Advances and challenges in remote sensing for grass species classification in coastal wetlands

• Remote sensing data sources summarized, with strengths and limitations. • Tidal fluctuations and cloud cover noted as key challenges in wetland monitoring. • Time-series phenological metrics emphasized for grass species differentiation. • Object-based UAV data, multi-source fusion, and deep learning explored as promising tools. Coastal wetlands, vital for carbon sequestration and climate change mitigation, attracting global attention due to their unique ecological and socio-economic benefits, like blue carbon merit. Monitoring the distribution of grass species in these ecosystems provides critical insights into wetland health, productivity, and the ecological impacts. Remote sensing has emerged as an effective tool for large-scale, long-term, and near real-time monitoring of grass species dynamics, particularly in areas experiencing rapid changes, such as invasive species hotspots. Recent advancements in phenological analysis and deep learning have significantly improved remote sensing classification accuracy, facilitating more detailed grass species mapping. These improvements have contributed to the development of comprehensive monitoring frameworks. This review synthesizes the current state of grass species classification in coastal wetlands, focusing on key data sources, including passive optical remote sensing, synthetic aperture radar (SAR), and LiDAR. It also examines classification methodologies, such as mixed pixel interpretation, pixel-based classification, and object-based segmentation, while emphasizing recent progress in multi-source data integration, classification features, and phenological analysis. Furthermore, promising directions include the integration of UAV-based hyperspectral data and the application of deep learning within cloud computing platforms. By summarizing recent advancements, challenges, and future research directions, this review contributes to the development of more effective approaches for remote sensing-based grass species classification in coastal wetlands.