Stereo Vision Measurement Method Based On Calibration Of Structural Straight Line Feature

Convenient and efficient camera calibration is the basis of accurate and efficient visual measurement.In traditional calibration methods,prefabricated references are used for calibration.The reference needs to be fully distributed in the field of view,which limits the use of visual measurement methods when the field of view is large.To solve the problem,camera calibration algorithm based on the structural straight line features is proposed.Aiming at the measurement accuracy of stereo vision,the selection of baseline distance of stereo vision system and 3d reconstruction of spatial points are studied.The main research content includes the following three parts:(1)A camera calibration method based on linear feature of structure is proposed.In the calibration method in this paper,the projection invariance of the line is used to construct the optimization function to solve the camera distortion coefficient firstly.A method of vanishing point fitting is proposed.The constraint equation is constructed with vanishing point to solve the camera's internal parameter matrix.The simulation test is carried out and the calibration results are compared with the calibration method of Zhang Zhengyou calibration method.The results show that the proposed algorithm has a high precision.(2)By introducing the precision coefficient to precision model of binocular stereo vision measurement system,the influence of the camera orientation on the accuracy of image matching of the circular target image is discussed.The precision coefficient is introduced to describe the influence of the angle of the optical axis on the accuracy of the positioning of the image point.Simulation test and biaxial displacement test were carried out to verify the applicability of the accuracy coefficient formula.The results show that the accuracy coefficient formula can reflect the changing trend of the positioning accuracy of image points.The relationship between the baseline distance of stereo vision system and the measurement accuracy is discussed considering the effect of camera optical axis angle change on the positioning accuracy of image points.The ratio of the baseline distance to the object to be measured should be between 0.6 and 1.3.A 3d reconstruction method of spatial points based on maximum likelihood estimation was proposed with a method to solve 3d coordinates of spatial points directly without constructing optimization function iteration.The results show that the proposed method has good accuracy and stability.(3)The shaking table tests of masonry building was completed.The test results showed that the displacement obtained by stereo vision measurement method was consistent with that of LVDT,and the first-order main frequency of acceleration power spectrum obtained by stereo vision measurement method was almost the same as that obtained by accelerometer,which verified the validity and accuracy of the 3d reconstruction method based on maximum likelihood estimation.Complete T block concrete walls subjected to low-cyclic loading test.An orthogonal vanishing point is used to establish a structural reference surface to solve the external parameters of the camera.The test results show that the error of building the reference plane based on vanishing point is basically the same as that of establishing the reference plane based on Zhang Zhengyou calibration method,which shows the effectiveness of establishing the reference plane based on vanishing point.