Research On Three-dimensional Relative Displacement Measurement Technology Of Cracks Based On Computer Visio

Dams are an important national infrastructure in China.It can not only resist flood disaster,but also make full use of water resources to ensure the production needs of industry and agriculture.Dam cracks seriously affect the safety,integrity and durability of the dam structure.Therefore,crack monitoring is of great significance to ensure the safety of dam building structure.With the help of computer vision,image processing and multi-view techniques,the concept of crack equivalent 3D model is proposed and a crack three-way relative displacement detection system is built.In this paper,the identification method of marking corner and the calculation method of depth information recovery are studied,and the relative displacement of cracks in three dimensions in space is measured by using crack images.These studies provide a reference for evaluating the safety of dam cracks.This paper mainly involves the following aspects:(1)The idea of measuring the three-way relative displacement of cracks through image processing is proposed,the three-way relative displacement detection system of cracks is built,and the construction process of the equivalent three-dimensional model of cracks is described.Paste marks on the surface of the concrete structure on both sides of the crack.Identify and calculate the relative three-dimensional coordinates of the feature points on the marks in space and build the equivalent three-dimensional model of cracks.The differences of equivalent three-dimensional models of cracks at different time points were compared and calculated,which were replaced by the variations of the cracks in three dimensional directions in space,so as to realize three-dimensional change detection of cracks in space.(2)Referring to the traditional three-way fracture measurement method.Replace the traditional metal three-way punctuation with a QR code.The image is processed by graying,Gaussian filtering denoising and adaptive threshold segmentation,and the binary graph of mark is obtained initially.The region position of QR code in the image is obtained by morphological processing.Finally,the boundary extraction is used to obtain the corner coordinates of QR code.The repeatability test of corner recognition proves that the method has high repeatability for the recognition of corner coordinates of QR codes in images.(3)A depth information calculation method with range Angle is proposed.The relative coordinates of the identified feature points in the three-dimensional coordinate system are calculated and obtained by imposing realistic scale and threshold constraints on the spatial geometric relationship between the midpoint and its projection in the imaging plane.Experiments show that the introduction of range corner points can effectively improve the utilization rate of image data and the stability of calculation results.In this paper,the effective measurement range of the crack three-way relative displacement detection system is measured under the existing experimental equipment and environmental conditions,the causes of the calculation errors are analyzed,and the corresponding improvement methods are put forward.