Real-Time Marine Mammal Detection System Unveiled
A new real-time marine mammal detection system combines AI, robotics, and sensor fusion, using multi-modal sensors and AI-driven detection. While focused on marine life, its architecture is similar to those needed for vessel and dark target identification, with insights into robust, real-time multi-sensor data fusion.
The real-time detection system emerges from a collaboration between the University of Plymouth and MSeis, supported by Innovate UK. The partnership leverages expertise in marine conservation, robotics, marine autonomy, and sensor technology. The system integrates advanced computer vision models into MSeis's dual-sensor thermal and optical camera platforms. It aims to automate the detection and identification of marine mammals and floating anomalies like debris, enhancing accuracy and efficiency. Expert spotters inform the system's development, ensuring reliable environmental monitoring and marine risk mitigation. The technology addresses a growing need for marine wildlife protection amidst expanding offshore industries like renewables and subsea operations. Existing methods often rely on manual observation, which this system seeks to improve upon. Other AI-powered systems like Tetra Tech's Neptune and RPS's Neptune also exist for marine mammal detection, demonstrating the increasing reliance on AI in this field. The system's architecture shares similarities with those used for vessel and dark target identification. "Dark targets" refer to vessels not broadcasting an automated identification signal, often associated with illegal activities. Satellite technology, including optical, radio frequency, and synthetic aperture radar sensors, plays a crucial role in detecting these vessels.
