Research On Collision Detection And Response Method For Maize Canopy Reconstruction

This paper presents a suitful collision detection method for checking out theintersections and penetrations that prone to three-dimensional maize canopy reconstructed,and puts forward a typical simplified response method to modifying the collisions based onthe geometric characteristics and dynamic features of maize. Adopting the above methods,one collision detection and response system for maize model is designed to find out theintersections among virtual canopy and to implement reasonable response.The following represents our main study:(1) Based on the morphological character of maize leaves, axis aligned bounding box(AABB) and oriented bounding box (OBB) are carried on maize leaves model in separatelyto achieve quickly bounding boxes culling, while the process of OBBs' construction andtraverse get some simplified. Later, one collision detection method based on AABB-OBB isproposed for maize canopy. Among them, AABB is used to split the whole plant into severalareas and to implement overlap test firstly. Then OBB tree is constructed to reduce thetraverse lists.(2) Due to the collisions always involving the edge of leaves, one optimization forprimitive test is proposed. Under current data structure, the triangles along the edge of maizeleaves are given priority to implementing primitive tests, by which the untouched pairs can beculled quickly and the test efficiency can get some improvement.(3) According to the dynamic characteristics of maize leaves, one simplified responsemethod for corn leaves is put forward based on current geometric mesh. At first, leaves aredivided into three blocks by which the active object is determined, while the responsedirection is calculated based on the information of edge-cross and vein. Then, according tosome simple constraints proposed to maintaining the geometric detail, the vein feature pointsas the skeleton are computed by variable step size followed by others points. After severaltraverse, the intersections among pairs of maize leaves can be solved as much as possible. Forthe complex intersections among maize canopy, the strategies with the estimation of responsespace and the sort of consumption are adopted to eliminate the collisions, enhancing therealism of the canopy. (4) Integrating the above methods, one collision detection and response system for maizecanopy is developed, in which compute unified device architecture (CUDA) is introduced toaccelerate the computation. Using this system, it is easy to check out the collisions in virtualcanopy, and to carry out sounds responses as much as possible.