Non-invasive Underwater Marine Organism Observation Method based on Multibeam Sonar

The ocean harbors abundant biological resources, prompting the United Nations to initiate a decade-long effort in "ecosystem restoration." Understanding the behavior of marine organisms is essential for effective marine resource management and conservation. The traditional methods for tracking and monitoring marine life behavior include manual monitoring based on optical vision and tracking monitoring based on sensors. Non-invasive acoustic methods represent a crucial approach for underwater organism tracking, as they can operate effectively in complex conditions such as low light and turbid waters. This paper focuses on the tracking of underwater biological targets using multibeam forward-looking sonar. It combines particle filtering with the FCM algorithm and an optimized resampling particle filter algorithm for the identification and tracking of marine life in sonar images. Practical tracking was conducted through pool experiments, successfully tracking marine organisms such as turtle, fish, and scallops, achieving a tracking accuracy of over 80% with a distance error of less than 0.1 meters. This effectively demonstrates the capability of acoustic methods for underwater tracking.