Guest Editor: Jie Chen, Zhixiang Huang, Runfan Xia 

The large-scale multi-class SAR image target detection dataset-1.0 (MSAR-1.0) has a total of 28,449 detection slices, and the source data was collected from the HISEA-1 satellite and the Gaofen-3 satellite.

The polarization modes of MSAR-1.0 dataset include HH, HV, VH and VV. The dataset scenarios include airports, ports, inshore, islands, offshore, urban areas, etc. The classes include aircraft, oil tank, bridge and ship, comprising 1,851 bridges, 39,858 ships, 12,319 oil tanks and 6,368 aircrafts. Figure 1 is a few examples of slices from the MSAR-1.0 dataset.

Figure 1. A few examples of slices from the MSAR-1.0 dataset. (a,b) are oil tank; (c,d) are ship; (e,f) are bridge; (g,h) are aircraft

The slice size of MSAR-1.0 dataset is 256*256 pixels, and the slice size of some bridges is 2048*2048 pixels. The format is three-channel gray image and 24-bit depth JPG. The annotation format is XML, and the target class and location information are recorded. The location information consists of Xmin, Xmax, Ymin and Ymax. Examples of MSAR-1.0 dataset slicing label files are shown in figures 2 and 3 of the dataset usage instructions below. The label file conforms to the format requirements of mainstream detection networks such as Yolo series, PolarMask, SSD, and Faster-RCNN.

To meet more scientific needs, the MSAR-1.0 dataset provides multiple HISEA-1 images of large scenes, including aircrafts, oil tanks, bridges, ships, airport runway, etc. The annotation format is XML, and the target class and location information are recorded. The location information consists of Xmin, Xmax, Ymin and Ymax. Examples of large scene image label files of MSAR-1.0 dataset are shown in figure 4 and 5 of dataset usage instructions.

For the data set usage instructions, "Large-scale Multi-class SAR Image Target Detection Dataset-1.0 Usage Instruction".

Reference format of MSAR-1.0 dataset:

[1] Jie Chen, Zhixiang Huang, Runfan Xia, Bocai Wu, Lei Sheng, Long Sun, Baidong Yao. Large-scale multi-class SAR image target detection dataset-1.0[OL]. Journal of Radars, 2022. https://radars.ac.cn/web/data/getData?dataType=MSAR.

[2] Xia, R.; Chen, J.; Huang, Z.; Wan, H.; Wu, B.; Sun, L.; Yao, B.; Xiang, H.; Xing, M. CRTransSar: A Visual Transformer Based on Contextual Joint Representation Learning for SAR Ship Detection. Remote Sensing. 2022, 14, 1488.