Precise Calibration Of The Vr-based Ccd Study

Using of radar,laser on a moving objtect is the main technical of objects positioning, and there are specific requirements in the equipment, this scene and environment. This research is based on the combination of virtual reality technology and image processing: after reconstruction the object by VR,use CCD shoots the object after it moves, and then get its precise three-dimensional information, complete object localization.This paper have introduced and analysised Tsai and zhang calibration and their advantages and disadvantages, and an improved two-step calibration method has been given based on this,extract the sub-pixel- corner based Harris detector, establish a suitable nonlinear CCD model takes distortion in to account ,use improved LM optimization algorithm to obtain accurate parameters of the CCD. After then,shootting image pairs using this CCD,and matching them. Because the images can be divided into short and long baseline,for the purpose of getting higher accurate matching points, this paper presents a new matching algorithm: for the former, we use the gray-related laws to take the initial matches, and then use Lmeds and RANSAC which are robust algorithm to remove the wrong match points, get more match points using the guidance of matrix F.For the latter, using SIFT algorithm. And then using SFM algorithm to complete 3D reconstruction,display the reconstruction object by VRML browser.Present a new objects positioning method Innovatively using of the combination of VR and image processing,according to the reconstruction object, the CCD should be fixed in the location of the world coordinates, and then use single-CCD shoots the objects after it moves, positioning of objects precisely useing the way of VR.Using this method of objects positioning in this paper is relatively simple, it have a very broad prospects in video surveillance, 3D visual simulation, virtual scene, and other fields.