Three improvements of the ray tracing algorithm for CSG trees

Abstract Ray tracing is one of the most powerful techniques for image synthesis today. The most beautiful pictures in computer graphics were made by this method. Unfortunately, the computational expense and the computation time are very high. For reduction of computational effort in this paper three improvements for the ray tracing algorithm will be introduced. Two of them are extensions of existing improvements, the third is a new idea: (1) use of enclosures in the object tree; (2) bounding the ray length and (3) dynamical temporary object trees. These improvements are independent and therefore can be used separately or together. They are for ray tracing objects described in a CSG model. The improvements can also be used for secondary rays.