processing, histogram modification, spatial filter, etc.(4) The splitting techniques of the elevated bridge basal crack image is studied based on the fractal theory. Several clustering image segmentation algorithm are introduced and the K-means cluster is improved. In combination with fractal dimensions of the crack images, a clustering image segmentation algorithm is proposed based on fractal theory.Compare with the classical K-means cluster segmentation algorithm, the result suggests that the new algorithm can be more effective in achieving the detection of crack image. The fractal characteristics and geometrical characteristics of the elevated bridge crack are analyzed in this thesis.(5) In order to facilitate the detection and analysis of the elevated bridge basal crack, an image detection software is developed. Through processing the collected crack images, this software can recognize cracks accurately and export the characteristic data of the cracks.Through the present research, the preliminary research of the application of fractal theory has been carried out in the detection of cracks at the bottom of the elevated bridge, which proved the feasibility and validity of the fractal theory and laid the theoretical foundation for the subsequent engineering applications.