TY - GEN
T1 - Wideband Radar Target Recognition Based on Polarization Features and Double Layer K-LightGBM
AU - Deng, Bowen
AU - Lang, Ping
AU - Fu, Xiongjun
AU - Dong, Jian
AU - Ma, Zhifeng
AU - Cui, Zongding
N1 - Publisher Copyright:
© 2021 IEEE.
PY - 2021
Y1 - 2021
N2 - To improve the shape-similar target recognition performance, a polarization features-based double layer K-LightGBM model is proposed in this paper. First, a target features dataset is built, which is based on polarization invariants and two kinds of polarization decompositions of targets. Then, a double layer K-LightGBM model is set up by improving StackNet with K LightGBM stacked. Finally, the dataset is fed into the proposed model for training and testing. The experimental results show that the proposed method in this paper has better performance in terms of generalization and denoising, compared to many existing state-of-the-art methods, including LightGBM.
AB - To improve the shape-similar target recognition performance, a polarization features-based double layer K-LightGBM model is proposed in this paper. First, a target features dataset is built, which is based on polarization invariants and two kinds of polarization decompositions of targets. Then, a double layer K-LightGBM model is set up by improving StackNet with K LightGBM stacked. Finally, the dataset is fed into the proposed model for training and testing. The experimental results show that the proposed method in this paper has better performance in terms of generalization and denoising, compared to many existing state-of-the-art methods, including LightGBM.
KW - LightGBM
KW - StackNet
KW - polarization decomposition
KW - polarization invariants
KW - radar target recognition
UR - http://www.scopus.com/inward/record.url?scp=85181049718&partnerID=8YFLogxK
U2 - 10.1109/Radar53847.2021.10027876
DO - 10.1109/Radar53847.2021.10027876
M3 - Conference contribution
AN - SCOPUS:85181049718
T3 - Proceedings of the IEEE Radar Conference
SP - 1293
EP - 1296
BT - 2021 CIE International Conference on Radar, Radar 2021
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2021 CIE International Conference on Radar, Radar 2021
Y2 - 15 December 2021 through 19 December 2021
ER -