首页 > 论文 > 激光与光电子学进展 > 57卷 > 20期(pp:201011--1)

结合双流3D卷积和监控图像的降水临近预报

Precipitation Nowcasting Based on Dual-Flow 3D Convolution and Monitoring Images

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

针对大部分降水临近预报产品无法兼顾高覆盖率、高准确率及低成本的问题,提出一种基于室外监控图像和深度神经网络能预报未来1 h降水强度的方法。设计双流3D卷积神经网络来提取图像降雨信息的高维特征。该网络在低计算代价下自适应产生局部信息,并通过双损失函数从整体和局部统筹网络,提取降雨信息的时间特性和空间特性。实验结果表明,在降水强度预报领域,基于双损失函数的神经网络优于单损失函数。所提网络的误警率、命中率、临界成功指数、准确率在多数情况下优于其他模型。在模型效果可视化方面,所提网络能有效提取降水图像的特征信息。所提降水临近预报方法有能力进行精细且低成本的降水临近预报。

Abstract

At present, most of the precipitation nowcasting production is unable to consider the problems of high coverage, high accuracy, and low cost. Therefore, we herein propose a method based on outdoor monitoring images and deep neural network to forecast the rainfall intensity in the next 1 h. We design a dual-flow 3D convolutional neural network to extract high-dimensional features of rainfall information in images. The local information is adaptively generated at a low computational cost, and the temporal and spatial characteristics of rainfall information are extracted by the proposed network which integrates the whole network and the local network using a double loss function. The experimental results show that the neural network based on the dual loss function is better than that based on the single loss function in precipitation intensity forecasting. Percent of doom, false alarm rat, critical success index, and the accuracy of the proposed network are better than those of other models in most cases. In terms of visualization of the model effect, the proposed network can effectively extract the feature information of the precipitation images. Therefore, the proposed precipitation nowcasting method is capable of fine and low-cost precipitation prediction.

广告组1 - 空间光调制器+DMD
补充资料

中图分类号:P457.6

DOI:10.3788/LOP57.201011

所属栏目:图像处理

基金项目:中国博士后科学基金面上项目、海峡博士后交流资助计划、福州市科技局社会发展项目;

收稿日期:2019-12-24

修改稿日期:2020-02-25

网络出版日期:2020-10-01

作者单位    点击查看

杨素慧:福建农林大学计算机与信息学院, 福建 福州 350002
林志玮:福建农林大学计算机与信息学院, 福建 福州 350002福建农林大学林学院, 福建 福州350002福建农林大学林学博士后流动站, 福建 福州350002
赖绍钧:福州市气象局, 福建 福州 350014
刘金福:福建农林大学计算机与信息学院, 福建 福州 350002福建农林大学海峡自然保护区研究中心, 福建 福州 350002生态与资源统计福建省高校重点实验室, 福建 福州 350002

联系人作者:林志玮(cwlin@fafu.edu.cn); 刘金福(fjljf@126.com);

备注:中国博士后科学基金面上项目、海峡博士后交流资助计划、福州市科技局社会发展项目;

【1】Song Z J, He J X, Li X H, et al. Review of research progress in new generation satellite-borne precipitation products [J]. Meteorological Science and Technology. 2018, 46(4): 631-637.
宋子珏, 何建新, 李学华, 等. 星载降水测量雷达降水产品研究进展 [J]. 气象科技. 2018, 46(4): 631-637.

【2】Vu T M, Raghavan S V, Liong S Y, et al. Uncertainties of gridded precipitation observations in characterizing spatio-temporal drought and wetness over Vietnam [J]. International Journal of Climatology. 2018, 38(4): 2067-2081.

【3】Hu R H. Research on application of parallel computing in nowcasting system [D]. Guangzhou: South China University of Technology. 2015.
胡荣华. 并行计算在临近天气预报系统中的应用研究 [D]. 广州: 华南理工大学. 2015.

【4】Wang Y C, Kong R, Chen M X. Operational application progress of convective weather proximity prediction technology . C]∥The Fourth National Symposium on Disastrous Weather Forecast Technology, January 1, 2007, Beijing, China. Beijing: National Meteorological Center and Department of Prediction and Disaster Reduction, China Meteorological Administration. 2007, 35-37.
王迎春, 孔荣, 陈明轩. 对流天气临近预报技术的业务应用进展 . C]∥第四届全国灾害性天气预报技术研讨会, 1月1日, 2007, 北京, 中国. 北京: 中国气象局国家气象中心与预测减灾司. 2007, 35-37.

【5】Zhang Y Y. Spatial downscaling of TRMM precipitation data in the red river basin [D]. Kunming: Yunnan University. 2018.
张月圆. 红河流域TRMM降水数据空间降尺度研究 [D]. 昆明: 云南大学. 2018.

【6】Clark P, Roberts N, Lean H, et al. Convection-permitting models: a step-change in rainfall forecasting [J]. Meteorological Applications. 2016, 23(2): 165-181.

【7】Zhang Y, Wu S G, Zhang Y P, et al. Identification and effect verification of convective cloud precipitation in rainstorm processes based on SWAN mosaic products [J]. Meteorological Monthly. 2019, 45(2): 180-190.
张勇, 吴胜刚, 张亚萍, 等. 基于SWAN雷达拼图产品在暴雨过程中的对流云降水识别及效果检验 [J]. 气象. 2019, 45(2): 180-190.

【8】Xu W J, Su D B, Jia L L, et al. Summer drop size distribution characteristics in Beijing from laser-based optical disdrometer [J]. Laser & Optoelectronics Progress. 2012, 49(12): 120103.
徐文静, 苏德斌, 伽丽丽, 等. 利用激光降水粒子谱仪观测北京夏季雨滴谱特征 [J]. 激光与光电子学进展. 2012, 49(12): 120103.

【9】Gou Y B. The optimization and evaluation of quantitative precipitation estimation based on multi-radar mosaic [D]. Beijing: Chinese Academy of Meteorological Sciences. 2014, 7-9.
勾亚彬. 基于雷达组网拼图的定量降水估测算法优化及效果评估 [D]. 北京: 中国气象科学研究院. 2014, 7-9.

【10】Chen D D, Zhao J, Wang B L, et al. Comparison of cloud top height observation between FY-2F satellite and millimeter-wave cloud radar [J]. Meteorological Science and Technology. 2019, 47(3): 495-501.
陈冬冬, 赵静, 王柏林, 等. FY-2F卫星和毫米波云雷达云高观测的个例对比分析 [J]. 气象科技. 2019, 47(3): 495-501.

【11】Goyal S, Kumar A, Mohapatra M, et al. Satellite-based technique for nowcasting of thunderstorms over Indian region [J]. Journal of Earth System Science. 2017, 126(6): 79.

【12】Cheng C L, Chen M, Chen M X, et al. Comparative experiments on two high spatiotemporal resolution blending algorithms for quantitative precipitation nowcasting [J]. Acta Meteorologica Sinica. 2019, 77(4): 701-714.
程丛兰, 陈敏, 陈明轩, 等. 临近预报的两种高时空分辨率定量降水预报融合算法的对比试验 [J]. 气象学报. 2019, 77(4): 701-714.

【13】Fang T Z, Qin P Y, Wang L M, et al. High temporal and spatial resolution green vegetation coverage generation and its application in soil erosion monitoring [J]. Acta Ecologica Sinica. 2019, 39(15): 5679-5689.
方天纵, 秦朋遥, 王黎明, 等. 高时空分辨率植被覆盖获取方法及其在土壤侵蚀监测中的应用 [J]. 生态学报. 2019, 39(15): 5679-5689.

【14】Guo L, Lin Y T, Zhang Z H, et al. Mechanism of laser coloration of stainless steel and color prediction based on neural network [J]. Chinese Journal of Lasers. 2016, 43(11): 1102008.
郭亮, 林远添, 张震华, 等. 不锈钢激光着色机理及基于神经网络的颜色预测 [J]. 中国激光. 2016, 43(11): 1102008.

【15】Kurihata H, Takahashi T, Ide I, et al. Rainy weather recognition from in-vehicle camera images for driver assistance . [C]∥IEEE Proceedings. Intelligent Vehicles Symposium, June 6-8, 2005, Las Vegas, NV, USA. New York: IEEE. 2005, 205-210.

【16】Roser M, Moosmann F. Classification of weather situations on single color images . [C]∥2008 IEEE Intelligent Vehicles Symposium, June 4-6, 2008, Eindhoven, Netherlands. New York: IEEE. 2008, 798-803.

【17】Shi X J, Chen Z R, Wang H, et al. -09-19)[2019-12-23] . https:∥arxiv. 2015, org/abs/1506: 04214.

【18】Shi X J, Gao Z H, Lausen L, et al. Deep learning for precipitation nowcasting: a benchmark and a new model . [C]∥Proceedings of the 31st International Conference on Neural Information Processing Systems, December 4-9, 2017, Long Beach, California, USA. New York: Curran Associates. 2017, 5622-5632.

【19】2019-03-18)[2019-11-12] [EB/OL]. http:∥www.cma.gov.cn/root7/auto13139/201903/t20190319_517664.html. 2018.
年中国气候公报[EB/OL], (2019-03-18)[2019-11-12] . http:∥www.cma.gov.cn/root7/auto13139/201903/t20190319_517664.html. 2018.

【20】Lin Z C. Study on estimation of short duration PMP [D]. Nanjing: Nanjing University of Information Science & Technology. 2018, 16.
林智琛. 短历时可能最大降水估算研究 [D]. 南京: 南京信息工程大学. 2018, 16.

【21】Fuzhou Marine. 2019-09-11)[2019-11-12] [EB/OL]. http:∥hyj.fuzhou.gov.cn/zz/jsfwykp/yykp/201909/t20190911_3037804.htm. 2019.
-09-11)[2019-11-12] . http:∥hyj.fuzhou.gov.cn/zz/jsfwykp/yykp/201909/t20190911_3037804.htm. 2019.

【22】Luo Y, Zhao W, Zhai J Q. Dichotomous weather forecasts score research and a new measure of score [J]. Quarterly Journal of Applied Meteorology. 2009, 20(2): 129-136.
罗阳, 赵伟, 翟景秋. 两类天气预报评分问题研究及一种新评分方法 [J]. 应用气象学报. 2009, 20(2): 129-136.

【23】Tran D, Bourdev L, Fergus R, et al. Learning spatiotemporal features with 3D convolutional networks . [C]∥2015 IEEE International Conference on Computer Vision (ICCV), December 7-13, 2015, Santiago, Chile. New York: IEEE. 2015, 4489-4497.

【24】Yang H, Yuan C F, Li B, et al. Asymmetric 3D convolutional neural networks for action recognition [J]. Pattern Recognition. 2019, 85: 1-12.

引用该论文

Yang Suhui,Lin Zhiwei,Lai Shaojun,Liu Jinfu. Precipitation Nowcasting Based on Dual-Flow 3D Convolution and Monitoring Images[J]. Laser & Optoelectronics Progress, 2020, 57(20): 201011

杨素慧,林志玮,赖绍钧,刘金福. 结合双流3D卷积和监控图像的降水临近预报[J]. 激光与光电子学进展, 2020, 57(20): 201011

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF