Three Dimensional Segmentation of Object and Its Characteristic Points Definition

EasyChair Preprint 7102

Abstract

Reconstructing three dimensions (3D) of object is a big challenge for researchers in many years. To solve these problems, it is necessary to identify objects and obstacles on the 3D description of the working stage, to determine their real dimensions and spatial coordinates. To obtain a 3D description of the working stage, different systems of technical vision are commonly used: multi-camera, stereo systems, laser illuminated systems, etc. In this case, the system of vision with a structural illumination is considered, at the output of which three coordinates of each point of the working stage are formed. In this paper, algorithms for image segmentation are examined and a way of describing spatial objects is proposed. The algorithm for segmentation was developed by us and was presented in details in this study. The aim of segmentation is to partition the original set of points into disjoint subsets. Examination of complex shape body showed that the algorithm can be extract detailed geometry of the model. Accuracy of the algorithm in analyzing geometry structure was presented detailed in this study.

Keyphrases: 3D segmentation, image segmentation, working stage

@booklet{EasyChair:7102,
  author    = {Anh Van Nguyen and The Hung Tran},
  title     = {Three Dimensional Segmentation of Object and Its Characteristic Points Definition},
  howpublished = {EasyChair Preprint 7102},
  year      = {EasyChair, 2021}}