Research On Key Techniques For Large Structure Displacement Measurement Based On Photogrammetry

Monitoring the static and dynamic displacement of large structures is of great significance to the health assessment and maintenance of the structure.Due to the characteristics of non-contacting,high precision,full-field,and real-time measurement,photogrammetry has obvious advantages in displacement measurement for large structures.However,the complex environmental condition outside the laboratory is a great challenge to the application and popularization of this method.This dissertation addresses the key techniques of video photogrammetry in the outdoor application,which includes stereo vision calibration,camera shaking correction,and reference coordinate system establishment.In order to verify the effectiveness of the proposed methods,a three-dimensional displacement measurement experiment has been carried out on the blades of a 2 MW wind turbine.The major contributions of this dissertation are listed as follows.(1)The camera-IMU fusion algorithm is introduced to the stereo camera calibration for the extrinsic parameters,which reduces the influences of qualities of scene feature points on calibration based on epipolar geometry.Considering the error in the yaw angle measured by the IMU,an angle differential method is introduced to improve its measurement accuracy.The rotation matrix between the two cameras can be determined by the posture information provided by the IMU,and then the translation vector can be solved according to the finite feature points and the known rotation matrix and intrinsic parameters.The experimental results show that the accuracy of the proposed method is comparable to that resulted from the well-known Zhang's method,and it can be applied to measurement with a large field of view(FOV).(2)Traditional calibration methods are difficult to apply in the field of large FOV measurement in case that the feature points are limited or unevenly distributed.In this dissertation,a camera calibration method utilizing the UAV has been proposed.Based on the principle that the feature points cove the entire FOV and are evenly distributed,a UAV carrying markers is controlled to fly in the FOV while the stereo vision system capture images simultaneously.The extrinsic parameters of stereo vision are solved according to the geometric constraints of the image points,and then the sparse beam adjustment method is used to optimize the parameters.The effectiveness of the proposed method is verified through experiments,and the influences of the number and distribution of reference points on the calibration accuracy are discussed.(3)Three major problems in measuring deformations of large-scale structures using the stereo-DIC method are analyzed.First,an extrinsic parameter optimization method is proposed by introducing the DIC method into the SIFT feature point matching algorithm.Second,since the reference frame is usually established on the left camera,the measured displacement components do not agree with the measured structure.A method of reconstructing the reference frame using the geometric information of the measured structure is proposed.Third,in order to eliminate errors caused by camera shaking in field measurement,the fixed points in the scene are selected in solving the motion model parameters before and after the camera movement to correct the image coordinates of the measuring points.Finally,the effectiveness of the above three methods are verified by taking the building as the measurement object.(4)Taking a 2MW wind turbine as the measurement object,since there are almost no available feature points on the blade surface,the feature points in the FOV are increased by pasting some speckle patterns on the blade surface.Three measuring points along the radial direction of the blade are selected,and the pixel displacements of the measuring points are retrieved with the matching strategy by reversely rotating the reference subsets.Four methods were used to calibrate the parameters of the measuring system.The optimal set of parameters was selected to calculate the 3D displacement of the wind turbine blades,and their frequency spectra are analyzed.The experimental results show that the method proposed in this dissertation can accurately measure the 3D displacement of the rotating blade,which provides a potential method for the fault diagnosis of full-scale wind turbines in operation.