projective_texture_mapping

投影矩阵的教程 Introduction Projective texture mapping is a method of texture mapping described by Segal [3] that allows the texture image to be projected onto the scene as if by a slide projector. Figure 1 shows some example screen shots from the projspot demo, available in the NVIDIA OpenGL SDK. Projective texture mapping is useful in a variety of lighting techniques, including shadow mapping [4]. This document provides some background and describes the steps involved in projective texture mapping in OpenGL. Projective texture mapping refers both to the way texture coordinates are assigned to vertices, and the way they are computed during rasterization of primitives. We usually think of texture mapping as “the application of a texture image to a primitive,” and while it certainly is that – there is more math going on than most folks think. If you have ever written your own rasterizer with support for mipmap filtered, perspective-correct, projective texture mapping, you no doubt became aware of the many subtle issues involved. We will begin by discussing the way that texture coordinates are computed during rasterization, and then we will discuss methods for assigning the texture coordinates to the vertices. We do not discuss filtering here, but there is a paper on Anisotropic Filtering at the NVIDIA developer web site that provides a good introduction to that topic. Figure 1.