Study On Image-based Bridge Crack Detection And Analysis

In recent years,driven by the National "Thirteenth Five-Year Plan",the highway transportation infrastructure construction in China has made remarkable achievements.According to the data of the Ministry of Transport,by the end of 2018,there were832,500 bridges across the country,of which more than 88.41% were small and medium spans bridges.Furthermore,they were mostly reinforced concrete bridges with simple structures which were easy to be constructed.Therefore,the normal operation of transportation is of great significance.However,the traditional bridge structure crack detection mainly relies on manual detection by professional technicians,its efficiency and accuracy need to be improved,resulting in the following problems: the detection data is unreliable and inaccurate;the detection data cannot be structured and stored;Bridge disease cannot be accurately positioned in three dimensions;bridge disease quantitative analysis is not accurate.Therefore,the new visual inspection method to complete the bridge crack detection and evaluation will be of great significance.Based on digital image processing technology and photogrammetry principles,this paper summarizes and compares the processing algorithms of bridge crack images,and proposes improved methods to achieve crack image graying,filtering,denoising and edge detection.Then,an algorithm for extracting the width and length of the crack is proposed,and the geometric parameters of the crack are obtained,and the crack morphology is classified.Then,based on the photogrammetry theory,a threedimensional model of the experimental beam was established,and the spatial parameters of the crack were extracted.Finally,based on the above-mentioned crack parameters,the relationship between each parameter and the bearing capacity of the beam body was analyzed.The main research contents are as follows:1.Pretreatment of cracks,edge detection and parameter extraction.The preprocessed crack images of the collected concrete bridges are pre-processed,and the gray image,filtering and denoising of the crack images are realized,so that the cracks are enhanced and the noise is weakened.Then the improved spatial domain detection algorithm and morphology detection algorithm are combined to detect and mark the crack edges.Finally,based on the crack characteristics,a refined circumscribed rectangle method is proposed to calculate the crack length,and the intersection point method is used to calculate the crack width.The geometric parameters of the crack are extracted successfully.2.Spatial positioning of cracks.In order to realize the precise positioning of the cracks in the three-dimensional space of the bridge structure,the photogrammetry technology theory is elaborated in detail,the photogrammetry modeling process is proposed,and the accuracy of the 3D model reconstruction of the beam body is verified through experiments.The image establishes a three-dimensional model of the real scene,and the spatial location parameters of the crack are extracted by writing a plug-in.3.Quantitative analysis of crack parameters.Taking rectangular beams with rectangular section as the object,on the basis of summarizing and analyzing the whole process of concrete beams,the calculation method of crack height and width is proposed.Based on the experimental data,the relationship between the maximum width of the crack and the maximum height of the beam and the bending capacity of the beam is analyzed and established.In this paper,using rectangular reinforced concrete beams as the experimental object,the feasibility of crack detection using image processing technology and photogrammetry principle is discussed,and the relationship between crack parameters and the bearing capacity of the beam is established based on the detection data.The evaluation method provides a complete idea and has a high application prospect.