AI-based technology to monitor behavioural responses in marine invertebrates

The continuous release of contaminants from anthropogenic activities poses a serious threat to the marine environment, especially in coastal areas. In this context, the use of bioindicators is crucial to monitor the health of marine coastal ecosystems. Alteration of marine organisms behaviour is useful to provide insights into pollution levels, representing an environmental early warning. However, detecting and quantifying behavioural responses presents critical issues, - i.e. time consuming, operator-dependent – that affect data accuracy and reliability. To address these challenges, the present study focuses on the development of an AI-based technology, able to detect behavioural responses in marine invertebrates for an automatic environmental monitoring. Specifically, the AI-based technology has been validated towards crustaceans (barnacles) behavioural responses, namely valves opening. Preliminary results on barnacle valves opening frequency and active organisms percentage were obtained through in laboratory and in field considering the influence of different parameters (i.e. temperature, food source, suspended organic matter) on this behavioural alteration. Data obtained through the AI-based technology have been validated by manual detection. Results confirm the accuracy, efficiency and reliability of behavioural responses of marine crustaceans under different environmental scenarios, promoting its use as an early warning tool for assessing marine environmental health also by using an AI-based technology.