三维表面纹理高度重建中的光线不均校正

因三维表面纹理能比二维纹理更好地表现物体的纹理信息,而且随场景光照及视角的变化而变化,所以被广泛用于虚拟现实以及计算机游戏等技术之中。Photometric Stereo作为一种有效的获取三维表面纹理信息的技术而被人们所广泛关注。均匀的光照条件是Photometric Stereo捕获和重建三维表面纹理成功的关键条件。在现实应用中,不均匀光照会导致三维表面纹理在捕获和重建过程中发生失真和畸变。针对这种失真和畸变进行了研究,并提出了一种解决此类问题的方法。实验结果表明,该方法简单可行,有效。

Texture Classification of 3D Surface Textures Via Directional Quincunx Lifting

This thesis presents a new approach to classify 3D surface textures by using lifting transform with quincunx subsampling. Feature vectors are generated from eight different lifting prediction directions. We classify 3D surface texture images based on minimum Euclidean distance between the test images and the training sets. The feasibility and effectiveness of our proposed approach can be validated by the experimental results.