TJ590 : 3D Environment reconstruction Using Optical Sensors Installed on Mobile Robot
Thesis > Central Library of Shahrood University > Mechanical Engineering > MSc > 2018
Authors:
Saeed Nouri [Author], Alireza Ahmadifard[Supervisor], Amir Hossein Nayebi Astaneh[Supervisor]
Abstarct: Outdoor maps and navigation information delivered by modern services and technologies like Google Maps and Garmin navigators have revolutionized the lifestyle of many people. Motivated by the desire for similar navigation systems for indoor usage from consumers, advertisers, emergency rescuers/responders, etc. In order to provide environment map, it is necessary to measuring depth simultaneously with the intensity and color of scene points. In this thesis, the depth is measured by a scan-line laser and the texture of scene points by RGB-camera is driven. These sensors fixed on a mobile robot’s manipulator. In order to find the localizing of a mobile robot in the scene, the stereo vision mechanism is used. Registration of 3D points scanned by laser scanner with the texture, the pose of camera relative to laser scanner should be determined.and color of that point is needed. Rotation matrix and translation vector of camera system fixed on the robot relative laser scanner is constant and is extracted in the calibration process. To find the robot position in 3D word coordinate system is used as a stereo vision system including two fixed cameras in the scanned environment. In a scenario mobile robot in the room to capture depth and texture information from different angles of the scene. The depth and texture information using by translation matrix of laser coordinate system to camera coordinate system and pose of the robot can be registered and plotted in 3D reference coordinate system. In the proposed method, due to the fact that the robot's position is not dependent on the information of the robot connected sensors, there is no accumulation error due to the robot locating itself. The resolution of the 3D reconstruction created from the environment in this study is 0.35 in 0.5 degrees. The results indicate that the reconstruction of the environment has very good accuracy.
Keywords:
#3D rendering #2D laser depth sensor #producing point cloud #indoor environments. Link
Keeping place: Central Library of Shahrood University
Visitor: