The Collision Detection Technology Research And Application Of Virtual Surgery

As an important application of virtual reality in modern medicine, virtual surgery,also named surgery simulation, has extensive application prospect in pre-operativeplanning, training for operating skill of surgeon, surgical navigation, design of artificialprosthesis, and so on. With multiple key techniques such as computer graphics, imageprocessing, pattern recognition, and automatic control, and biomedicine, combined invirtual surgery, virtual operating environment in computer is constructed, wherein boththe surgical instruments and the organs and tissues are modeled precisely, so thatphysical phenomena like deformation, cutting and blooding can be simulated vividly.Surgery simulation systems based on virtual reality techniques aim to afford effectivesurgery training platform, and to assist real surgery performance as well.One of the crucial problems in research and development of virtual surgery systemis collision detection. Fast and accurate collision detection is the important premise ofsimulating tissue deformation, cutting and seaming, have create influence onimproving the reality of virtual surgery environment and the immersion of end users.On the other side, with the rapid development of computer hardware, the complexityand real-time property of the virtual surgery demand collision detection techniqueswith higher performance. However, most of the current collision detection algorithmsare proposed for rigid objects, while research on collision detection for flexible objectsis still immature. Collision in virtual surgery occurs mainly between surgicalinstruments and organs or tissues. The tissues and organs are usually modeled asflexible objects of large scale, while the instruments are always modeled as simplerigid models. How to detect quickly and precisely the collision between the two typesof objects has become the bottle-neck of the whole virtual reality system.Collision detection is an underlying problem in the fields of computer graphics,robot motion planning and virtual reality. Its primary mission is to detect the collisionor intersect between two or multiple objects, and to compute the area of intersection orthe depth of penetration. As effective techniques to resolve the intrinsic timecomplexity in collision detection, space decomposition and hierarchical bounding volume based techniques are generally applied, aiming to speed up collision detectionby excluding as soon as possible those primitive pairs that are clearly impossible tointersect.In order to resolve the problem of collision detection in virtual surgery, relatedtechniques are studied, and algorithms based on bounding volume hierarchy and thosebased on space decomposition are introduced and analyzed. According to thecharacteristics of virtual surgery systems, a fast collision detection method based onBVH are proposed, which improves the original algorithm in the three aspects,including construction of the hierarchical bounding volume tree, traversal of the tree,and accurate intersection detection. Simulation experiments show that the improvedalgorithm is capable of detecting the collision between flexible objects and rigidobjects effectively, meeting the virtual surgery’s demands efficiently.Moreover, an improved collision detection algorithm based on AABB is proposedto guarantee the real-time detection. The space decomposition method is utilized todecompose the object space to find out the grids that are possible to collide, for whichthe respective hierarchical bounding volume trees using axis-aligned bounding box(AABB) are constructed further. By traversing these hierarchical bounding volumetrees, the potential primitive pairs that are possible to collide are submitted to the phaseof precise intersection test. Results of experiments and practical application show thatthe hybrid algorithm based on space decomposition and hierarchical bounding volumetree can improve the efficiency of collision detection obviously, and can tackle theproblem of collision detection in large-scale and complicated virtual surgery system.