Sewer inspection by remote CCTV camera is standard practice world-wide. The technique, evolved over 25 years uses a CCTV camera and light source carried along the sewer pipeline by a tractor unit on the end of a cable. The process is dependent on operator skill in 'driving' the tractor and observing sewer defects.
The ISAAC research explored applications of existing machine and computer vision techniques for sewer inspection to improve survey quality and/or inspection process efficiency.
Computer vision techniques examined include:
- Omni-directional cameras. This technology allows omni-directional views to be added in real time to the standard camera view that looks ahead along the sewer pipe axis.
- Imposing scale onto panoramic images. This allows metric measurement from images.
- Manipulation of stored images. This allows alignment of newly captured sewer pipe image and archived ones from previous inspection(s) to aid assessment of rate of sewer deterioration.
- Laser-based orientation. This uses a projected laser cross and camera to establish orientation of an inspection unit in the sewer.
- Reprojection. This examined reprojecting an image captured by a camera which is offset from the pipe centre so the image appears captured from camera position on the pipe axis.
This research was supported by the European Framework 5 Information Society Technologies Programme and involved the following collaborators:
- Czech Technical University, Czechoslovakia
- Prague Water Supply and Sewerage Company, Czechoslovakia
- WRc, UK
- Pearpoint Limited (manufacturer of opto-elctronic systems), UK
- Anglian Water, UK
- National Research Center for Information Technology, Germany
This project started in January 2002 and was completed in February 2003.
This research will improve the quality of CCTV surveys for sewer inspection and the associated diagnosis of the results.