A dataset of object detection of tailings ponds in Henan Province, China, 2016-2021

Tailings pond is a necessary infrastructure for mining, and also a risk source of man-made debris flow with high potential energy. There is a danger of dam break, which may cause significant personnel, economic losses and environmental disasters. The object detection of tailings pond in remote sensing image is to identify and locate the tailings pond in remote sensing image. It plays an important role in finding out the distribution and quantity of tailings pond and conducting dynamic monitoring. It is a basic work of emergency supervision of tailings pond. Compared with the traditional methods, the method of using deep learning to detect tailings pond objects in remote sensing images has significantly improved in accuracy, stability and efficiency, but requires high-quality training dataset. At present, the open-source object detection dataset of tailings pond is lack, and it is marked with horizontal bounding box, so there are some limitations in object detection on remote sensing images. Therefore, we have constructed a object detection dataset of tailings pond in Henan Province, China, and open it for sharing. This dataset has the following characteristics: (1) The current largest object detection dataset of tailings pond contains 1183 slices and 1728 instances; (2) The dataset provides a total of four different years of sample data in 2016, 2018, 2020 and 2021, all based on domestic optical satellite images; (3) The object is labeled with oriented bounding box, so the image background interference is less. The dataset can be used to conduct technical research on the development of object detection model of deep learning tailings pond and automatic and intelligent detection of tailings pond, which is of great significance for promoting the development of automatic extraction technology of tailings pond and safety supervision of tailings pond.