A method of HRRP denoising and recognition Based on CAE

DOI:	10.11809/bqzbgcxb2023.01.028
Keywords:	high resolution range profiles; auto encoder; convolutional neural network; radar target; ship
Abstract:	Aiming at the problem of noise pollution in the research of HRRP recognition, this paper proposes a HRRP recognition method based on convolutional auto encoder (CAE). Combined with the reconstruction function of CAE and with the classification performance of convolutional neural network (CNN), the method takes the noiseless data as labels and uses CAE to learn the noise characteristics of noisy HRRP, so as to achieve the denoising and reconstruction of HRRP and then uses CNN to identify the reconstructed HRRP. The simulation results show that, under the noise environment of a peak signal to noise ratio of 10, 20 and 40 dB, the recognition accuracy of HRRP can reach 76.48%, 95.14% and 9833% respectively, which reduces the influence of noise on HRRP recognition to a certain extent and ensures the recognition accuracy.
Published:	2023-01-28
PDF HTML