Automated assessment of buried pipeline defects by image processing

Many underground water pipelines are old and approaching their service lives in a great number of cities. With the promotion of sustaining buried infrastructure, it's necessary to pay much attention on how to effectively extract defect characteristics of damaged pipelines. Detection of defects in underground pipes is a crucial step to assess the deterioration degree of pipeline for municipal operators. Based on the image processing theory, a defect feature extracting method under HSV color space is proposed in this paper. QFCM (Quick Fuzzy C-Mean clustering) segmentation arithmetic is applied to extract characteristics parameters. The proposed algorithm can identify defects from background, and the types of defects in the buried pipes can be categorized in the estimation stage. Then, different methodologies of parameters extraction are applied in different types of pipe defects, features like area, angle, length and width of defects can also be calculated. And then, a method of assessing the accuracy of feature extraction algorithm is discussed. Finally, the proposed detection approach has been experimentally tested using a group of images acquired by CCD camera from real inspection scenarios. The experimental results proved that it is feasible and effective to apply the system in feature extraction of pipe defects of the underground water-pipelines.