| With the rapid development of China’s economy,steel structure architectures gradually play an important role in various infrastructure projects.In terms of bridges,traditional wooden bridges and concrete bridges have been substituted by steel bridges.However,the defects of engineering construction and the use of a long time are likely to cause chronic diseases on the bridge surface.If these defects are not monitored and treated in time,they may affect the bridge’s serviceable life and even threaten the lives of the people,causing immeasurable losses.In order to avoid this situation,research on the daily maintenance and disease detection of steel structure bridges is constantly deepening.In recent years,with the popularization of computer technology and artificial intelligence,more and more intelligent detection has been applied to the research of bridge diseases.At the same time,people have higher requirements for the efficiency and accuracy of the detection.Aiming to detect the common rust,coating detachment and crack diseases on the bridge surface,this paper proposes an algorithm for identifying the surface diseases of steel structure bridges based on image processing technology,and develops a disease detection software system based on the MFC framework and OpenCV library.The details are listed as follows:(1)For the detection of rust disease,a threshold segmentation algorithm based on HSV chromaticity space is proposed.Convert the image to HSV chromaticity space,then analyze the component histogram of the target area,select an appropriate threshold for disease segmentation,finally mark the rusted part.After testing,the detection and recognition rate of this algorithm reaches 93%,and the average detection time is 0.35s.(2)For the detection of coating shedding diseases,an algorithm based on edge detection and threshold segmentation is proposed.After the image is converted to gray space and filtered,the Canny operator is used to detect the edge of the dropped part.Then the closed operation is used to connect the disconnected area.Finally,the target part is marked based on the connected domain,and the diseased area of the image and the actual area are calculated at the same time.After testing,the detection and recognition rate of the algorithm reaches 96%,and the average detection time is 0.37s.(3)For the detection of cracks,a segmentation algorithm based on the combination of single-scale Retinex and the Maximum Between-class Variance Method is proposed.After the image is filtered and single-scale Retinex enhanced,it not only highlights the traces of the cracks,but also retains the edge information.Finally,the binary image is segmented and the average width of the cracks is calculated.After testing,the detection and recognition rate of this algorithm reaches 90%,and the average detection time is 0.25s.(4)Based on the VS and OpenCV libraries,a software system for identifying the surface disease of steel structure bridges was developed.The rust,coating peeling and crack disease areas of the steel structure bridge were detected and described in text.At the same time,the detailed parameters of the disease were obtained qualitatively.Experiments indicate that the algorithm in this paper has a good effect on the recognition and feature extraction of the three common diseases of steel structure bridges.At the same time,specific analysis is carried out according to the actual situation.The performance of the software is good and can meet the needs of actual engineering. |
