Mobile visual detecting system with a catadioptric vision sensor in pipeline

Abstract In this paper, we focus on detecting the defects of a pipeline’s internal surface in a unified global coordinate system. An effective catadioptric vision based system for real time measurement in a metallic pipeline is presented. It consists of a catadioptric vision sensor, a laser locator and a mobile device. Due to the strong reflection and pitting corrosions of the pipeline inner surface, a series of structure lights are used in the catadioptric vision sensor to relieve the influence on the measurement accuracy of the metallic pipe. The 3D shape of the pipeline inner surface is achieved by periodic detection utlizing the mobile device where the catadioptric vision sensor is mounted. And the laser locator provides the position of the vision sensor in each measurement in a unified global coordinate system. Then, the 3D shape reconstruction in the global coordinate system considering all candidate laser positions in the sensor coordinates system is performed. Experimental results show that the designed system can well reconstruct the contours of the pipeline and it is feasible to be applied in the practical application of the pipeline’s internal measurement.