Realization of Point Cloud Maps Using ROS & Visual Sensor on Raspberry Pi 3 Based Mobile Robot

摘要

This paper present Point Cloud maps using ORBSLAM with visual sensor or monocular camera like the webcam, including map reuse, loop closing and relocalization capabilities. The system integrated with ROS works in real-time on raspberry pi based mobile robot from small hand-held indoor sequences. This system performs relocalization and loop closing in real-time. The ORB SLAM method is used and it detects targets which will help the mobile robot to explore unknown environments. The entire design of this mobile robot realization will be designed using a framework that is the Robot Operating System (ROS) with the help of tools that are already available in it. For results in the realization of point cloud using the ORB SLAM method can detect objects optimally with an object size of 31.5x56 cm at a distance of 70 cm and get an error value of 1.21% and an accuracy value of98.79%, and the system can work in unknown environments and get a visualization map that comes from the camera sensor.