Using Range Information Propagation Transform in Object Recognition
基于距离信息扩展原理的物体识别方法

Using sparse range image and gray level image in combination.features extracted from the gray level image can be assigned conditionally with corresponding range values by means of a projective mapping.It is called rank information propagation transformation.The range information propagated in this way can be used to calculate the Euclidean invariant of corresponding surface planar patterns and hence to generate initial hypothesis.Since initial hypotheses are extracted based on reliable information.and hybrid control strategy with feedback is used.the proposed method can be used in practice situations.In this paper the principle of the method and its implemental issues are addressed.and some preliminary experimental results are presented to demonstrate its effectiveness and efficiency.