Determination of Internal Elevation Fluctuation from CCTV Footage of Sanitary Sewers Using Deep Learning

Abstract

The slope of sewer pipes is a major factor for transporting sewage at designed flow rates. However, the gradient inside the sewer pipe changes locally for various reasons after construction. This causes flow disturbances requiring investigation and appropriate maintenance. This study extracted the internal elevation fluctuation from closed-circuit television investigation footage, which is required for sanitary sewers. The principle that a change in water level in sewer pipes indirectly indicates a change in elevation was applied. The sewage area was detected using a convolutional neural network, a type of deep learning technique, and the water level was calculated using the geometric principles of circles and proportions. The training accuracy was 98%, and the water level accuracy compared to random sampling was 90.4%. Lateral connections, joints, and outliers were removed, and a smoothing method was applied to reduce data fluctuations. Because the target sewer pipes are 2.5 m concrete reinforced pipes, the joint elevation was determined every 2.5 m so that the internal slope of the sewer pipe would consist of 2.5 m linear slopes. The investigative method proposed in this study is effective with high economic feasibility and sufficient accuracy compared to the existing sensor-based methods of internal gradient investigation.

Multi-scale hybrid vision transformer and Sinkhorn tokenizer for sewer defect classification

“…The field has, however, become more transparent as many have started to directly compare different methods on the same datasets, in an effort to offset the lack of public detection and segmentation datasets [17,36,34]. Recently, the field has also started investigating other parts of the sewer inspection process [30,32,17,[37][38][39][40][41], such as Haurum et al [37] proposing a multi-task classification approach for simultaneously classifying defects, water level, pipe material, and pipe shape, and Wang et al [30] proposed a framework to accurately determine the severity of defects related to the operation and maintenance of the pipes. The field has also adopted recent trends from the general computer vision field such as selfsupervised learning [39], synthetic data generation [25,24,[42][43][44], neural architecture search [45], and usage of the Transformer architecture [17,46], indicating that the automated sewer inspection field is catching up to the general computer vision domain.…”

Research on the Application of MEMS Gyroscope in Inspecting the Breakage of Urban Sewerage Pipelines

“…The CCTV method and QV are widely used in Sewerage network inspection [5]. CCTV inspection of pipelines is costly, laborintensive, time-consuming, and error-prone, e.g., CCTV inspection requires labor-intensive and costly pre-inspection pilot draining and blocking of the inspected pipe section, and the inspected sewer mileage generates hours of video, which needs to be handled by trained and certified inspectors, a great demand for automatic identification technology [6,7]. Furthermore, the range of pipelines that can be inspected at any one time is quite narrow, for a large city, only about ten percent of them can be inspected each year [8,9].…”