Research On Thinning Algorithm Of 3D Target

3D thinning is an iterative process that erodes an object layer by layer from outside to inside at the same time keeps the original topology until only a skeleton is left. The skeleton extracted by 3D thinning is an important foundation for subsequent image analysis and feature extraction, widely used in the field of image analysis and visual measurement. In the production process, the thinning results can provide the dimension characteristics and dimensional datum of the target which can accurately describe the actual current state of the target, so as to achieve the all-round non-contact measurement in production process.The skeleton extracted accurately by 3D thinning asked to preserve topology structure of the target and keep the original connectivity, at the same time 90° rotation invariance was required in 3D vision measurement after 3D thinning in the production process. Focusing on the problems that the results of the existing thinning algorithms failed to preserve the connectivity, destroyed the original topology structure and had no rotation invariance, a novel thinning algorithm based on topology preservation and rotation invariance was proposed in this paper. First of all, the isotropic deleting templates are designed which make the templates be consistent in all directions, then ensure that the algorithm has rotation invariance; Secondly, according to the need to define the authenticity delete points, and demonstrate the proposed algorithm can satisfy the continuity conditions. By the hypothesis testing method, detecting whether the number of object and background component in its 26-neighborhood changed after all candidate deleting points were deleted, to confirm the true and false of the deleting point. It preserves the original topology structure, then ensure the connectivity is unchanged after rotation.Experimental results show that the new method is able to preserve connectivity and topology structure, and results are unchanged after rotation.