Calculation Method Of Crack Width Of Concrete Bridge Based On Image Technology

At present,there are more than 1,000,000 bridges in service in China,and bridge inspection to ensure the healthy operation of bridges is becoming more and more important.However,due to the shortcomings of traditional manual inspection such as low efficiency and poor safety,it can no longer meet the needs of bridge inspection,and Chinese and foreign scholars have gradually shifted the research focus of bridge inspection to intelligent detection based on image technology.In bridge inspection,crack width is one of the important evaluation criteria for the healthy operation of bridges,and China also puts forward clear requirements for crack width in bridge specifications,so it is a high-efficiency,low-cost,high-precision detection method to calculate parameters such as bridge crack width combined with image technology.In this thesis,the acquisition platform is first used to collect the bridge crack image information,and the collected image information is transmitted back to the ground control center in real time,and then a set of concrete bridge crack width calculation method is proposed based on image technology,and the algorithm is implemented in Python language,through which the returned bridge crack image information is calculated and finally the crack width and other parameters are obtained.The research focuses include the preprocessing of crack image information,the extraction and calculation of key information of cracks,and the specific research content is as follows:(1)A set of image preprocessing methods were designed based on the apparent features of bridge crack image information,which laid the foundation for the follow-up research in this thesis.The preprocessing method combines the existing algorithms to optimize the image enhancement,image smoothing and image segmentation algorithms.Firstly,for image enhancement processing,an optimized local histogram equalization algorithm is proposed,and the key information in the image using the algorithm is more prominent.Then,for image smoothing,a bilateral mean-to-average filter is proposed,and the SSIM and PSNR evaluations of the output images using this filter are high,and the algorithm efficiency is increased by 72.5%.Finally,for image segmentation,an adaptive threshold segmentation algorithm based on normal distribution is proposed,which can be used to maximize the retention of key information such as crack area while ensuring computational efficiency.(2)A crack single-pixel skeleton curve extraction method is proposed,and the algorithm is compiled using the Spyder compiler.In this method,the coordinate system of the preprocessed crack image information is established in the global range,the height to width ratio of the connected area is calculated to determine the crack trend,and then the crack skeleton curve is obtained by calculating row by row(column)in the connected area,and finally the number of connected domains and the result validity coefficient are used to evaluate the extracted results.The results show that the skeleton curve extracted by this algorithm performs well in terms of continuity and extraction accuracy,and meets the requirements of bridge crack width calculation.(3)A crack classifier is designed based on the extracted crack skeleton.Firstly,the linear elimination test of the crack skeleton is performed to remove the marked pseudocracks,and then the mesh cracks are identified by calculating the histogram of the crack skeleton curve in the X and Y directions,and finally the linear cracks are classified based on the inclination angle of the rotating outer rectangle.The test results show that the recognition rate of the classifier is 92.8% and the accuracy is 78.1%,which is suitable for the initial screening of image information.(4)A crack width calculation method is proposed and implemented in Python language.In this method,the upper and lower edge curves and fracture skeleton curves of cracks are mathematically modeled by three-point interpolation,and the Euclidean distance between the normal equation of the middle point of the skeleton curve and the intersection of the upper and lower edge curves is calculated as the crack width at the pixel.After encapsulating the algorithm,the cracks at different positions of the bridge are verified,and the results show that the algorithm has good robustness during operation,the average return time is about 9.47 s,and the accuracy of the result is 98.53%.