融合LightGBM的ResNeXt气象目标细粒度识别方法

doi:10.12305/j.issn.1001-506X.2024.12.12

摘要/Abstract

摘要：

为精确识别气象目标与混杂其中的非气象目标，提出一种融合轻量级梯度提升机(light gradient boosting machine，LightGBM)与残差网络的残差网络(residual network of residual network: next generation, ResNeXt) 的气象目标识别方法。首先，制作块状样本数据集，以此数据集为驱动，建立以ResNeXt为基础的气象目标识别网络模型，实现以块状数据样本为识别单位的气象目标粗粒度识别，识别精度可达99.6%以上；然后，再将此粗粒度结果与参考数据的差异值纳入LightGBM分类器，得到以雷达采样单元为识别单位的细粒度识别结果。结合实际观测数据，证明所提方法融合了LightGBM细粒度识别与ResNeXt高精度识别的能力，能够完成气象目标与杂波的判别，判别结果与参考结果高度一致。结合实际观测数据，证明所提方法融合了LightGBM细粒度识别与ResNeXt高精度识别的能力，能够完成气象目标与杂波的判别，判别结果与参考结果高度一致。

关键词: 气象雷达, 气象目标识别, 残差网络, 轻量级梯度提升机, 融合, 深度学习

Abstract:

In order to accurately identify meteorological targets and non-meteorological targets mixed in the meteorological targets, a meteorological target recognition method that combines light gradient boosting machine (LightGBM) and residual network of residual network: next generation(ResNeXt) network is proposed. Firstly, a block sample dataset is produced, and this dataset is used as a driver to establish a meteorological target recognition network model based on the ResNeXt. This model realizes a coarse-grained recognition of meteorological targets with block data samples as the recognition unit with the recognition accuracy of more than 99.6%. Then, the difference between the coarse-grained recognition results and the reference data is included in the LightGBM classifier so the fine-grained recognition results with the radar sampling unit as recognition unit are obtained. Combined with the actual observation data, it is proved that the proposed method combines the ability of fine-grained recognition of LightGBM with the ability of high-precision recognition of ResNeXt, which can distinguish meteorological targets from clutter, and the distinguishing results are highly consistent with the reference results.

Key words: eteorological radar, meteorological target recognition, residual network, light gradient boosting machine (LightGBM), fusion, deep learning

中图分类号:

TN957.5

欧阳彤, 汪玲, 朱岱寅, 李勇. 融合LightGBM的ResNeXt气象目标细粒度识别方法[J]. 系统工程与电子技术, 2024, 46(12): 4034-4043.

Tong OUYANG, Ling WANG, Daiyin ZHU, Yong LI. Fine-grained recognition method for meteorological targets in ResNeXt fused with LightGBM[J]. Systems Engineering and Electronics, 2024, 46(12): 4034-4043.

图/表 17

图1

图2

图3

图4

图5

表1

图6

图7

图8

图9

图10

图11

图12

图13

图14

图15

图16

参考文献 31

1	张培昌, 顾松山, 王振会, 等. 气象万千探本索源——南京信息工程大学"大气探测学科"发展历程回顾与展望[J]. 大气科学学报, 2021, 44 (2): 165- 177.
	ZHANG P C , GU S S , WANG Z H , et al. Review and prospect of atmospheric sounding in Nanjing university of information science and technology[J]. Transactions of Atmospheric Sciences, 2021, 44 (2): 165- 177.
2	BINETTI M S , CAMPANALE C , MASSARELLI C , et al. The use of weather radar data: possibilities, challenges and advanced applications[J]. Earth, 2022, 3 (1): 157- 171. doi: 10.3390/earth3010012
3	ZHAO C H , ZHANG Y J , ZHENG D , et al. An improved hydrometeor identification method for X-band dual-polarization radar and its application for one summer Hailstorm over Northern China[J]. Atmospheric Research, 2020, 245 (1): 105075.
4	ZHANG Q , ZHANG P , ZHANG Y , et al. Development of dual-polarization weather radar products display platform[J]. Journal of Physics: Conference Series, 2021, 1812 (1): 012016. doi: 10.1088/1742-6596/1812/1/012016
5	宋文婷, 李昀英, 黄浩, 等. 基于S波段双偏振雷达资料的降水粒子类型识别算法及应用[J]. 大气科学学报, 2021, 44 (2): 209- 218.
	SONG W T , LI Y Y , HUANG H , et al. Precipitation particle type recognition algorithm based on S-band dual polarization radar data and its application[J]. Transactions of Atmospheric Sciences, 2021, 44 (2): 209- 218.
6	李海, 任嘉伟, 尚金雷. 一种基于模糊神经网络-模糊C均值聚类的双偏振气象雷达降水粒子分类方法[J]. 电子与信息学报, 2019, 41 (4): 809- 815.
	LI H , REN J W , SHANG J L . Hydrometeor classification method in dual-polarization weather radar based on fuzzy neural network-fuzzy C-means[J]. Journal of Electronics & Information Technology, 2019, 41 (4): 809- 815.
7	SOLOL Z , SZTURE J , ORELLANA-ALVER J , et al. The role of weather radar in rainfall estimation and its application in meteo-rological and hydrological modelling-a review[J]. Remote Sensing, 2021, 13 (3): 351. doi: 10.3390/rs13030351
8	MANIRAGUHA F, VODACEK A, NDASHIMYE E, et al. Ground clutter mitigation and insect signature detection for polarimetric C-band Doppler weather radar[C]//Proc. of the IEEE Global Humanitarian Technology Conference, 2021: 289-296.
9	WANG Y , WU D , YU Q H , et al. A weather signal detection algorithm based on EVD in elevation for airborne weather radar[J]. Digital Signal Processing, 2021, 116, 103118. doi: 10.1016/j.dsp.2021.103118
10	WANG Y, WU D, ZHU D Y, et al. Detection of weather signal in ground clutter for airborne weather radar based on spatial processing in elevation[C]//Proc. of the IEEE 5th International Conference on Signal and Image Processing, 2020: 655-659.
11	SASIDHARAN S , ANANDAN V K , MISHRA S , et al. Identification, characterization and removal of anomalous propagation and ground clutter echoes using polarimetric Doppler weather radar products[J]. Journal of Electromagnetic Waves and Applications, 2023, 37 (2): 176- 189. doi: 10.1080/09205071.2022.2118085
12	STRAKA J M. Hydrometeor fields in a supercell storm as deduced from dual-polarization radar[C]//Proc. of the 18th Conference on Severe Local Storms, 1996: 551-554.
13	GOLBON-HAGHIGHI M H , ZHANG G . Detection of ground clutter for dual-polarization weather radar using a novel 3D discriminant function[J]. Journal of Atmospheric and Oceanic Technology, 2019, 36 (7): 1285- 1296. doi: 10.1175/JTECH-D-18-0147.1
14	MA J L , HU Z Q , YANG M L , et al. Improvement of X-band polarization radar melting layer recognition by the Bayesian method and ITS impact on hydrometeor classification[J]. Advances in Atmospheric Sciences, 2020, 37 (1): 105- 116. doi: 10.1007/s00376-019-9007-z
15	JATAU P , MELNIKOV V , YU T Y . A machine learning approach for classifying bird and insect radar echoes with S-band polarimetric weather radar[J]. Journal of Atmospheric and Oceanic Technology, 2021, 38 (10): 1797- 1812.
16	MAKINEN T , RITVANEN J , PULKKINEN S , et al. Bayesian classification of nonmeteorological targets in polarimetric Doppler radar measurements[J]. Journal of Atmospheric and Oceanic Technology, 2022, 39 (10): 1561- 1578. doi: 10.1175/JTECH-D-21-0177.1
17	欧阳彤, 汪玲, 朱岱寅, 等. 基于LightGBM的气象目标分类技术[J]. 雷达科学与技术, 2023, 21 (6): 621- 629. doi: 10.3969/j.issn.1672-2337.2023.06.005
	OUYANG T , WANG L , ZHU D Y , et al. Meteorological target classification technology based on LightGBM[J]. Radar Science and Technology, 2023, 21 (6): 621- 629. doi: 10.3969/j.issn.1672-2337.2023.06.005
18	LU Y P, KUMAR J. Convolutional neural networks for hydrometeor classification using dual polarization Doppler radars[C]// Proc. of the International Conference on Data Mining Workshops, 2019: 288-295.
19	YU Q H, WU D, ZHU D Y, et al. CNN-based weather signal detection algorithm for airborne weather radar[C]//Proc. of the IEEE 5th International Conference on Signal and Image Processing, 2020: 660-664.
20	高涌荇, 王旭东, 汪玲, 等. 基于RCNN的双极化气象雷达天气信号检测[J]. 系统工程与电子技术, 2022, 44 (11): 3380- 3387. doi: 10.12305/j.issn.1001-506X.2022.11.12
	GAO Y X , WANG X D , WANG L , et al. Weather signal detection for dual polarization weather radar based on RCNN[J]. Systems Engineering and Electronics, 2022, 44 (11): 3380- 3387. doi: 10.12305/j.issn.1001-506X.2022.11.12
21	XIE S, GIRSHICK R, DOLLAR P, et al. Aggregated residual transformations for deep neural networks[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 1492-1500.
22	KE G L, MENG Q, FINLEY T, et al. LightGBM: a highly efficient gradient boosting decision tree[C]//Proc. of the Advances in Neural Information Processing Systems, 2017: 3149-3157.
23	盖荣丽, 蔡建荣, 王诗宇, 等. 卷积神经网络在图像识别中的应用研究综述[J]. 小型微型计算机系统, 2021, 42 (9): 1980- 1984. doi: 10.3969/j.issn.1000-1220.2021.09.030
	GAI R L , CAI J R , WANG S Y , et al. Research review on image recognition based on deep learning[J]. Journal of Chinese Computer Systems, 2021, 42 (9): 1980- 1984. doi: 10.3969/j.issn.1000-1220.2021.09.030
24	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 770-778.
25	SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[EB/OL]. [2023-08-12]. https://arxiv.org/pdf/1409. 1556.
26	SZEGEDY C, LIU W, JIA Y Q, et al. Going deeper with convolutions[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2015.
27	LOFFE S, SZEGEDY C. Batch normalization: accelerating deep network training by reducing internal covariate shift[C]//Proc. of the International Conference on Machine Learning, 2015: 448-456.
28	ABIEN F A. Deep learning using rectified linear units (ReLU)[EB/OL]. [2023-08-12]. https://arxiv.org/pdf/1803. 08375.
29	FRIEDMAN J H . Greedy function approximation: a gradient boosting machine[J]. Annals of Statistics, 2001, 29 (5): 1189- 1232. doi: 10.1214/aos/1013203450
30	PASZKE A, GROSS S, MASSAA F, et al. Pytorch: an imperative style, high performance deep learning library[C]//Proc. of the 33rd Conference on Neural Information Processing Systems, 2019: 8026-8037.
31	LECUN Y , BOTTOU L , BENGIO Y , et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86 (11): 2278- 2324. doi: 10.1109/5.726791

参数	数值
学习率	0.01
动量值	0.9
衰减权重	0.001

[1]	蔡伟, 王鑫, 蒋昕昊, 杨志勇, 陈栋. 基于解耦的小样本目标检测方法研究[J]. 系统工程与电子技术, 2024, 46(9): 2941-2950.
[2]	郭宝锋, 焦丽婷, 李胜, 朱晓秀, 薛东方, 孙慧贤. 空间目标T/R-R型雷达多视角融合成像算法[J]. 系统工程与电子技术, 2024, 46(9): 3019-3030.
[3]	陈晓萱, 徐书文, 胡绍海, 马晓乐. 基于卷积与自注意力的红外与可见光图像融合[J]. 系统工程与电子技术, 2024, 46(8): 2641-2649.
[4]	应文健, 程雨森, 王旋, 孙世岩. 基于研制阶段数据融合的舰炮制导弹药测试性评估方法[J]. 系统工程与电子技术, 2024, 46(8): 2730-2737.
[5]	安孟昀, 殷加鹏, 黄建开, 李永祯, 王雪松. 一种基于谱连续的极化-多普勒气象雷达信号重构方法[J]. 系统工程与电子技术, 2024, 46(7): 2285-2293.
[6]	汪强龙, 高晓光, 吴必聪, 胡子剑, 万开方. 受限玻尔兹曼机及其变体研究综述[J]. 系统工程与电子技术, 2024, 46(7): 2323-2345.
[7]	孙先涛, 江汪洋, 陈文杰, 陈伟海, 智亚丽. 基于感兴趣区域的物体抓取位姿检测[J]. 系统工程与电子技术, 2024, 46(6): 1867-1877.
[8]	王全全, 顾志豪, 吴城坤, 宛汀. 基于HHO-SVM的抗SSDF攻击协作频谱感知方法[J]. 系统工程与电子技术, 2024, 46(6): 2146-2154.
[9]	徐奕禹, 陈长风, 袁雪林, 陈正坤, 周志健. 基于高度计辅助的GNSS欺骗干扰检测[J]. 系统工程与电子技术, 2024, 46(5): 1484-1492.
[10]	吴钇达, 王彩云, 王佳宁, 李晓飞. 基于ISVM-DS的红外多传感器融合识别方法[J]. 系统工程与电子技术, 2024, 46(5): 1555-1560.
[11]	李海, 田众, 钱君. 基于ECOC平衡随机森林的雷达降水粒子分类[J]. 系统工程与电子技术, 2024, 46(5): 1599-1606.
[12]	陈雪梅, 刘志恒, 周绥平, 余航, 刘彦明. 基于HRNet的高分辨率遥感影像道路提取方法[J]. 系统工程与电子技术, 2024, 46(4): 1167-1173.
[13]	邱建杰, 蔡益朝, 李浩, 黄权印. 基于动态估计反馈的灰色理论航迹关联算法[J]. 系统工程与电子技术, 2024, 46(4): 1401-1411.
[14]	邹涵, 张天骐, 马焜然, 杨宗方. 基于多特征融合的MIMO-OFDM系统单-混信号调制识别算法[J]. 系统工程与电子技术, 2024, 46(4): 1456-1465.
[15]	张天文, 张晓玲, 邵子康, 曾天娇. 基于掩模注意型交互的SAR舰船实例分割[J]. 系统工程与电子技术, 2024, 46(3): 831-838.