Nowadays, most of image-guided weapons adopt 2D images. Lacking of a goal's deepness information, the weapons can't identify and track the goal effectively. On the contrary, deepness images, attained by a 3D laser vision sensor, contain the goal's deepness, but enormous 3D information exceeds the ability of a computer's timely solving images. Based on the idea of "replacing time with space", the technique of Level of Detail (i.e.LOD) supplies the difficult problem a middle course. Based on some definite rules, LOD furthest simplifies the 3D mesh models of a goal, making a computer display images and graphics as fast and efficiently as possible, without influencing the quality of images and graphics, so the technique can greatly improve the ability of image-guided weapons' identifying and tracking a goal, efficient combing a goal's deepness images with it's 3D mesh models and constructing the LOD of a goal.The thesis makes Garland algorithm not only can simplify models having sudden vertexes, but also can gain a preferable result, by improving on the vertex error of the Garland algorithm; makes progressive mesh algorithm efficiently fix on the LOD of a goal through the distance from a view port to a goal, by introducing mesh density; establishes the LOD of a terrain by programming with the aid of both VC6.0 and OpenGL; researches a goal's 3D reconstruction and LOD establishment, based on the goal's deepness images; accomplishes a experiment measuring the distance from a goal to a 3D laser distance measurer. |