基于视频的天气现象识别及其应用研究

刘冬韡; 史军; 俞玮; 王亚东; 郭巍; 杜明斌

doi:10.19912/j.0254-0096.tynxb.2023-0602

PDF(1520 KB)

太阳能学报 ›› 2024, Vol. 45 ›› Issue (8) : 441-447. DOI: 10.19912/j.0254-0096.tynxb.2023-0602

基于视频的天气现象识别及其应用研究

刘冬韡¹, 史军¹, 俞玮¹, 王亚东², 郭巍¹, 杜明斌¹

作者信息 +

VIDEO-BASED WEATHER PHENOMENON RECOGNITION AND ITS APPLICATION RESEARCH

Liu Dongwei¹, Shi Jun¹, Yu Wei¹, Wang Yadong², Guo Wei¹, Du Mingbin¹

Author information +

文章历史 +

摘要

针对传统天气现象观测因采用专用设备导致布设和维护成本高、获取难的问题,提出一种利用广泛布设的视频实现对天气现象观测的方法。通过将因特网获取的16327张天气现象图片放入深度神经网络中训练,建立一个天气现象分类的预训练模型。在此基础上加入上海徐家汇和洋山港气象站2021年视频图像,对模型进行微调,开发基于站点视频的天气现象识别模型。利用2022年1—10月份的视频图像数据对模型进行检验,模型识别结果的F1评分分别为0.74和0.67,而人工识别结果分别为0.67和0.61,表明所建立的模型性能接近或优于人眼识别的效果。通过例举两个应用案例,证明该项技术具有较好的应用前景。

Abstract

In this study a newly method is proposed to identify weather phenomena with widely dispersed video data to address the issues of high deployment and maintenance costs by traditional weather phenomenon observation equipment. In this method, a deep neural network is firstly used to train 16327 weather phenomenon photographs from internet for building a pre-training categorization model, and then a fine-tuning procedure is applied with those video photographs from two meteorological observation station to improve the identifying accuracy. The model was testeel by using the video image data from January to October 2022. The results indicate that the performance of the proposed model is comparable to or superior to that of human eye recognition with the F1 scores of 0.74 and 0.67 for the model recognition in the two different station compared with 0.67 and 0.61 for the manual recognition. A cases analysis shows that the proposed model is mostly perfect for identifying sunny and fog but unsatisfactory for rainfall. Further, two application cases show that the model can be used to retrieve the sunshine duration and identify weather phenomena in real time using the online video images. Most important, this method can be employed as a substituted scheme for the present weather phenomena instrument to achieve low-cost weather information observation.

导出引用

刘冬韡, 史军, 俞玮, 王亚东, 郭巍, 杜明斌. 基于视频的天气现象识别及其应用研究[J]. 太阳能学报. 2024, 45(8): 441-447 https://doi.org/10.19912/j.0254-0096.tynxb.2023-0602

Liu Dongwei, Shi Jun, Yu Wei, Wang Yadong, Guo Wei, Du Mingbin. VIDEO-BASED WEATHER PHENOMENON RECOGNITION AND ITS APPLICATION RESEARCH[J]. Acta Energiae Solaris Sinica. 2024, 45(8): 441-447 https://doi.org/10.19912/j.0254-0096.tynxb.2023-0602

中图分类号： TP391

参考文献

[1] 于瑛, 郭佳豪, 姚星, 等. 天气类型对逐时太阳总辐射模型适应性的影响[J]. 太阳能学报, 2022, 43(12): 186-193.
YU Y, GUO J H, YAO X, et al.Influence of weather types on adaptability of hourly global solar radiation models[J]. Acta energiae solaris sinica, 2022, 43(12): 186-193.
[2] 李芬, 杨勇, 胡超, 等. 基于天气类型识别的跟踪式光伏发电系统能量评估及优化[J]. 太阳能学报, 2020, 41(11): 112-119.
LI F, YANG Y, HU C, et al.Energy evaluation analysis and optimization of tracking photovoltaic generation system based on weather type recognition[J]. Acta energiae solaris sinica, 2020, 41(11): 112-119.
[3] 刘冬韡, 穆海振, 贺千山, 等. 一种基于实景图像的低能见度识别算法[J]. 应用气象学报, 2022, 33(4): 501-512.
LIU D W, MU H Z, HE Q S, et al.A low visibility recognition algorithm based on surveillance video[J]. Journal of applied meteorological science, 2022, 33(4): 501-512.
[4] LU C W, LIN D, JIA J Y, et al.Two-class weather classification[C]//2014 IEEE Conference on Computer Vision and Pattern Recognition. Columbus, OH, USA, 2014: 3718-3725.
[5] 张红艳, 李茵茵, 万伟. 改进K近邻和支持向量机相融合的天气识别[J]. 计算机工程与应用, 2014, 50(14): 148-151, 167.
ZHANG H Y, LI Y Y, WAN W.Weather identification based on improved K nearest neighbor and support vector machine[J]. Computer engineering and applications, 2014, 50(14): 148-151, 167.
[6] 李骞, 范茵, 张璟, 等. 基于室外图像的天气现象识别方法[J]. 计算机应用, 2011, 31(6): 1624-1627.
LI Q, FAN Y, ZHANG J, et al.Method of weather recognition based on decision-tree-based SVM[J]. Journal of computer applications, 2011, 31(6): 1624-1627.
[7] 于浩, 王孝通, 徐冠雷. 基于贝叶斯分类的雾天和雨天两类天气图像自动识别[J]. 舰船电子工程, 2016, 36(9): 73-75, 94.
YU H, WANG X T, XU G L.Automatic recognition of two different weathers of foggy and rainy based on Bayesian classification[J]. Ship electronic engineering, 2016, 36(9): 73-75, 94.
[8] XIA J M, XUAN D W, TAN L, et al.ResNet15: weather recognition on traffic road with deep convolutional neural network[J]. Advances in meteorology, 2020, 2020: 6972826.
[9] XIAO H X, ZHANG F, SHEN Z P, et al. Classification of weather phenomenon from images by using deep convolutional neural network[J]. Earth and space science, 2021, 8(5): e2020EA001604.
[10] ZHAO B, HUA L L, LI X L, et al.Weather recognition via classification labels and weather-cue maps[J]. Pattern recognition, 2019, 95: 272-284.
[11] 王亚朝, 赵伟, 徐海洋, 等. 基于轻量型卷积神经网络的非固定场景天气识别算法[J]. 电子测量技术, 2019, 42(17): 152-156.
WANG Y Z, ZHAO W, XU H Y, et al.Non-fixed scene weather classification algorithm based on lightweight convolutional neural network[J]. Electronic measurement technology, 2019, 42(17): 152-156.
[12] ZHANG Z, MA H D, FU H Y, et al.Scene-free multi-class weather classification on single images[J]. Neurocomputing, 2016, 207: 365-373.
[13] 左杰格, 柳晓鸣, 蔡兵. 基于图像分块与特征融合的户外图像天气识别[J]. 计算机科学, 2022, 49(3): 197-203.
ZUO J G, LIU X M, CAI B.Outdoor image weather recognition based on image blocks and feature fusion[J]. Computer science, 2022, 49(3): 197-203.
[14] 郭志强, 胡永武, 刘鹏, 等. 基于特征融合的室外天气图像分类[J]. 计算机应用, 2020, 40(4): 1023-1029.
GUO Z Q, HU Y W, LIU P, et al.Outdoor weather image classification based on feature fusion[J]. Journal of computer applications, 2020, 40(4): 1023-1029.
[15] XIE K Z, HUANG L, WEI Z Q, et al.WCATN: Unsupervised deep learning to classify weather conditions from outdoor images[J]. Engineering applications of artificial intelligence, 2022, 113: 104928.
[16] ZHANG Z, MA H D.Multi-class weather classification on single images[C]//2015 IEEE International Conference on Image Processing (ICIP).Quebec City, QC, Canada, 2015: 4396-4400.
[17] 高嵩, 毕宝贵, 李月安, 等. MICAPS4预报业务系统建设进展与未来发展[J]. 应用气象学报, 2017, 28(5): 513-531.
GAO S, BI B G, LI Y A, et al.Implementation and development plan of MICAPS4[J]. Journal of applied meteorological science, 2017, 28(5): 513-531.
[18] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[J]. arXiv e-prints, 2014: arXiv: 1409.1556.
[19] SZEGEDY C, IOFFE S, VANHOUCKE V, et al.Inception-v4, inception-ResNet and the impact of residual connections on learning[C]//Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence. San Francisco, California, USA, 2017: 4278-4284.
[20] TAN M, LE Q.EfficientNet: rethinking model scaling for convolutional neural networks[C]//International Conference on Machine Learning. PMLR, 2019: 6105-6114.
[21] TU Z Z, TALEBI H, ZHANG H, et al.MaxViT: multi-axis vision transformer[C]//Computer Vision-ECCV 2022: 17th European conference. Tel Aviv, Israel, 2022: 459-479.
[22] 廖星宇. 深度学习入门之 PyTorch[M]. 北京: 电子工出版社, 2017: 13-14.
LIAO X Y.PyTorch, an introduction to deep learning[M]. Beijing: Publishing House of Electronics Industry, 2017: 13-14.
[23] 周志华. 机器学习[M]. 北京: 清华大学出版社, 2016: 30.
ZHOU Z H.Machine learning[M]. Beijing: Tsinghua University Press, 2016: 30.
[24] 马宁, 任芝花, 王妍, 等. 全国降水天气现象平行观测对比分析[J]. 干旱区研究, 2022, 39(1): 54-63.
MA N, REN Z H, WANG Y, et al.Comparative analysis of parallel observations of precipitation phenomena in China[J]. Arid zone research, 2022, 39(1): 54-63.