合成孔径雷达SAR技术具有穿透力强、识别精度高等优点，在海上船舶目标识别、军事侦察等领域发挥着非常重要的作用。由于合成孔径雷达本身的光学特性以及图像采集过程中的噪声干扰，SAR图像存在着类别不平衡等问题。本文研究不平衡SAR图像的识别技术，从SAR图像特点出发，结合深度网络学习算法，建立不平衡SAR图像的快速识别模型，具有重要的应用价值。
Marine synthetic aperture radar (SAR) technology has the advantages of strong penetration and high identification accuracy, and it plays a very important role in Marine ship target identification, military reconnaissance and other fields. Due to the optical characteristics of SAR and noise interference in the process of image acquisition, there are some problems such as category imbalance in SAR images. The research direction of this paper is the recognition technology of unbalanced SAR images. Based on the characteristics of SAR images, combined with deep network learning algorithm, the fast recognition model of unbalanced SAR images is established, which has important application value.
2023,45(5): 174-177 收稿日期：2022-08-23
DOI：10.3404/j.issn.1672-7649.2023.05.034
分类号：TN95
基金项目：湖南省教育厅科研项目 (21C0137）
作者简介：向诚(1973-),男,硕士,副教授,主要从事人工智能与嵌入式技术研究
