MSTAR
收藏资源简介:
MSTAR数据集是由法国国防部信息管理处创建的,用于训练和测试SAR图像中的自动目标检测模型。该数据集包含10类军事地面车辆的SAR图像,共计21600对目标签名和阴影掩码。数据集的创建过程涉及使用MOCEM模拟器生成合成SAR图像,并将其与真实背景图像结合。MSTAR数据集主要应用于自动目标检测领域,旨在解决从合成数据到真实测量数据的一般化问题。
The MSTAR dataset was developed by the Information Management Office of the French Ministry of Defense for training and testing automatic target detection models on synthetic aperture radar (SAR) images. It contains SAR images of 10 classes of military ground vehicles, with a total of 21,600 pairs of target signatures and shadow masks. The dataset was created by generating synthetic SAR images using the MOCEM simulator and combining them with real background images. The MSTAR dataset is mainly applied in the field of automatic target detection, aiming to solve the generalization problem from synthetic data to real measured SAR data.
- 1FewSAR: A Few-shot SAR Image Classification Benchmark陆军工程大学指挥控制工程学院 · 2023年



