Drone and ground-truth data collection, image annotation and machine learning: A protocol for coastal habitat mapping...

Area of Science:

• Marine ecology
• Remote sensing
• Geospatial analysis

Background:

• Coastal habitats are vital ecosystems requiring accurate monitoring.
• Traditional mapping methods lack the spatial and temporal resolution needed for detailed habitat assessment.
• Aerial drone imaging provides a cost-effective solution for high-resolution coastal habitat mapping.

Purpose of the Study:

• To present a systematic method for shallow water habitat classification using drone imagery.
• To demonstrate the application of this method in a case study involving diverse coastal habitats.
• To enable detailed and efficient mapping for improved understanding and management of marine ecosystems.

Main Methods:

• Collection of aerial drone images and creation of orthomosaics.
• Acquisition of ground-truth data for supervised image annotation and map validation.
• Training of machine learning algorithms for automated habitat classification based on annotated imagery.

Main Results:

• Successful classification of coastal habitats including seagrass, seaweed, kelp, sediments, and rock.
• Demonstration of drone imaging's capability to differentiate various coastal vegetation types.
• Generation of detailed habitat maps with high spatial and temporal resolution.

Conclusions:

• The presented drone-based method provides an efficient and detailed approach for shallow water habitat mapping.
• This methodology supports the sustainable management of ecologically valuable marine ecosystems.
• High-resolution drone imagery combined with machine learning enhances coastal habitat monitoring capabilities.