Vision for a Power Distribution Line Maintenance Robot -Environments Input

Abstract

This paper describes a vision system for a power distribution line maintenance robot. The robot will perform a task, for example, a switch gear installation task automatically. To perform the task, 3D model of work environments are constructed in a computer. Target positions for robot commands are generated using the work environment 3D model. We investigated three methods to superimpose 3D frame models of working environment parts on the photographs: 1) Manually superimposing method by a mouth, 2) Manually pointing method, and 3) Automatically superimposing method.