Identification Of Pipeline Cracking Area In Underwater Environment

Chong, Chean Yee (2019) Identification Of Pipeline Cracking Area In Underwater Environment. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Identification Of Pipeline Cracking Area In Underwater Environment.pdf - Submitted Version Download (286kB)

Abstract

Underwater pipelines are usually used to transport oil and natural gas in huge quantities and over long distances. Cracking in underwater pipelines is hard to be detected and the cracking had caused a large number of past accidents throughout the world. The main problem for detecting the cracking area or location is the poor visibility in underwater environment. The objectives of this study are to develop image processing method in identifying the cracking area using real underwater pipeline’s images and analyse the pipeline cracking area in different types of artificial underwater environment. In the proposed system, the images of cracking pipeline were collected and undergoes image processing algorithm, which include grayscale image, image filtering, image thresholding and edge detection. Then, contour method was used to find the coordinates of the cracking area from the processed images. Two experiments were designed to fulfil the objectives. The first experiment is to identify the pipeline’s crack using the images from real underwater environment, while other experiments are to determine the pipeline cracking area using images from artificial underwater environment in different conditions. In the first experiment, suitable filter, threshold method and edge detector are chosen for the system. The percent error of the proposed system is determined in Experiment 2 for different conditions. For the result of Experiment 1, the proposed system is including Gaussian filter, simple thresholding and Canny edge detection. From Experiment 2, the performance of the proposed system is the best in the condition of low turbidity level with high lighting level. The condition of high turbidity level with low lighting level had the worst performance as the resulted percent error had reached more than 70% and half of the cracking parts cannot be found in this condition. Therefore, turbidity and lighting conditions is important for image processing.

Actions (login required)

View Item