Since most urban gas pipelines are buried under the ground, there are more demands for using trenchless technology in pipe inspection. Employing intelligent in-pipe robots, which can carry NDT devices and tools into pipelines to complete inspection and maintenance, appears to be one of the most attractive solutions now. While travelling into urban gas pipelines, there is no light source in dark pipelines, and inspection robot should carry the illumination instruments providing brightness. As the gas pipeline is a structured and constrained space, the images taken from robot vision system often cover advanced knowledge about environment structure and useful information of events in scene, which can be used for autonomous navigation of in-pipe robot. On the basis of analysis of the model of optics and the property of imaging process, the image of inner pipe wall can be divided into three different regions corresponding to bright region, gray region, and dark region. According to the change obtained from three regions in the image, the new methods of recognition for straight pipe, elbow pipe, `T' shaped branch, and blocks were developed. Then, combining these recognition methods, a navigation strategy with monocular vision system was presented for the robot autonomously navigating in gas pipeline. The simulation experiments show that the presented methods are available for recognition of straight pipe, elbow pipe, `T' shaped branch, and blocks in gas pipeline environment. This has benefits for improving the intelligence and automation of gas inspection robot.
Recognition of Environment for Navigation of Gas Pipeline Inspection Robot Based on Monocular Vision System
International Conference on Pipelines and Trenchless Technology (ICPTT) 2009 ; 2009 ; Shanghai, China
ICPTT 2009 ; 240-251
2009-09-15
Conference paper
Electronic Resource
English
Monocular Vision for Mobile Robot Localization and Autonomous Navigation
British Library Online Contents | 2007
|Monocular Vision Based Object Recognition and Tracking for Intelligent Robot
British Library Conference Proceedings | 2017
|Optical Flow-Based Monocular Vision/INS Integrated Navigation for Mobile Robot Indoors
Trans Tech Publications | 2014
|Monocular Vision based Navigation in GPS-Denied Riverine Environments
British Library Conference Proceedings | 2011
|