Image Processing–Based Framework for Determining Deterioration in Sewer Pipe Defects

Abstract

Municipal drainage systems play a key role in public health and are considered one of the main components of every modern city’s infrastructure. However, as a drainage system ages, its pipes gradually deteriorate at rates that vary based on the conditions of utilization. To prevent unexpected failures, municipalities have adopted a proactive approach that relies on regular condition assessments of their assets. At this juncture, a question that needs to be answered is how these assessment data can be used to optimize the frequency of inspections of the drainage pipes. In fact, when the assessment information is used in conjunction with deterioration models, city managers are able to develop data-driven maintenance, rehabilitation, and replacement plans based on the current condition of the assets and their risk of failure. As a result, understanding the rate at which defects evolve over time provides valuable information in terms of understanding the relationship between the various factors affecting defect development and pipe deterioration. This research presents an image registration framework for extracting crack development information from closed-circuit television (CCTV) videos of sewer pipes. Image processing techniques are used to estimate the relative change for a given defect from images taken at two different times. Because the parameters of cameras (e.g., model, location, angle of view) are generally not expected to be identical for consecutive inspection campaigns, the images to be compared were first scaled using a technique referred to as the area scaler (AS) to ensure all images have the same frame of reference. This scaling procedure is illustrated in a case study, containing 49 pairs of images that led to a relative error (with respect to the mean) generally not exceeding 5% when frames of the same defects contained a sufficient number of matching points.