Aquatic Animal Monitoring System

Abstract

Abstract Aquatic Animal Monitoring Systemis initiated as part of PTTEP's Ocean for Life strategy as we thrive in enhancing Ocean Health & Biodiversity Monitoring to ensure that PTTEP's offshore operations are friendly and safe to the surrounding environment and aquatic animals. The basis of the Aquatic Animal Monitoring Systemproject focuses on conservation survey and tracking of rare aquatic animals as well as marinebiodiversity. As part of the process, an underwater camera was installed on a jacket leg of PTTEP's platform to allow the video recording of underwater lives. The video footage was then analyzed by Artificial Intelligence (AI)software using an object detection method for determining the animal's categorization, then using machine learning algorithm for more accuracy. This concept can visualize aquatic animals around the platform and the surrounding environment. Moreover, the AI software can shorten the video by cutting off any non-life appearing period. Therefore, this technique can support a processor during the video analysis from the platform, contributing to a better work efficiency as it can save time, manpower, and most importantly cost. For the detection algorithm, all targets generatea large amount of data in the form of images with labels in order to train a software to memorize the target objects. The AI software was able to detect and identify nine species of aquatic animals which are fish, turtle, whale, dolphin, shark, seal, sea lion, stingray, and seahorse. With AI software in place, the video raw file can be shortened up to 85% by removing non-life periods in the original video and tracking only animal life in the video frame. This is a significant milestone for PTTEP in creating sustainable values to the ocean, which is considered as PTTEP's second home. Adopting artificial intelligence and machine learning technology to this project, it helps categorizing aquatic animal types and shorten a videofile. Moreover, it can save manpower and time.