A laboratory test-bed for an automated power line inspection system

Aerial inspection of overhead power lines using video surveillance techniques requires that the camera acquire and smoothly track the object of interest, usually the top of a wood pole on the 11 kV electricity distribution network. Methods for achieving this using active vision, where the camera sightline is steered automatically according to features within the image, are reported here. The focus of the paper is the computer simulation and experimental test-bed used to develop, test and assess the performance of the methods. The characteristics of the acquisition method and the visual feedback loop for tracking are described. A computer simulation is used to predict the effect of various error sources on sightline accuracy and to visualise how the image would look to the observer during operation. The construction of a laboratory test rig is described and results for closed-loop visual tracking are presented.