Automating the detection of hydrological barriers and fragmentation in wetlands using deep learning and InSAR

作者：Clara Hübinger, Etienne Fluet‐Chouinard, Gustaf Hugelius, Francisco J. Peña, Fernando Jaramillo · 发表于：Remote Sensing of Environment · 年份：2024 · DOI：10.1016/j.rse.2024.114314 · 被引用次数：16 · 研究领域：Flood Risk Assessment and Management、Soil erosion and sediment transport、Coastal wetland ecosystem dynamics

The loss of hydrological connectivity and fragmentation of natural wetlands are widespread drivers of wetland degradation. Understanding where and how natural connectivity is impaired is essential for managing, protecting and remediating these ecosystems. Wetland Interferometric Synthetic Aperture Radar (Wetland InSAR) can provide information on surface flow orientation in wetlands at a high spatial resolution, which can be used for the detection of hydrological barriers in the wetland. However, the broad application of this approach is constrained by the labour-intensive manual delineation of barriers based on mapped water levels. This study presents the first deep learning-based methodology for the automated detection of hydrological barriers. We trained a deep convolutional network to segment edge features of hydrological barriers in 22 image pairs captured by ALOS PALSAR-1 L-Band InSAR between 2006 and 2011. The training dataset consisted of manually labelled and delineated barriers showing abrupt changes in water surface elevation and wrapped interferograms with high coherence. The model was trained and tested on six wetland sites of varying fragmentation levels and wetland types in the United States, Cuba, Mexico, Colombia and Venezuela. Across these sites, the convolutional network detected hydrological barriers with up to 80% accuracy. The model performed particularly well in detecting linear hydrological barriers such as roads, levees and channels. Notably, we found ...