Automated pipe inspection using ANN and laser data fusion

Standard CCTV (close circuit television) is currently used in many pipe inspection applications, such as sewers. This human-based approach is prone to error because of the exorbitant amount of data to be assessed, and smaller anomalies or defects are likely to be overlooked reducing the chance of detection of faults at an early stage. Laser profilers for pipe inspection have been recently proposed to overcome CCTV problems. Positional as well as intensity information, related to potential defects, can be extracted from the laser-camera acquired images. While most of these systems are based on the geometrical analysis of pipes, here the intensity distribution of the reflected light is also exploited. This paper describes the strategies developed for the automation of defect classification in pipes and explores new methods to fuse intensity and positional information and shows how they can be used to improve multi-variable defect classification. A neural network-based classification method is presented. Experimental results are provided.