補償光子在漸進性光子映射中之應用

Abstract

漸進性光子映射演算法(Progressive Photon Mapping)由傳統全域照明技術之一的光子映射演算法(Photon Mapping)所改進而來，改變了傳統光子映射演算法的架構，使得光子只需儲存擊點(hit point)的資訊。 本篇論文根據光子映射演算法提出一個易實作且直覺的補償光子(Compensation Photon)應用在漸進性光子映射演算法。在互動式的場景中，當場景中物體位置變化時藉由補償光子可以不用每個畫面都須重新設置擊點的位置，並且重新潑灑所有光子。補償光子將移動的物體影響至場景中的能量抵銷，包含了物體移動範圍內的能量以及物體影響至場景的能量，使得下一個畫面中不用重新潑灑所有光子，只需潑灑物體移動的範圍以及影響到的區域所需的光子即可。不用計算與前一個畫面相同能量的區域，因此可以使得漸進性光子映射演算法在互動式場景中達到加速的效果

Progressive photon mapping is a simple and robust progressive global illumination algorithm based on photon mapping. It changes traditional photon mapping's framework , such that we only need to store the hit point information of each photon.

In this thesis, we introduce a compensation photon to the progressive photon mapping, which is simple and easy to implement. In interactive environments, when an object moves, via the compensation photon, there is no need to set all hit point and re-emit all the photons. Using compensation photon to offset the power in the scene coming from the moving object, we can find the areas there have the same power as the previous frame. There is no need to re-emit all the photons to render the entire frame. We just emit to the areas where compensation photon has offset. Therefore we save the time which is taken by emitting unnecessary photons to render the scene.