Abstract:A new method, is presented which is calculated to use only contour line from a single image in order to compute the dimensions of object and its pose from a single image of perspective. From the perspective view, the perspective element, basic unit, and combined unit are researched, and the correlative arithmetic is presented. The pivotal arithmetic about the fragmentary visual information is also researched. At last, handle that information through classification and synthesize, construct the body of that picture in the 3D scene.