3D Indoor Environmental Reconstruction System that Allows to Determine the Trajectory of a Mobile Robot Through Analysis and Image Processing

Abstract

This paper presents the results of 3D reconstruction of indoor environments using  as a unique technique the reconstruction technique  of stereo vision and  Matlab software for obtaining point clouds, subsequent stitching  using the ICP algorithm and trajectory calculation using the same reconstruction making an estimate of location in the  indoor area  . To obtain the information of the environment  to  be reconstructed, a low-cost non-autonomous mobile robot commanded by bluetooth was implemented, mounted on this a stereo camera MYNT EYE S1030 was available for the capture of stereo images and having a certain trajectory in order to obtain the data of the environment in 360 grams with respect to thehorizontal axis. The results, based on the statistical analysis carried out on this system,  showed variations in the fidelity of the physical environment compared to the virtual environment with an average error of 3.27 meters, and  is intended to be used in the topography and mapping of indoor environments.