MIS Check-Dam Dataset

MIS Check-Dam Dataset是由印度理工学院孟买分校创建的一个公开数据集，专注于农业用途中的人工灌溉结构——水坝的检测与映射。该数据集包含1082个水坝实例，分布在1037张640x640像素的卫星图像中，这些图像来自Google静态地图API，由领域专家进行标注。数据集的创建旨在简化灌溉普查过程，解决传统方法耗时、昂贵且易出错的问题。此外，该数据集还支持对象检测和实例分割任务，为深度学习在遥感图像处理中的应用提供了重要资源。

The MIS Check-Dam Dataset is a public dataset developed by the Indian Institute of Technology Bombay, focusing on the detection and mapping of dams—artificial irrigation structures utilized in agricultural practices. It comprises 1082 dam instances distributed across 1037 satellite images with a resolution of 640×640 pixels. These images were acquired via the Google Static Maps API and annotated by domain experts. The dataset was created to streamline irrigation census workflows and mitigate the limitations of traditional approaches, which are typically time-consuming, cost-intensive, and susceptible to errors. Furthermore, the dataset supports both object detection and instance segmentation tasks, serving as a critical resource for advancing deep learning applications in remote sensing image processing.