耦合光流矢量与移动距离约束的大气污染传输路径识别算法

贺威文; 邹滨; 徐勇; 张钦挺; 许杰; 吴坚

PDF(13358 KB)

中国环境科学 ›› 2025, Vol. 45 ›› Issue (9) : 4825-4836.

大气污染与控制

耦合光流矢量与移动距离约束的大气污染传输路径识别算法

贺威文¹, 邹滨¹, 徐勇^1,2, 张钦挺¹, 许杰³, 吴坚⁴

作者信息 +

Atmospheric pollutant transport pathway identification algorithm coupling optical flow vectors with movement distance constraints

HE Wei-wen¹, ZOU Bin¹, XU Yong^1,2, ZHANG Qin-ting¹, XU Jie³, WU Jian⁴

Author information +

文章历史 +

摘要

提出一种耦合光流矢量与移动距离约束的大气污染传输路径识别算法——Farneback-Distance Transmission Path Recognition Algorithm (FBD-TPRA).该算法基于一维深度卷积神经网络(1D-CNN)反演生成的高时空分辨率(空间分辨率为2km,时间分辨率为1h)PM_2.5浓度卫星遥感反演数据,通过光流法提取污染气团帧间运动特征,引入移动距离约束,实现污染气团传输路径的连续精细重构.湖南省长沙市三次典型大气污染事件(2024年1月14日､10月12日和10月27日)验证实验显示,FBD-TPRA算法所识别的传输路径能够合理解释城市浓度变化的时序关系(仙桃-荆州-岳阳-长沙､岳阳-常德-益阳-长沙､岳阳-益阳-长沙);相比广泛应用的HYSPLIT模拟路径,FBD-TPRA算法识别的传输路径与其在净位移方向上具有较高的一致性,余弦相似度最高达0.99,但在路径形态上存在差异,动态时间规整(DTW值)分别为2.99°､9.58°和3.69°.研究表明,FBD-TPRA算法通过结合高时空分辨率污染物浓度数据和移动距离约束,增强了对污染物传输过程的刻画能力,为大气污染传输路径的识别提供了新的技术支持.

Abstract

Proposing a novel algorithm for identifying atmospheric pollutant transport pathways by coupling optical flow vectors with movement distance constraints—Farneback-Distance Transmission Path Recognition Algorithm (FBD-TPRA). Based on high spatiotemporal-resolution (2km spatial resolution, 1h temporal resolution) PM_2.5 concentration satellite remote sensing inversion data generated via a one-dimensional deep convolutional neural network (1D-CNN), this algorithm extracts inter-frame motion features of polluted air masses using optical flow and incorporates movement distance constraints to achieve continuous and refined reconstruction of pollutant transport pathways. Validation experiments conducted during three typical air pollution events in Changsha, Hunan Province (January 14, October 12, and October 27, 2024) demonstrate that the transport paths identified by the FBD-TPRA algorithm can reasonably explain the temporal relationships of urban concentration variations (Xiantao-Jingzhou- Yueyang-Changsha, Yueyang-Changde-Yiyang-Changsha, and Yueyang-Yiyang-Changsha). Compared with the widely used HYSPLIT model, the FBD-TPRA achieves strong consistency in net displacement direction(cosine similarity up to 0.99), but differ in path morphology, as reflected by dynamic time warping (DTW) values of 2.99°, 9.58°, and 3.69°, respectively. The study indicates that the FBD-TPRA algorithm enhances the characterization of pollutant transport processes by integrating high-resolution spatiotemporal pollutant concentration data and movement distance constraints, providing a new technical approach for identifying atmospheric pollution transport pathways.

导出引用

贺威文, 邹滨, 徐勇, 张钦挺, 许杰, 吴坚. 耦合光流矢量与移动距离约束的大气污染传输路径识别算法[J]. 中国环境科学. 2025, 45(9): 4825-4836

HE Wei-wen, ZOU Bin, XU Yong, ZHANG Qin-ting, XU Jie, WU Jian. Atmospheric pollutant transport pathway identification algorithm coupling optical flow vectors with movement distance constraints[J]. China Environmental Science. 2025, 45(9): 4825-4836

中图分类号： X51

参考文献

[1] 姚森,张晗宇,王晓琦,等.2016年1月京津冀地区大气污染特征与多尺度传输量化评估 [J].环境科学, 2021,42(2):534-545. Yao S, Zhang H Y, Wang X Q, et al. Air pollution characteristics and quantitative evaluation of multi-scale transport in the Beijing- Tianjin-Hebei Region in January, 2016 [J]. Environmental Sciences, 2021,42(2):534-545.
[2] Ramanathan V, Feng Y. Air pollution, greenhouse gases and climate change: Global and regional perspectives [J]. Atmospheric Environment, 2009,43(1):37-50.
[3] 饶晓琴,张碧辉,江琪,等.东南亚生物质燃烧对云南边境污染传输影响 [J].中国环境科学, 2023,43(9):4459-4468. Rao X Q, Zhang B H, Jiang Q, et al. Effects of biomass burning in Southeast Asia on pollution transport in Yunnan border area [J]. China Environmental Science, 2023,43(9):4459-4468.
[4] 薛文博,付飞,王金南,等.中国PM_2.5跨区域传输特征数值模拟研究 [J].中国环境科学, 2014,34(6):1361-1368. Xue W B, Fu F, Wang J N, et al. Numerical study on the characteristics of regional transport of PM_2.5 in China [J]. China Environmental Science, 2014,34(6):1361-1368.
[5] 王芳,陈东升,程水源,等.基于气流轨迹聚类的大气污染输送影响 [J].环境科学研究, 2009,22(6):637-642. Wang F, Chen D S, Cheng S Y, et al. Impacts of air pollutant transport based on air trajectory clustering [J]. Environmental Science Research, 2009,22(6):637-642.
[6] 康宝荣,刘立忠,刘焕武,等.关中地区秋冬季颗粒物二次有机碳的估算 [J].中国环境科学, 2019,39(9):3663-3670. Tang B R, Liu L Z, Liu H W, et al. Estimation of secondary organic carbon in PM_2.5 and PM₁₀ in Guanzhong area in autumn and winter [J]. China Environmental Science, 2019,39(9):3663-3670.
[7] Chang X, Wang S X, Zhao B, et al. Contributions of inter-city and regional transport to PM_2.5 concentrations in the BeijingTianjin-Hebei region and its implications on regional joint air pollution control [J]. Science of the Total Environment, 2019, 660:1191-1200.
[8] 彭玏,赵媛媛,赵吉麟,等.京津冀大气污染传输通道区大气污染时空格局研究 [J].中国环境科学, 2019,39(2):449-458. Peng G, Zhao Y Y, Zhao J L, et al. Spatiotemporal patterns of air pollution in air pollution transmission channel of Beijing-Tianjin- Hebei from 2000 to 2015 [J]. China Environmental Science, 2019, 39(2):449-458.
[9] 罗彪,刘暄.基于后向轨迹模式下的合肥市大气污染来源研究 [J].中国科学技术大学学报, 2019,49(4):321-328. Luo B, Liu X. Investigation on the source of air pollutants in Hefei city based on backward trajectory model [J]. Journal of University of Science and Technology of China, 2019,49(4):321-328.
[10] 陈佳欣.典型情景下武汉市PM_2.5和O₃污染形成机制和控制策略模拟研究 [D].武汉:武汉华中科技大学, 2023. Chen J X. Simulation study of PM_2.5 and O₃ pollution formation mechanisms and control strategies in Wuhan under typical scenarios [D]. Wuhan: Wuhan Huazhong University of Science and Technology, 2023.
[11] 潘勇,郑捷,肖航.长三角地区典型PM_2.5污染过程和跨区域传输对宁波污染贡献评估模拟 [J].环境科学, 2023,44(2):634-645. Pan Y, Zheng J, Xiao H. Simulation evaluation of the contribution of typical PM_2.5 pollution and trans-regional transport to Ningbo pollution in the Yangtze River Delta [J]. Environmental Science, 2023, 44(2): 634-645.
[12] 王继康,花丛,桂海林,等.2016年1月我国中东部一次大气污染物传输过程分析 [J].气象, 2017,43(7):804-812. Wang J K, Hua C, Gui H L, et al. Transport of PM_2.5 of the haze pollution episode over Mid-Eastern China in January 2016 [J]. Meteorology, 2017,43(7): 804-812.
[13] 张天浩.高时空分辨气溶胶遥感反演及其在大气污染排放和传输中的应用研究 [D].武汉:武汉大学, 2020. Zhang T H. High spatial and temporal resolution remote sensing based aerosol optical depth retrieval and its application in atmospheric environment [D]. Wuhan: Wuhan University, 2020.
[14] 王郭臣,王珏,信玉洁,等.天津PM₁₀和NO₂输送路径及潜在源区研究 [J].中国环境科学, 2014,34(12):3009-3016. Wang G C, Wang J, Xin Y J, et al. Transportation pathways and potential source areas of PM₁₀ and NO₂ in Tianjin [J]. China Environmental Science, 2014,34(12):3009-3016.
[15] 任浦慧,解静芳,姜洪进,等.太原市大气PM_2.5季节传输路径和潜在源分析 [J].中国环境科学, 2019,39(8):3144-3151. Ren P H, Jie J F, Jiang H J, et al. Transport pathways and potential sources of PM_2.5 in different seasons in Taiyuan city [J]. China Environmental Science, 2019,39(8):3144-3151.
[16] 孙雪,刘子贺,李轲,等.基于后向轨迹模式的区域大气污染及传输特征分析 [J].环境监测管理与技术, 2022,34(3):25-29. Sun X, Liu Z H, Li K, et al. Analysis of regional air pollution and transmission characteristics based on backward trajectory model [J]. Environmental Monitoring Management and Technology, 2022,34(3): 25-29.
[17] 崔宏,刘肖,秦巧燕.汾渭平原典型污染城市PM_2.5来源分布及传输分析 [J].环境工程技术学报, 2022,12(5):1593-1600. Cui H, Liu X, Qin Q Y. PM_2.5 source distribution and transmission analysis in typical polluted cities of Fenwei Plain [J]. Journal of Environmental Engineering Technology, 2022,12(5):1593-1600.
[18] Chen J P, Yin J H, Zang L, et al. Stacking machine learning model for estimating hourly PM_2.5 in China based on Himawari 8 aerosol optical depth data [J]. Science of the Total Environment, 2019, 697.
[19] 吴迪,杜宁,王莉,等.基于GTWR-XGBoost模型的四川省PM_2.5小时浓度估算 [J].环境科学, 2023,44(7):3738-3748. Wu D, Du Y, Wang L, et al. Estimation of PM_2.5 hourly concentration in Sichuan Province based on GTWR-XGBoost model [J]. Environmental Sciences, 2023,44(7):3738-3748.
[20] 杨欢.基于光流的微小型飞行器目标检测和姿态估计方法研究 [D].北京:北京理工大学, 2016. Yang H. Optical flow method based moving objects detection and attitude information estimation of UAVs [D]. Beijing: Beijing Institute of Technology, 2016.
[21] 武龙冬.基于光流的机器人视觉避障技术研究 [D].沈阳:辽宁大学, 2017. Wu L D. Research on obstacle avoidance technology of robot vision based on optical flow [D]. Shenyang: Liaoning University, 2017.
[22] 赵悦,储开斌,张继,等.面向复杂环境的特征匹配算法 [J].计算机应用与软件, 2025,42(1):264-270,293. Zhao Y, Chu K B, Zhang J, et al. Feature point matching algorithms for complex environments [J]. Computer Applications and Software, 2025,42(01):264-270,293.
[23] 熊健,冯徽徽,丁莹,等.长沙地区PM_2.5传输路径和潜在源区时空分析 [J].环境科学学报, 2021,41(9):3663-3672. Xiong J, Feng H H, Ding Y, et al. Spatial and temporal analysis of PM_2.5 transmission paths and potential sources regions in Changsha Region [J]. Journal of Environmental Science, 2021,41(9):3663-3672.
[24] 胡晨霞,邹滨,李沈鑫,等.城市微环境PM_2.5浓度空间分异特征分析 [J].中国环境科学, 2018,38(3):910-916. Hu C X, Zou B, Li S X, et al. Spatial heterogeneity analysis of PM_2.5 concentrations in intra-urban microenvironments [J]. China Environmental Science, 2018,38(3):910-916.
[25] 曾欣,舒磊,张靖,等.长株潭区域一次重度污染过程气象成因 [J].中国环境监测, 2023,39(5):183-192. Zeng X, Shu L, Zhang J, et al. Meteorological cause of a heavy pollution process in Changsha-Zhuzhou-Xiangtan region [J]. Environmental monitoring in China, 2023,39(5):183-192.
[26] 李细生,陈媛,张克非,等.长株潭大气重污染的成因分型和冷空气侵入型路径 [J].四川环境, 2023,42(6):37-44. Li X S, Chen Y, Zhang K F, et al. Causes classification and cold air intrusion path of heavy air pollution in Chang Zhu Tan [J]. Sichuan Environment, 2023,42(6):37-44.
[27] 丁莹,冯徽徽,邹滨,等.长株潭城市群气溶胶时空分布与传输规律 [J].中国环境科学, 2020,40(5):1906-1914. Ding Y, Feng H H, Zou B, et al. Spatial-temporal distribution and transport characteristic of aerosol in Changsha-Zhuzhou-Xiangtan urban agglomeration [J]. China Environmental Science, 2020,40(5): 1906-1914.
[28] 周莉,兰明才,张舰齐,等.2017年冬季长沙一次重度霾污染天气过程气象成因分析 [J].气象与环境学报, 2019,35(3):45-51. Zhou, L, Lan M C, Zhang J Q, et al. Characteristics and causes of a heavy haze event in Changsha during winter of 2017 [J]. Journal of Meteorology and Environment, 2019,35(3):45-51.
[29] 朱家亮,贺科,张敏,等.长沙市大气污染特征的气象条件分析 [J].环保科技, 2022,28(6):54-61. Zhu J L, He K, Zhang M, et al. Analysis of meteorological conditions of air pollution characteristics in Changsha City [J]. Environmental Technology, 2022,28(6):54-61.
[30] Horn B, Schunck B. Determiningopticalflow [J]. ArtificialIntelligence, 1981,17(13):185-203.
[31] Farnebäck G. Two-frame motion estimation based on polynomial expansion [C] //Image Analysis: 13th Scandinavian Conference, SCIA 2003 Halmstad, Sweden, June 29-July 2, 2003 Proceedings 13. Springer Berlin Heidelberg, 2003:363-370.
[32] 蒋维楣,刘红年,张宁,等.空气污染气象学 [M].南京:南京大学出版社, 2021:120-125. Jiang W M, Liu H N, Zhang N, et al. Air pollution meteorology [M]. Nanjing: Nanjing University Press, 2021:120-125.
[33] 赵旭辉,童欢欢,王含月,等.合肥市冬季颗粒物污染传输通道分析及通量研究 [J].中国环境监测, 2022,38(3):43-52. Zhao X H, Tong H H, Wang H Y, et al. Study on the transmission channels and fluxes of particulate pollution in Hefei in winter [J]. Environmental monitoring in China, 2022,38(3):43-52.
[34] Wang Y Q, Zhang X Y, Draxler R R. TrajStat:GlS-based software that uses various trajectory statistical analysis methods to identify potential sources from long-term air pollution measurement data [J]. Environmental Modelling&Software, 24(8): 938-939.
[35] 湖南生态环境.刚刚,长沙启动重污染天气黄色预警 [EB/OL]. https://mp.weixin.qq.com/s/i1d1EkHJeY6hj0xKejr_uQ,2024-1-14. Hunan ecological environment. Just now, Changsha started the yellow warning of heavy pollution weather [EB/OL]. https://mp.weixin.qq. com/s/i1d1EkHJeY6hj0xKejr_uQ,2024-1-14.
[36] 湖北生态环境.污染气团过境雾霾来势汹汹预计湖北省空气质量16日整体转好 [EB/OL]. https://mp.weixin.qq.com/s/h2lrczo-7E- X9BbCRMFlGg,2024-1-15. Hubei ecological environment. It is estimated that the air quality in Hubei Province will be improved on the 16th day [EB/OL]. https://mp.weixin.qq.com/s/h2lrczo-7E-X9BbCRMFlGg,2024-1-15.
[37] 湖南生态环境.长沙启动重污染天气黄色预警!请市民注意防范 [EB/OL].https://mp.weixin.qq.com/s/ScCUu2xHVSPJuXCBB_nqTA,2024-10-13. Hunan ecological environment. Changsha start heavy pollution weather yellow warning! Please pay attention to prevent [EB/OL]. https://mp.weixin.qq.com/s/ScCUu2xHVSPJuXCBB_nqTA,2024-10-13.
[38] 湖南生态环境.刚刚!我省2市启动重污染天气黄色预警 [EB/OL]. https://mp.weixin.qq.com/s/T_DFLiYUPveOAuMPtXLkUw,2024-10-28. Hunan ecological environment. Just! Yellow warning of heavy pollution weather launched in two cities of our province [EB/OL]. https://mp.weixin.qq.com/s/T_DFLiYUPveOAuMPtXLkUw,2024-10-28.