基于自适应3D－Hough变换方法估计三维物体的位置参数

本文提出一种基于３Ｄ－Ｈｏｕｇｈ变换检测深度信息图像中平面特征及估计三维物体位置参数的新算法．利用３Ｄ－Ｈｏｕｇｈ参数空间中位移和旋转相分离的特征，多层次地估计物体的位置参数．因为该算法直接从输入的深度信息图像获得几何特征，所以该算法具有较好的鲁棒性．

An Approach to 3D-Pose Determination Using Adaptive 3D-Hough Transform

In this paper, we introduce a new way for plane detection and determination of the location and orientation of general 3D objects from range image. This technique is based on a 3D extension of the well-known straight line Hough transform. The technique is very efficient and robust, since the dimension of the feature space is low and it uses range image directly. The feature space separates the translation and rotation effects, and the hierarchical algorithm obtained to detect object rotation is promising.