AI Model for Automated Detection and Mapping of Intertidal Vegetation

There is a dearth of accessible information and innovative tools to map and obtain biomass data for analysis, conservation, and sustainable resource management of intertidal vegetation in the United States and globally. A major technical problem is the integration of biological and physical data at small spatial scales (2 meters). We propose Unoccupied Aerial Systems (UAS) can be used for quality control, co-registration, training, and validation of satellite or high-altitude or space-based aerial imagery. Specifically, we propose to develop training data using Unoccupied Aerial Systems (UAS) to bridge the gap between small- and large-scale ecological and remote sensing data, and to use Artificial Intelligence (AI) and Machine Learning (ML) to accurately map, classify, and estimate biomass of intertidal vegetation such as macroalgae. This Phase I study will assess the feasibility to develop a deep learning model based on UASacquired data and satellite imagery in the detection, classification, and biomass estimation of intertidal vegetation at local, as well as state-wide or regional scales.