Adaptive mesh generation for progressive radiosity: A ray‐tracing based algorithm.

Abstract

The radiosity method is one of the most popular rendering algorithms. It allows to simulate interreflections of light accurately between surfaces as energy transfers are well designed. However, this algorithm and its derivatives need to break scenes into a relatively large number of small elements to approximate the illumination function. Even with a very large number of elements, not all the illumination effects can be simulated. In fact, there are always artefacts when modelling sharp shadows, besides shadows falling between mesh vertices can be missed entirely. To reduce the computational cost of such methods and to increase the accuracy of the radiosity solution, adaptive mesh generation is well suited. In this paper, we present a ray‐tracing based algorithm for adaptive mesh generation which resolves all the illumination problems without lengthening computation time too much. This method allows a small number of initial elements and increases element density in critical locations while solving the illumination problem.

Parallel Progressive Radiosity with Adaptive Meshing

“…That is actually not true for each patch with a large area. To achieve a good accuracy, adaptive meshing was integrated with radiosity in [5,8,10,11,16,24]. In this approach, each patch derived from the initial subdivision (meshing) may be further refined during the course of the computation if the accuracy of each patch radiosity does not meet the standard.…”

“…The intermediate result after a few iterations can be used for display and a user can revise the design based on this intermediate solution without waiting for the entire computation to end. A further improvement by which to increase the image quality of radiosity in such an iterative algorithm is to incorporate adaptive subdivision [5,16].…”