Wetlands Mapping with Deep ResU-Net CNN and Open-Access Multisensor and Multitemporal Satellite Data in Alberta’s Parkland and Grassland Region

Wetlands are a valuable ecosystem that provides various services to flora and fauna. This study developed compared deep shallow learning models for wetland classification across the climatically dynamic landscape of Alberta’s Parkland Grassland Natural Region. approach mapping entailed exploring multi-temporal (combination spring/summer fall months over 4 years—2017 202) multisensory (Sentinel 1 2 Advanced Land Observing Satellite, ALOS) data as input in predictive models. image consisted S1 dual-polarization vertical-horizontal bands, S2 near-infrared shortwave infrared ALOS-derived topographic wetness index. The explored ResU-Net (DL) model two models, namely random forest (RF) support vector machine (SVM). We observed significant increase average F1-score ResNet prediction (0.82) SVM RF 0.69 0.69, respectively. showed occurrence mixed pixels, particularly marshes swamps confused upland classes (such agricultural land). Overall, it was evident CNN predictions performed better than outcome this demonstrates potential exploiting open-access satellite imagery generate credible products large landscapes.