Use this URL to cite or link to this record in EThOS:
Title: An investigation of sewer pipe deformation by image analysis of video surveys
Author: Xu, K.
Awarding Body: University College of Swansea
Current Institution: Swansea University
Date of Award: 1995
Availability of Full Text:
Access from EThOS:
Close Circuit Television (CCTV) surveys of sewers are widely used in the UK to assess the structural integrity of sewer pipes. The video images are examined visually, and classified into five grades according to the degree of damage that can be observed. For severely damaged pipes, this technique is adequate, but there is considerable doubt in classifying pipes with very slight damage. In addition, the archiving of very many video tapes is expensive, and repeated access to video images of particular sections of pipe is difficult and time consuming. Therefore, an automatic sewer-pipe inspection system is required, based on the CCTV survey, which can extract and assess the structural condition of sewer pipes to ensure accuracy, efficiency and economy of sewer pipe examination. To this end, the objective of the thesis is to investigate the practical use of computer vision for automatic pipe-joint assessments, and the main effort is concentrated on software development. Initial damage in sewer pipes is associated with changes in shape of the pipe profile, with the undamaged profile being circular. Preliminary work was conducted to investigate the use of the pipe joint as a measure of the pipe shape. Automated pipe-joint shape assessments were investigated using existing software, but this could not handle images from video tapes. However, a manual technique proved that pipe joints could be used to assess pipe shape change. The main achievement of this work is the investigation of image processing algorithms, and associated software development, for pipe-joint boundary extraction which work with relatively poor contrast and noisy background as well as boundary shape recognition and analyses which deal with incomplete boundary outlines. Also a reference circle for the undamaged profile was estimated for use in pipe-joint shape discrimination. For most video pictures, reasonable results would be obtained with these algorithms. Two algorithms have been investigated for crack detection, one based on boundary curvature analysis, the other on a new boundary segment analysis technique. Also a neural network was introduced into pipe-joint shape discrimination. A Sewer Image Processing System (SIPS) has been established, using Microsoft Windows application software, which is based on the image processing techniques developed in this work.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available