基于深度学习的雷达目标识别算法评估系统设计

The Design of Radar Target Recognition Algorithm Evaluation System Based on Deep Learning

摘要:

在雷达成像装备试验中，传统真实场景测试方法构建难度大、场景有限、试验风险高，急需解决目标识别算法测试不充分、评估不全面的问题本文针对现有问题设计了一套目标识别算法测试系统，可提供合成孔径雷达图像和逆合成孔径雷达图像的处理、标注以及目标识别算法的自动运行、环境配置、性能评估等功能。相较于传统的试验方法，该系统具备成本低、试验时间短、可控性强、可扩展等优点。

Abstract:

In the radar imaging equipment test, the traditional real scene test method is difficult to construct, the scene is limi-ted, and the test risk is high, so it is urgent to solve the problems of insufficient testing and incomplete evaluation of the target recognition algorithm. Aiming at the existing problems, this paper designs a test system for the target recognition algorithm, which can provide the processing and labeling of SAR image and inverse SAR image, as well as the automatic operation, environment configuration and performance evaluation of target recognition algorithm. Compared with the traditional test method, the system has the advantages of low cost, short test time, strong controllability and extensibility.