| The technology of self-localization is one of the key techniques of robot autonomous navigation. Usually a mobile robot need a location sensor for positioning and need a map for its navigation. Using visual sensor to get environmental information, and realizing the positioning and mapping tasks in one step is a hotspot in the field of robotics research in the past few years. In this paper, the localization technology for autonomous robot in unknown environment is studied.This article firstly reviews robot simultaneous localization and mapping (SLAM) theory development and research status of history. On this basis, to different SLAM algorithm are simulated and compared, and put forward the DM3730based on embedded platform and visual sensor device realizes the robot localization and map reconstruction. For the visual part, Kinect sensors include an visible light color camera and a infrared monochromatic CMOS camera making use of structure light to measure depth, which can restore the environment’s depth information and visual information. Firstly, there are a lot of black hole in original depth images taken by Kinect visual sensors. this article uses the neighboring repair methods to amend depth-image in order to obtain more accurate depth information of environmental feature points. Secondly, the establishment of the environment map using feature map, using Oriented FAST and Rotated BRIEF (ORB) characteristics of the operator to achieve environmental characteristic extraction, and according to the characteristics of the ORB operator, put forward the Locality-sensitive hashing (LSH) can realize the matching of feature points, when dealing with high-dimensional data achieved good effect and the match speed. Finally, the collecting environmental information and robot pose information to the FastSLAM algorithm and realizes automatic mobile robot positioning function.In the experiments, the mobile robot motion control is realized by using command pattern, combining odometer information, electronic compass and visual sensor, etc., to realize mobile robot self-localization in the office environment, The experimental results show that the visual SLAM system based on Kinect and ORB feature is feasible and has an acceptable accuracy, with a good potential for indoor robot navigation tasks in an unknown environment. |
PDF Full Text Request