摘要
描述了一种基于广义Ｈｏｕｇｈ变换的部分遮挡物体识别方法．该方法分建模及识别两个过程，在识别过程中，景物图象经特征抽取、组合特征匹配、变换矩阵计算，进而根据广义Ｈｏｕｇｈ变换在参数空间求出与景物对应的模型类及相应的变换矩阵．由于用了广义Ｈｏｕｇｈ变换，物体识别过程中不必求出物体的全部特征，因而能较好地识别部分遮挡的物体，且该方法对噪声不敏感．为了说明该方法。
This paper describes a method for recognizing partially occluded objects, which is based on the generalized Hough transform. The method includes two stages: modeling and recognizing. During recognizing, through features extracting, matching grouped features, pose clustering in parameter space based on the generalized Hough transform, the classes and pose of the object may be obtained from a scene image. Using the generalized Hough transform, eliminates the need to compute all features of the object, so the method is preferable for recognizing partially occluded objects and is insensitive to noise. To illustoate the method, a simple example is included, some experiment results are given as well. Some tools that occlude each other can be recognized successfully with a PC system under experimental conditions.
