Rehabilitation of our nation’s aging sewer pipes is of primary concern for municipalities to ensure normal operations and services on a continuing basis. This paper describes a novel computational approach for automated sewer inspection that circumvent drawbacks of AI-based methods. The model uses advanced cellular automaton techniques for CCTV image classification and anomaly identification. It operates in a discrete image space and consists essentially of representing images as a collection of cells and tracing changes in their boundaries to define anomalies, assess their severity, and grade and score them based on the NASSCO PACP standards. The major advantages of this approach are its processing speed and accuracy to detect defects and assess their severity, effectiveness and simplicity to produce three-dimensional BIM models of defect characteristics (e.g., size, location and severity), lack of bias and training data sets, and its applicability to any pipe material. It also makes it easier, faster, safer, and less costly to inspect sewers. Such capabilities greatly enhance the ability of utilities to accurately detect and classify defects in sewer networks to properly evaluate their physical and functional condition and optimize and prioritize rehabilitation and replacement decisions. The effectiveness and versatility of the model are demonstrated using sewer inspection data from various utilities. Enhancement of capital planning and asset management is a principal benefit of the proposed methodology. It also plays a significant role in extending an asset’s service life.
