Research On Bridge Crane Deformation Detection System Based On Close-Range Photogrammetry

Bridge crane is a special equipment which is widely used in industry.The safety inspection of bridge crane is of great significance for preventing accidents and ensuring the safety of people’s lives and property.At present,safety detection is mainly made for the camber,lower flexibility and side-bend of the main girder.There are still some shortcomings in the detection methods,such as potential hazards,complicated operation and high cost.This paper presents a method of bridge crane deformation detection based on close-range photogrammetry.Firstly,according to the design requirements of high precision,low cost and simple operation,we formulate the overall scheme of deformation detection based on close-range photogrammetry.The key technologies of the system are analyzed,including point-projection instrument,image acquisition and image processing.Secondly,we analyze the technical indicators of the point-projection instrument,and design the point-projection instrument based on the requirements of the detection efficiency and the stability.On the basis of meeting the function of projecting white light mark,we add the lase-ranging module and heat dissipation module.According to the requirements of portability and easy disassembly,we design the structure of the Integration platform.Thirdly,we present the image acquisition method of close-range photogrammetry.The error source of the non-measuring camera is analyzed,and the camera is calibrated by two-dimensional DLT algorithm.We analyze the imaging principle of digital camera.RAW file is used as the source file of image processing.According to resolution,information and clarity,the image quality is evaluated.Fourthly,we study the image processing technology of detection system.The method of image pre-processing is designed,including graying,denoising,binarization,and image enhancement.Through two steps we detect the edge of the white light mark.First,coarse positioning at pixel level is carried out by optimized Canny operator,and then accurate positioning at the sub-pixel level is carried out by Zemike operator.Based on the positing principle of the white light mark center,we propose a positing center method based on circle fitting.The error compensation method of the positing is designed to ensure the accuracy of the center coordinate reaches 0.05pixel.Based on the object space coordinate of marking center,the lower edge points of main girder are ascertained,and the lower edge curve is obtained,and then the deformation parameters are calculated.Finally,we develop a software for image data processing and design the verification experiment,including the deformation detection experiment based on white light mark,deformation detection experiment based on artificial mark and deformation detection experiment based on total station.The accuracy of the deformation detection system which we established reaches millimeter-level,which meets the requirements of crane safety detection.Compared with the other two detection methods,it has certain advantages in cost and convenience,and which has a large potential in improving accuracy.The deformation detection system of bridge crane,we designed,lays a theoretical and experimental foundation for the engineering,and also provides a reference for similar deformation detection system.