The first Spatiotemporal dataset of open-pit mines in the middle reaches of the Yangtze river

Open-pit mine (OM) interpretation and mapping are essential for quantifying the environmental impacts of mineral resource extraction and have received increasing attention in the field of sustainable development. OMs are characterized by diverse geometric shapes, spectral variability, and irregular distribution. However, no long-term sample datasets or mapping products are currently available for OMs in the middle reaches of the Yangtze River. In this study, we used Sentinel-2 satellite imagery and deep learning techniques to construct LT-OM-MID, the first 10-meter-resolution, long-term elemental sample dataset for OMs in this region, covering the years 2018, 2020, 2022, and 2024. LT-OM-MID includes 350 typical OM case areas (TOMCA), a total of 13,275 OM samples, and corresponding annual mapping products. On TOMCA, the OM class achieved a maximum overall accuracy (ACC) of 61.92% based on partial ground-truth validation. Quantitative and qualitative comparisons with visual interpretation and geological survey data confirm the robustness and efficiency of the mapping framework. LT-OM-MID provides crucial data support for government efforts in OM monitoring, management, and environmental restoration.