The Calibration Of Fringe Projection 3D Measurement System And The Improved Calibration Algorithm

Machine vision is a comprehensive technology of optics, electronics, computers and mechanics which has been widely used in various areas of industrial measurement. Optical 3D measuring is a typical application of machine vision. Fringe projection 3D measurement is the representative technique of optical 3D measuring. It has many advantages such as high degree of automation, fast access to data and high accuracy of measurement and has been greatly promoted the development of modern manufacturing. Now, the development trend of optical 3D measuring obtains great attention from various research institutes and manufacturing enterprises.CCD camera is the visual part of the fringe projection 3D measurement. The calibration of the CCD camera and the acquisition of CCD camera intrinsic and external parameters are the key technique of the fringe projection 3D measurement. In this paper the CCD camera of the visual imaging system is calibrated based on the fringe projection 3D designed by Xi'an Technological University Institute of Precision Optical Measurement.The results of this study are given as follows.1) The projector module, CCD camera module and digital process module of the fringe projection 3D measuring system were assembled based on the research of former researchers. The principle of the fringe projection 3D measuring was researched. The imaging mathematical model of CCD camera were established based on the transformation among the image coordinate system, world coordinate system, camera coordinate system and the distortion error between the ideal image and the real image. The calculation method of the CCD camera imaging model was studied. The nonlinear Levenberg-Marquardt optimization algorithm used in the solution of camera model was deducted.2) A PVC lattice array target plate was printed according to the characteristics of the CCD camera. The edge extraction of the image obtained by CCD camera was implemented using the center of the spot as feature point. The image coordinate of the center of the spot was extracted accurately using coordinate transformation and mathematical statistics. The feature point extraction algorithm was verified according to the experimental results.3) The Tsai calibration algorithm was deducted and modified attributed to the deficiency of this algorithm. The imaging center of the CCD camera and scale factor of the CCD image plane were obtained based on the coordinate of the lattice array target plate and their 3D space coordinate. Calibration experiments were implemented using Tsai algorithm and the improved algorithm to obtain intrinsic parameters and external parameters of the CCD camera. The calibration error of these two algorithm were calculated using the intrinsic parameters and external parameters of the CCD camera, the CCD camera imaging model and the coordinates of the lattice array target plate. The results showed that the calibration error of the Tsai algorithm wasδ1= 0.705mm andδz1=0.651mm mm, the calibration error of the improved algorithm wasδ2=0.148mm andδz2= 0.093mm and the improved calibration algorithm was correct and feasible.