"Road Segmentation Dataset for Open-Pit Iron Mines"

"This dataset comprises over 1,300 high-resolution images along with their corresponding pixel-level annotation labels, specifically designed for road segmentation in open-pit iron mines. The imagery was captured by satellite over a large-scale, fully operational iron mine located in northern China, ensuring that the data reflects real-world conditions and complexities. Each image has been uniformly cropped and formatted, enabling seamless integration into standard deep learning pipelines. This well-structured dataset supports a range of computer vision tasks, including road extraction and semantic segmentation. It is particularly suitable for training, validation, and testing phases of model development, offering a reliable foundation for advancing algorithms in challenging industrial environments. With its precise annotations and consistent preprocessing, the dataset serves as a valuable resource for researchers and practitioners working on remote sensing applications, infrastructure monitoring, and autonomous navigation within mining contexts."