Research Article
Multi-view Polarization HRRP Target Recognition Based on Convolutional Neural Network
@INPROCEEDINGS{10.1007/978-3-030-67514-1_56, author={Mengyuan Ma and Kai Liu and Xiling Luo and Tao Zhang}, title={Multi-view Polarization HRRP Target Recognition Based on Convolutional Neural Network}, proceedings={IoT as a Service. 6th EAI International Conference, IoTaaS 2020, Xi’an, China, November 19--20, 2020, Proceedings}, proceedings_a={IOTAAS}, year={2021}, month={1}, keywords={Polarization radar Target recognition Convolutional neural network High-resolution range profile Scattering center}, doi={10.1007/978-3-030-67514-1_56} }- Mengyuan Ma
Kai Liu
Xiling Luo
Tao Zhang
Year: 2021
Multi-view Polarization HRRP Target Recognition Based on Convolutional Neural Network
IOTAAS
Springer
DOI: 10.1007/978-3-030-67514-1_56
Abstract
This paper proposes a multi-view polarization high-resolution range profile (HRRP) target recognition method based on convolutional neural network (CNN-based MVPHRRP), which combines high-resolution technology with polarization technology to extract radar signal features. Using the feature layer fusion method, the intensity of scattering centers, the ratios of odd and even scattering extracted by Pauli decomposition constitute a three-dimensional feature tensor. On the basis of retaining the time-domain distribution characteristics, the multi-polarization characteristics and the target’s structural composition are fitted. Then we build a CNN for radar target recognition. The simulation results show that the use of CNN-based MVPHRRP for radar target recognition has a good effect, and the classification accuracy is less affected by the signal-to-noise ratio (SNR). Increasing the number of multi-views plays a positive role in improving the recognition performance, and the introduction of multi-view multi-polarization information can effectively increase the average recognition probability of target recognition.
