Intelligent Detection Of Pier Structure Surface Disease

In recent years,China has completed a large number of bridge pillars,substation frame columns,large chimneys and other pier structure infrastructure construction,and most of these facilities are made of reinforced concrete materials.With the extension of the service time of the pier structure and the influence of environmental factors,it often leads to certain cracks,exposed bars and other diseases.In severe cases,it will cause structural collapse and seriously damage people ’s property and life safety.Therefore,it is very important to maintain and supervise the pier structure of reinforced concrete materials.The traditional manual detection method requires a lot of manpower for the detection of concrete pier structures,and the detection period is long,the detection cost is high,and the efficiency is low.Therefore,the built climbing robot system is used to obtain the image of pier structure disease,and an intelligent detection method for pier structure surface disease is proposed.In this paper,the surface image of the pier collected by China Railway Construction Bridge Engineering Group is taken as the research object.By studying the image deduplication algorithm,cylindrical back projection technology,disease recognition algorithm and image mosaic technology,the intelligent detection of the surface disease of the pier structure is designed and realized,which is of great significance to the health monitoring and maintenance of the pier structure.This paper first introduces the hardware part of the robot system,and designs the interface between the robot control system and the camera host computer,realizes the control of the robot system movement and the camera,and stably obtains the apparent image of the pier structure.Secondly,aiming at the problem that the overlapping area of adjacent images is too large in the annular shooting path of the robot,which leads to the redundancy of image data,a perceptual hash deduplication algorithm based on superpixel segmentation is introduced to screen out redundant images,which effectively improves the efficiency of surface disease detection of pier structure.The image denoising and enhancement algorithm is used to complete the preprocessing of the deduplication image,thus reducing the influence of environmental factors such as illumination.Then,aiming at the problem of image distortion caused by the surface characteristics of pier structure,a cylindrical expansion method based on two-dimensional surface fitting is proposed to realize the distortion correction of image.The experimental results show that the measurement error of this method can be controlled within 6%.Then,the concrete disease identification method based on Meta-QNN is used to complete the network architecture design,so as to identify the concrete disease information efficiently and accurately,and the Unet network is used to segment the identified crack image,which lays a foundation for the measurement and analysis of subsequent cracks.Finally,the image stitching technology based on ORB feature extraction is used to realize the stitching of disease images,and the Zhang Suen skeleton refinement method and the shortest distance method are used to complete the marking and quantification of crack disease information.Through experimental verification,the method studied in this paper can effectively complete the intelligent detection of pier structure surface diseases,improve the detection efficiency,and provide data support for later structural health monitoring and maintenance.