Automatic Visual Leakage Detection and Localization from Pipelines in Chemical Process Plants Using Machine Vision Techniques

Abstract

Abstract Liquid leakage from the pipelines is a critical issue in large-scale process plants. Damages in the pipelines affect the normal operation of the plant and increase the maintenance cost. Furthermore, they cause an unsafe and hazardous situation for the operators. Therefore, detection and localization of leakages is a crucial task for maintenance and condition monitoring. Recently, using infrared (IR) cameras for leakage detection in large-scale plants was found as a promising approach. IR cameras can capture the leaking liquid if it has a higher (or lower) temperature than its surroundings. In this paper, a method based on IR video data and machine vision techniques is proposed to detect and localize liquid leakages in the chemical process plant. Since the proposed method is a vision-based method and does not consider the physical properties of leaking liquid, it is applicable for any type of liquid (water, oil, etc.) leakage. In this method, subsequent frames are subtracted and are divided into the blocks. Then, principle component analysis is performed in each block to extract features from the blocks. All subtracted frames within the blocks are individually transferred to the feature vectors, which are considered as a basis for classifying the blocks. For classification, the k-nearest neighbor algorithm is used to classify the blocks as normal (non-leakage) or anomalous (leakage). Finally, the positions of leakages are determined in each anomalous block. In order to evaluate the approach, two data sets with two different formats, consisting of video footages of a laboratory demonstrator plant captured by an IR camera, are considered. The results show that the proposed method is a promising approach to detect and localize leakages from pipelines using IR videos. The proposed method has a high accuracy and reasonable detection time for leakage detection. Furthermore, the possibilities of extension of the proposed method to a real industrial plant and the limitations are discussed at the end.