Towards decentralized ecological monitoring: A local-first web-based AI platform for sea turtle detection

Accurately monitoring sea turtle populations is crucial for informing effective conservation strategies; however, traditional methods for assessing spatial distribution and abundance are time-consuming, labour-intensive, and prone to observer bias. Unoccupied Aerial Vehicles (UAVs) have become valuable tools for collecting high-resolution imagery with minimal disturbance to marine fauna. However, the lack of standardised and user-friendly platforms for processing UAV data limits their broader application. In this paper, we present a lightweight browser-based web application designed to streamline the analysis of video collected by UAVs to monitor sea turtles. The application requires no additional software installations beyond a modern web browser and operates entirely on the client side, preserving data privacy. It supports the integration and execution of artificial intelligence models locally, facilitating the automated detection and classification of turtles in video footage. This tool, freely available, bridges the gap between UAV data collection and effective conservation-oriented decision-making by enabling rapid, standardised, and scalable analysis. Our approach promotes community-driven development and the reuse of AI models, making environmental monitoring practices more accessible and collaborative. • A browser-based platform enables local sea turtle detection from UAV footage without cloud uploads or software installation. • Two annotated datasets totaling 32,000 high-resolution frames support AI model training for sea turtles monitoring. • RT-DETR enables near real-time inference, matching expert precision while reducing analysis time by 60%. • Open-source architecture supports community model sharing and could extend to other marine megafauna conservation applications.