Presentation Information
[TPZePaqP-03]Blue discovery 2.0: the development of Self-made AI-USV and Multispectral UAV to Investigate Seagrass Meadow and DO Monitoring in Penghu Archipelago, Taiwan
*Cheng-Hao Lu1 (1. National Penghu University of Science and Technology (Taiwan))
Seagrass meadows serve as vital blue carbon ecosystems and critical coastal habitats, yet globally and locally, there remains a lack of an efficient, low-cost monitoring system capable of large-scale and repetitive surveys. Traditional seagrass assessments heavily rely on manual intertidal wading or boat-based scuba diving. These conventional methods suffer from severe operational bottlenecks: limited air tank capacity, strict diving safety constraints, and brief effective working windows. Consequently, surveying deep or expansive seagrass beds is exceptionally difficult, carrying escalating human risks and prohibitive costs as depth increases.
To overcome these constraints, this study introduces an innovative, cost-effective, and non-invasive integrated framework that pairs a multispectral Unmanned Aerial Vehicle (UAV) with a custom-built Unmanned Surface Vehicle (USV). The self-made USV represents a next-generation scientific research vessel equipped with multi-channel communication systems, high-precision RTK GNSS positioning, bathymetric mapping capabilities, and multi-parameter water quality sensors for real-time dissolved oxygen (DO) and gas monitoring. Crucially, the USV features an onboard edge-computing AI model trained for high-accuracy, real-time seagrass identification.
Field deployments conducted across the Penghu archipelago demonstrate the successful synergy of airborne and surface automation. The multispectral UAV provides expansive spatial overviews, while the AI-USV performs close-range, high-resolution under-surface verification, bathymetric profiling, and localized water quality mapping. This integrated approach successfully generated precise spatial distribution maps of seagrass coverage alongside synchronized DO and environmental data. Supported by systematic, long-term robotic surveys, this framework delivers an empirical baseline for marine habitat management and ecological restoration strategies. While the initial development focuses on establishing foundational data for blue carbon accounting in Taiwan, the ultimate trajectory of this platform is to scale into a universal, high-efficiency tool for monitoring broader nearshore ecosystems.
To overcome these constraints, this study introduces an innovative, cost-effective, and non-invasive integrated framework that pairs a multispectral Unmanned Aerial Vehicle (UAV) with a custom-built Unmanned Surface Vehicle (USV). The self-made USV represents a next-generation scientific research vessel equipped with multi-channel communication systems, high-precision RTK GNSS positioning, bathymetric mapping capabilities, and multi-parameter water quality sensors for real-time dissolved oxygen (DO) and gas monitoring. Crucially, the USV features an onboard edge-computing AI model trained for high-accuracy, real-time seagrass identification.
Field deployments conducted across the Penghu archipelago demonstrate the successful synergy of airborne and surface automation. The multispectral UAV provides expansive spatial overviews, while the AI-USV performs close-range, high-resolution under-surface verification, bathymetric profiling, and localized water quality mapping. This integrated approach successfully generated precise spatial distribution maps of seagrass coverage alongside synchronized DO and environmental data. Supported by systematic, long-term robotic surveys, this framework delivers an empirical baseline for marine habitat management and ecological restoration strategies. While the initial development focuses on establishing foundational data for blue carbon accounting in Taiwan, the ultimate trajectory of this platform is to scale into a universal, high-efficiency tool for monitoring broader nearshore ecosystems.
Comment
To browse or post comments, you must log in.Log in