基于随机Hough变换的深度图像分割  

Range Image Segmentation Based on Randomized Hough Transform

作　　者：丁益洪[1] 平西建[1] 胡敏[1] 

机构地区：[1]信息工程大学信息工程学院,郑州450002

出　　处：《计算机辅助设计与图形学学报》2005年第5期902-907,共6页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金 ( 60 2 72 0 0 4)

摘　　要：提出了基于随机Hough变换的深度图像分割算法,该算法采用随机Hough变换在深度图像中寻找平面,具有对噪声不敏感的优点通过对一常用深度图像数据库(ABW图像库)的分割实验,并将实验结果同4种经典的深度图像分割算法在同一数据库中的分割结果作了比较分析,表明该算法对噪声不敏感。A novel range image segmentation algorithm based on RHT (Randomized Hough Transform) is presented in the paper. The algorithm finds planar regions by utilizing RHT and has the advantage of insensitivity to noise. The proposed algorithm has been tested on a popular range image database (ABW image database) with comparisons to four typical range image segmentation algorithms. The experiment results show that the proposed algorithm has a strong anti-noise ability and its segmentation performance is superior to the four algorithms tested.

关 键 词：深度图像分析 深度图像分割 随机HOUGH变换 平面检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]