雷达探测中,目标识别是关键问题。借助人工智能技术提高雷达目标识别性能是近几年的研究热点。针对神经网络运用于雷达目标识别时存在的网络训练困难和数据量欠缺等问题,将网络集成的思想引入高分辨雷达目标识别, 通过将传统目标识别方法与网络集成技术结合,降低网络的复杂度,减少对数据量的要求,并利用角域划分,建立分角域网络集成高分辨雷达目标识别架构,确保系统在低数据量下仍可获得较好的识别效果。雷达实测数据证明了该方法的有效性。

Radar automatic target recognition(ATR)is a difficult challenge in radar detection. Radar ATR based on artificial intelligence has become one of the research hotspots all over the world in recent years. A new architecture is proposed in this paper to solve the problem of poor data and network complexity by combination of traditional high range resolution profile(HRRP) radar ATR methods and neural network ensemble. By using the target angle information, recognition performance is improved while the workload of manual annotation and the train data are reduced too, and the processing flow is proposed. The results of radar real data verify that the new method is effective.

TN957. 52