Video Processing Techniques for Assisted CCTV Inspection and Condition Rating of Sewers

Nima Sarshar, Mahmoud Halfawy, and Jantira Hengmeechai

ABSTRACT

Sewer historical condition information is critical for establishing deterioration patterns, and hence the development of statistical models to predict future condition of the sewers. Such information is the key enabler for developing proactive asset management and renewal planning strategies. The lack of adequate historical condition data has caused many of the decisions to be based on educated guesses and subjective assessment of the sewer condition. Most Canadian municipalities possess hundreds or thousands of hours of CCTV recordings that, unfortunately, cannot be feasibly organized and analyzed to extract meaningful condition data.

This paper describes the development of a software system to semi-automatically extract historical condition data information from archived sewer inspection CCTV files. The software represents and manages the sewer data in a centralized repository. Users can easily access inspection and condition data via a GIS interface. The software implements video processing algorithms to support semi-automatic evaluation of the CCTV footage by automatically detecting defect features and assisting the user to define scores and rating according to a pre-defined condition rating scheme. The software was successfully used to efficiently access, analyze, and evaluate sewer condition data from CCTV video files obtained from the City of Regina.


Permanent link: