Moving and Stationary Target Acquisition and Recognition (MSTAR)

该数据集名为MSTAR，包含了6,874张不同类型的军事车辆图像，这些图像是通过桑迪亚X波段雷达收集的。该数据集专为自动目标识别（ATR）任务而设计，涵盖了10种不同的军事车型。数据集在训练（15°）和测试（17°）中收集了不同角度的图像。所有图像都被中心裁剪至88x88像素，并将像素值限制在[0, 1]的范围内。该数据集的规模为6,874张图像，涉及的任务是自动目标识别（ATR）。

This dataset, named MSTAR, contains 6,874 images of various military vehicles collected via Sandia X-band radar. Purpose-built for automatic target recognition (ATR) tasks, it covers 10 distinct military vehicle models. Images were gathered at varying depression angles for the training (15°) and testing (17°) subsets respectively. All images are centrally cropped to 88×88 pixels, with pixel values normalized to the range [0, 1]. With a total of 6,874 images, this dataset is dedicated to automatic target recognition (ATR) tasks.

搜集汇总

背景与挑战

背景概述

MSTAR-Active-Learning是一个Python包，用于在MSTAR数据集上实现基于图的主动学习算法，支持SAR图像的半监督分类。该数据集包含预处理后的幅度和相位图像，适用于机器学习模型的训练和测试。