论文部分内容阅读
为了利用雷达信号的时频信息以提高雷达目标识别性能,提出了一种基于分数阶Fourier变换(FrFT)的雷达目标识别方法。利用FrFT提取目标高分辨距离像的时频特征。从类可分性角度,求取多个最优变换,利用主分量分析方法对距离像的FrFT进行特征降维,并使用神经网络进行分类识别。最后,采用D-S证据理论,对多个最优变换的识别结果进行决策融合。仿真结果证明了该方法的合理性和可行性。
In order to utilize the time-frequency information of radar signal to improve the radar target recognition performance, a radar target recognition method based on fractional Fourier transform (FrFT) is proposed. Using FrFT to extract time-frequency features of target high resolution range images. From the viewpoint of class separability, a number of optimal transformations are obtained. Principal component analysis is used to reduce the dimensionality of FrFT in range images and neural network is used to classify and recognize them. Finally, D-S evidence theory is used to fuse the recognition results of multiple optimal transformations. Simulation results show that the method is reasonable and feasible.