微小目标图像多普勒频谱识别研究

Research on Doppler spectrum recognition of tiny target images

为提高雷达对微小目标的识别精准度,基于深度信念网络对微小目标图像多普勒频谱识别方法展开研究。首先利用深度信念网络挖掘微小目标图像中的深层抽象特征,然后将对目标图像的多普勒频谱识别转换为对强杂波环境下微弱周期信号的识别,从而提高目标图像的信噪比。基于挖掘到的深层次特征,将目标图像周期信号转换为小频率信号,从而完成识别。仿真结果表明:该方法受强杂波影响较小,使得PSD图中峰值更加清晰明了,可在短时间内获得理想的识别结果,且优化了图像信噪比,从而实现了对微小目标的精准识别。

In order to improve the accuracy of radar recognition of tinny targets,this paper studies the doppler spectrum recognition method of small target image based on deep belief network.Firstly,deep belief network is used to mine the deep abstract features in micro-target image,and then the Doppler spectrum recognition of target image is transformed into weak periodic signal recognition in strong clutter environment,so as to improve the signal-to-noise ratio of target image.Based on the deep features mined,the periodic signal of the target image is converted into a small frequency signal to complete the recognition.The simulation results show that the method is less affected by the strong clutter,which makes the peak value in the PSD image clearer,and can obtain the ideal recognition results in a short time.In addition,the SNR of the image is optimized,so as to realize the accurate recognition of small targets.