基于尺度不变特征的从纹理回复形状  

SFT Based on Scale Invariant Feature

作　　者：王强[1] 彭思龙[1] 李振伟[1] 

机构地区：[1]中国科学院自动化研究所国家专用集成电路设计工程技术研究中心,北京100080

出　　处：《计算机工程》2009年第2期216-218,共3页Computer Engineering

摘　　要：提出一种估计纹理平面方向的方法。该方法提取纹理图像中具有尺度不变特征的关键点,求出以这些关键点为中心的局部模式尺度和方向正规化特征,按特征将模式分类。假定图像中同类局部模式的尺度变化仅由成像时的透视效果产生,利用纹理图像中局部模式尺度变化与射影变换的关系估计纹理平面的方向。实验结果证明,此方法简单有效。This paper presents a method for estimating the orientation of a textured surface. Keypoints with scale invariant features in a texture image are detected, and the local patterns with these keypoints as centers are described by the direction and scale normalized features. The patterns are clustered by the features. Under the assumption that all the scale variation of the similar patterns is attributed to the perspective effects the surface orientation is estimated. Experimental results show that the method is simple and effective.

关 键 词：尺度不变特征 单应 从纹理恢复形状 

分 类 号：TN911.72[电子电信—通信与信息系统]