Non-Invasive Submerged Pipeline Inspection Through Advanced Image Analysis and Deep Learning Algorithms for Quantitative Deterioration Assessment

An identification and amount of corrosion on the pipeline which is located at bottom of the sea can be achieved by utilizing information obtained from colored corroded pipes and by estimating through the oddy test. This study proposes an algorithm for enhancing the quality of deteriorated underwater images. The algorithm aims to reduce blurring effects and improve the quality of the images, which subsequently aids in the corrosion estimation and detection process. In order to estimate the corrosion age, the oddy test is utilized, which is typically used to determine the quality of materials that are suitable for storing, transporting, or displaying museum objects. A convolution neural network is employed to categories the corrosion rate detected at corroded pipes. An images used for object detection and training the CNN are obtained from the various source which can be accessed publicly. In the defined algorithm, a result of 98% accuracy is achieved, and the algorithm is capable of estimating the approximate age of the corroded pipeline.