Large-scale multi-class SAR image target detection dataset-1.0

Large-Scale Multi-Class SAR Target Detection Dataset-1.0 (MSAR-1.0) consists of a total of 28,449 detection slices, utilizing data from the HISEA-1 satellite and the Gaofen-3 satellite.The MSAR-1.0 dataset includes four polarizations: HH, HV, VH, and VV. The dataset covers diverse scenarios such as airports, ports, nearshore areas, islands, open seas, and urban areas. The target types include four categories: aircraft, oil tanks, bridges, and ships, comprising 1,851 bridges, 39,858 ships, 12,319 oil tanks, and 6,368 aircraft.The slice size of the MSAR-1.0 dataset is 256 × 256 pixels, with some bridge slices sized at 2048 × 2048 pixels. The images are three-channel grayscale images in 24-bit depth JPG format. The annotation format is XML, recording the target type and location information, where the location information consists of Xmin, Xmax, Ymin, and Ymax. Examples of the slice label files in the MSAR-1.0 dataset are shown in Figures 2 and 3 of the dataset usage instructions. The format is compatible with mainstream detection networks such as the YOLO series, PolarMask, SSD, and Faster-RCNN.To meet further research needs, the MSAR-1.0 dataset also provides several large-scene images captured by the HISEA-1 satellite, featuring target types including aircraft, oil tanks, bridges, ships, and airport runways. The annotation format is XML, recording the target type and location information, where the location information consists of Xmin, Xmax, Ymin, and Ymax.