The Research Of Collision Detection Methods In Visual System

In flight simulator, visual system presents a scene for pilots in the flight training process. Of all the information during a flight, 70 percent is given by vision, so, the enrichment, reality and accuracy of visual system will influence the quality of flight simulator and the effect of flight training. Collision detection, as an important part to form a visual system, has a very high station, a precise collision detection will make the visual system much more real and immersed. In the recent years, with the springing up of Virtual Reality and Distributed Interactive Simulation, the research of collision detection is becoming pop gradually, and is being paid more and more emphases.There are many methods for collision detection, but this paper has only done some deep research on two important methods, space-division technology and bounding-box technology. Although the two technologies have different ways to calculate, they both aim to enhance the efficiency of collision detection by reducing the objects participating in the collision detection and the number of basic geometrical elements. Space-division technology divides the whole space into several sub-spaces according to certain algorithm and the denseness of objects in the scene. After dividing, the computer will test only the objects in one sub-space at one time. Bounding-box technology will use a bounding-box instead of the object itself to take part in the test, the box will be slightly bigger but much simpler than the corresponding object. In the two methods, whether the choice of algorithm for space-division is reasonable and whether the reduction of the bounding-box's shape is proper will influence the efficiency of collision detection.Through research, analyses and experiments on the two technologies, on the consideration of the characteristics of the models in the visual system of flight simulator, and analyzing the merits and demerits of several traditional bounding-boxes, the paper proposed a simplified bounding-box. And the new method tests the collision detection by the projection's overlapping test and the intersection's altitude test, in this way, it divides the three-dimensional problem into a two-dimensional problem and a one-dimensional problem, so, it reduces the complexity of the problem.