Studies On Position And Pose Measurement System For Moving Objects

Line geometries are the most important parameters that characterize the safety status of railway infrastructure. Therefore, high-precision measurement of these parameters is key to ensure railway safety, and has an important bearing on the national economy and people's lives. At present, the dynamic vehicular measurement has become the main means of monitoring the safety of railway and rail transit. However, in the process of measurement, as a result of vibration, testing equipment will have uncertainty of six degrees of freedom, seriously affecting the accuracy of measurements. So it's crucial to detect the position and pose of the moving vehicle to compensate measurement errors. In this thesis, based on studies on related technology at home and abroad, and on account of the characteristics of train detection equipment, we designed fixed-point measurement system of the position and pose of moving vehicle.In this thesis, the design was based on the principle of binocular stereo vision and Bursa coordinate transformation model. The proposed system had a simple structure, high precision, and was easy to install. In order to meet the system requirements of high-speed measurement, the thesis utilized FPGA to complete the functions of image acquisition, and made use of PC in processing functions to complete post-processing of images. In the image processing, we used morphological theory to remove noise of images and achieve precise positioning of the feature points. Then we used gravity center method to obtain the sub-pixel image coordinates of the focus points. In this thesis, we used neural network for camera calibration. In accordance with the principle of three-dimensional photography, three-dimensional coordinates of feature points could be obtained by using image coordinates of the points. Finally, according to a coordinate transformation model we could get six degrees of freedom of vehicles. On the basis of the existing experimental conditions and according to the system design, we established the measurement platform and did the related experiments. Experimental data showed that the accuracy of system met the accuracy requirements.