Study On The Key Technique Of Single Camera 3D Coordinate Vision Measurement System Using A Light Pen

The portable 3D coordinate vision measurement system based on the probe imaging coordinate measuring technology presents many characteristics such as fast and flexible construction, simple operation, large measuring range, high accuracy, capable of measuring on spot and not affected by the difference of measured surface. So the system has wide prospective applications in many fields such as the measurement for large dimension on spot, the rapid 3D data capture from CAD model, costume design and many others. After a comprehensive review on the indoors and overseas current situation and developing tendency of portable 3D coordinate vision measurement system, a single camera 3D coordinate vision measurement system using a light pen has been designed and set up. In the dissertation, two aspects, the system mathematic model and its solution's uniqueness and the improvement of the system's measuring stability, are the emphases. The experimental result shows that this system has achieved a measuring stability precision of 0.365mm, 0.05mm and 0.072mm in the direction of Z, Y and X axis respectively in one meter measuring range.The dissertation covers the following aspects concerning the main work and creative research in the project.1. The system's composing and measuring principle has been discussed. The system mainly consists of a light pen, a CCD camera, a set of special calibration accessories and a image card built in computer installed measuring software. The light pen is a hand-held probe on which at least three point light sources and a spherical touch stylus are mounted in one line. The CCD camera captures the image of the point light sources and the 3D coordinates of measured point touched by the stylus are computed with the measuring software.2. Based on the principle of Perspective-Three-Point Problem (P3P), the system's mathematic model, Perspective-Collinear-Three-Point Problem (P3CP), has been proposed and established, and the uniqueness of P3CP's solution has been verified. The analytical equations of the measured point's 3D coordinates have been derived. Meanwhile, it has been determined that the number of point light sources is not less than three.3. The comparative experiment has been conducted on a CMM, and the experimental result has verified the correctness of the system's model. The repeating measurement results show that the actual constructed system's maximum instability measuring error is about 0.077mm, 0.01mm and 1.009mm in the direction of X, Y and Z axis respectively. So it is stressed that some treasures must be adopted to improve the measuring stability in the direction of Z axis. According to the actual measuring system and environment, some main factors affecting the measuring stability have been indicated, and the content of chapter3, 4, 5 in this dissertation is about how to improve the system's measuring stability.4. The dimension, shape and position accuracy of the light pen's structure, which is one of the main factors affecting the system's measuring stability and accuracy, has been designed. Through the theoretical analysis and practical experiment, the determining principle of the light pen's positioning dimension has been given. The tolerance value of these positioning dimensions has been assessed 0.001mm with the actual experiment. Combining the precision requirement andmachining cost, the concerning shape and position tolerance has been given.5. Because of the CCD camera imaging system's distortion error and the light pen's manufacturing and assembling error, the system's measuring stability with the P3CP model is not satisfied. So the actual image points are fitted in the image plane and the calculated measured control points are fitted with corresponding weighed values in the camera coordinate system with the Least Square fitting method of a line. After the above fitting processing, the measuring result has demonstrated that the system's measuring stability precision has reached 0.365mm, 0.05mm, 0.072mm in the direction of Z, Y, X axis respectively.6. The image processing method has been discussed. The method, which is for processing the original image of point light sources and picking up the image contour and its center's coordinate of the measured feature points in the CCD image plane, is another main factor affecting the measuring stability and accuracy. A new algorithm for recognizing and extracting the contour of elliptical target feature points, which is based on evaluation for elliptical profile error, has been designed creatively, and the actual image processing result has proved that this algorithm is stable and reliable.7. The improvement on the solving algorithm for the measured point's 3D coordinate and the redundant design of light pen structure have been discussed. In order to improve the system's measuring accuracy, more than three collinear point light sources are mounted on the light pen to compose the redundant system. Thus, the measured point's 3D coordinates can be calculated with not the collinear condition but the concerning numerical algorithm for Least Square Problem.8. The principle of the light pen structure parameter self-calibration and its algorithm have been discussed, and the proper mathematic model of self-calibration has been established. To realize the self-calibration, three measured points are necessary when the number of point light sources is 5. The computer simulation has proved that the self-calibration algorithm is non-sensitive to the initial value and has a fast convergence speed and a high accuracy.Through theoretical analysis, experiments and computer simulation, the single camera 3D coordinates vision measurement system using a light pen has been studied with emphasis on the system's composing and measuring principle, the system mathematic model and its solution's uniqueness, the system's measuring stability, the design of the light pen structure parameter's precision, the image points'and the measured control points'Least Square fitting of a line, the system's image processing, the improvement on the solving algorithm for the measured point's 3D coordinate and the redundant design of light pen structure, the light pen structure parameter's self-calibration and its algorithm and so on. In the dissertation, some creative progresses are made and some new methods and conceptions are proposed. The work lays a foundation for further development of the practical portable 3D coordinate vision measuring system with higher measuring stability.