Extraction of Sewer Pipe defects using Morphological segmentation based on Edge Detection

• Main Authors: Shi-Zhi Chen, 陳世植
• Other Authors: Tong-Ching Su
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/96101423734632130501

碩士 === 國立金門大學 === 土木與工程管理學系碩士班 === 100 === At present, image processing and artificial intelligence techniques have been used to develop diagnostic systems to assist engineers in interpreting sewer pipe defects on CCTV images to overcome human’s fatigue and subjectivity, and time-consumption. Based on the segmented morphologies on images, the diagnostic systems were proposed to diagnose sewer pipe defects. Opening top-hat operation coupled with Otsu’s thresholding is usually applied to morphology extraction. In this thesis, a novel approach of morphological segmentation based on edge detection (MSED) was also presented and applied to identify the morphology representatives for the sewer pipe defects on CCTV images. Before the implementations of opening top-hat operation or MSED, the median filters of 3×3 or 5×5 are employed to reduce image noise as well as keep informative textures. The 8 illustrations available at the Sewerage Rehabilitation Manual of Water Research Centre, UK were selected to be the testing images. Compared with the performances of opening top-hat operation and MSED, median filtering of 5×5 followed by opening top-hat operation is merely suitable in morphology extraction of open joint. However, median filtering of 3×3 followed by MSED could effectively extract the representative morphologies of fracture, spalling, deformation, hole, and collapse. This result demonstrates that MSED outperform opening top-hat operation. Besides, another 16 inspection images showing the sewer pipe defects of fracture and open joint were selected to be tested. The testing result indicates that the representative morphologies could not be extracted due to the inappropriate luminance or image contrast of the CCTV equipment. Hence, a well imaging condition should be built during CCTV inspection inside sewer pipes.