You are here

Automated Underground Pipe Inspection Using a Unified Image Processing and Artificial Intelligence Methodology

TitleAutomated Underground Pipe Inspection Using a Unified Image Processing and Artificial Intelligence Methodology
Publication TypeThesis
Year of Publication2000
AuthorsSinha, S. K.
Academic DepartmentDepartment of Systems Engineering
UniversityUniversity of Waterloo
CityWaterloo, Ontario, Canada
Thesis TypePh.D. Thesis

The National Science Foundation (NSF) has estimated the total U.S. investment in civil
infrastructure systems at US $ 20 trillion. The investments in underground infrastructure
systems represent a major component of this overall investment. Recent studies have
shown that the cost of replacing all water mains in the United States would run to US $
348 billion. The estimated cost to upgrade the water transmission and distribution system
is US $ 77 billion. In Canada, the estimated cost to bring the water mains and sewer
pipelines to an acceptable level is CDN $ 11.5 billion and 47 billion, respectively. Many
of these pipeline systems are eroding due to ageing, excessive demand, misuse, poor
construction, mismanagement and neglect. Due to their lack of visibility, rehabilitation of
underground pipeline is frequently neglected until a catastrophic failure occurs, resulting
in difficult and costly rehabilitation.
The enormity of the problem of deteriorating municipal pipeline infrastructure is
apparent. Since neglecting or rebuilding the pipeline system is not financially realistic,
asset managers require the capacity to monitor the condition of underground pipes. Thus,
reliable cost effective pipeline assessment methods are necessary so that pipeline
managers can develop long-term cost effective maintenance and rehabilitation programs.
These programs are necessary to ensure that critical pipeline sections are repaired or
replaced before they fail.
Closed circuit television (CCTV) surveys are used widely in North America to assess the
structural integrity of underground pipes. The video images are examined visually and
classified into grades according to degrees of damage. The human eye is extremely
effective at recognition and classification, but it is not suitable for assessing pipe defects
in thousand of miles of pipeline images due to fatigue and cost. In addition, manual
inspection for surface defects in the pipeline has a number of drawbacks, including
subjectivity, varying standards, and inspection time. These concerns have motivated us to
conduct a research for the development of an automated pipe inspection system, based on
the scanned images of underground pipes.

This thesis presents a system for the application of computer vision techniques to the
automatic assessment of the structural condition of underground pipes. Automatic
recognition of various pipe defects is of considerable interest since it has the potential to
solve problems of fatigue, subjectivity, and ambiguity, leading to economic benefits. The
main efforts of the research are placed on investigating algorithms and techniques for
image pre-processing, segmentation of pipe objects (i.e., cracks, holes, joints, laterals,
and collapse surface), crack detection, feature extraction and classification of defects. In
this study, an attempt has also been made to develop a framework for an integrated
pipeline network management. The proposed integrated pipeline management system is
necessary to help municipal managers to make consistent and cost-effective decisions
related to the preservation of underground pipeline systems.