Pipe inspection using intelligent analysis techniques with high noise-tolerance

Standard sewer inspection systems are based on closed circuit television (CCTV) cameras mounted on wheeled platforms. One of the disadvantages of camera inspection systems is that they can detect only a small part of all possible sewer damage that could conclude in collapses. The inspection outcome of standard CCTV systems relies not only on the quality of the acquired images, but also on the off-line recognition and classification conducted by human operators. The objective of this research is the development of intelligent sensor systems that will enable the automation of the pipe condition assessment. Optical techniques are proposed to complement the existing CCTV-based approach and to improve inspection results. Besides that, automated defect recognition algorithms based on Artificial Neural Networks are proposed. Experiments to test the tolerance of the automated. algorithm to artificially-generated noise have been conducted and results are presented.