扫 描 看 全 文
引用
陈镔捷，叶扬，林溢，游诗雪，邓劲松，杨武，王珂.2022.耦合注意力机制DNN的PM2.5估算及时空特征分析.遥感学报，26（5）： 1027-1038
Chen B J， Ye Y， Lin Y， You S X， Deng J S， Yang W and Wang K. 2022. Spatiotemporal estimation of PM2.5 using attention-based deep neural network. National Remote Sensing Bulletin， 26（5）：1027-1038
PM,2.5,作为指示环境质量的重要因子之一，不仅影响着灰霾天气的发生，还与公众健康息息相关，近年来受到广泛的关注。尽管PM,2.5,地面观测站点在不断地扩张，其覆盖范围依旧有限，难以反映全域PM,2.5,浓度的时空异质性。本研究运用卫星遥感气溶胶光学厚度数据，辅助因子除常规的气象因子等以外，还加入了针对中国人民生产生活习惯的农历日因子，提出一种耦合注意力机制的深度神经网络模型，对长三角区域2015年—2020年PM,2.5,浓度进行了逐日的高精度估算。模型交叉验证结果显示决定系数,R,2,高达0.85，斜率0.86，与地面站点观测值有较高的一致性，优于多元线性回归和随机森林模型。长三角区域PM,2.5,浓度时空特征分析结果表明，PM,2.5,浓度在空间上呈现北高南低的趋势；季节特征以冬季浓度最高，夏季浓度最低，春秋过渡。此外，长三角区域2015年—2020年整体PM,2.5,浓度呈下降趋势，其中以上海市最为明显，下降速率为3.30 μg/（m,3,·,a），其次为江苏省（2.65 μg/（m,3,·,a））；浙江省与安徽省下降速率都小于2 μg/（m,3,·,a），但由于安徽省PM,2.5,浓度远高于浙江省，提升空间更大，需要更多的关注。综上所述，利用卫星数据结合本研究提出的方法能弥补地面观测站点的不足，获得高精度全域PM,2.5,浓度时空分布特征，从而更科学地指导相关政策的规划与落地。
PM,2.5, as the primary indicator of environmental quality, not only affects the occurrence of haze but also is closely related to public health and has raised great attention recently. Although PM,2.5, ground monitoring stations are expanding, they are still on the sparse side to identify the spatiotemporal heterogeneity of PM,2.5, concentrations. With the development of remote sensing technology, satellite-based Aerosol Optical Depth (AOD) data provide an effective way to estimate large-scale PM,2.5, concentrations. This study aims to develop a novel deep neural network model for estimating PM,2.5, concentrations in the Yangtze River Delta (YRD).,In addition to satellite remote sensing AOD data, meteorological factors, digital elevation model data, normalized different vegetation index data, and the lunar calendar day representing Chinese production and living habits were integrated into the proposed attention-based Self-Adaptive Deep Neural Network (SADNN) in this study to estimate PM,2.5, concentrations in the YRD region from 2015 to 2020. Five-fold cross-validation was executed to evaluate the estimation accuracy of the SADNN. The multiple linear regression and random forest models were applied to compare with the SADNN.,The cross-validation results showed the proposed SADNN model had a high coefficient of determination value of 0.85 and a slope of 0.86, which were highly consistent with ground-level observations. The results also showed better performance than those of multiple linear regression and random forest models. The results for the spring festival in 2016 demonstrated the effectiveness of integrating the lunar calendar day and attention module into the model. The spatiotemporal patterns of PM,2.5, in the YRD were as follows: PM,2.5, concentrations were high in the north and low in the south, and the coastal and mountainous areas were better than inland and plain areas, respectively. On seasonal scales, winter was the most polluted season, while summer was the best. The overall PM,2.5, concentration in the YRD showed a decreasing trend from 2015 to 2020, especially in Shanghai Municipality, with the decreasing speed of 3.30 μg/(m,3,·,a), following the Jiangsu Province (2.65 μg/(m,3,·,a)). Zhejiang Province and Anhui Province had a lower decreasing speed of less than 2 μg/(m,3,·,a), and Anhui Province needed more effort and attention to improve the air quality due to its overall high PM,2.5, concentrations.,In conclusion, applying satellite remote sensing data and the proposed SADNN model to accurately estimate spatially continuous PM,2.5, concentrations can greatly make up for the lack of ground monitoring stations and scientific guide for environmental policy planning and implementation.
遥感气溶胶光学厚度（AOD）深度学习注意力机制长三角区域空气质量
remote sensingAerosol Optical Depth (AOD)deep learningattention modulethe Yangtze River Deltaair quality
Dockery D W, Pope C A, Xu X P, Spengler J D, Ware J H, Fay M E, Ferris B G Jr and Speizer F E. 1993. An association between air pollution and mortality in six U.S. cities. New England Journal of Medicine, 329(24): 1753-1759 [DOI: 10.1056/nejm199312093292401http://dx.doi.org/10.1056/nejm199312093292401]
Fang X, Zou B, Liu X P, Sternberg T and Zhai L. 2016. Satellite-based ground PM2.5 estimation using timely structure adaptive modeling. Remote Sensing of Environment, 186: 152-163 [DOI: 10.1016/j.rse.2016.08.027http://dx.doi.org/10.1016/j.rse.2016.08.027]
Fu Y Y, Ye Z R, Deng J S, Zheng X Y, Huang Y B, Yang W, Wang Y H and Wang K. 2019. Finer resolution mapping of marine aquaculture areas using WorldView-2 imagery and a hierarchical cascade convolutional neural network. Remote Sensing, 11(14): 1678 [DOI: 10.3390/rs11141678http://dx.doi.org/10.3390/rs11141678]
GBD 2019 Risk Factors Collaborators. 2020. Global burden of 87 risk factors in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. The Lancet, 396(10258): 1223-1249 [DOI: 10.1016/S0140-6736(20)30752-2http://dx.doi.org/10.1016/S0140-6736(20)30752-2]
He Q Q and Huang B. 2018. Satellite-based high-resolution PM2.5 estimation over the Beijing-Tianjin-Hebei region of China using an improved geographically and temporally weighted regression model. Environmental Pollution, 236: 1027-1037 [DOI: 10.1016/j.envpol.2018.01.053http://dx.doi.org/10.1016/j.envpol.2018.01.053]
Hu X F, Belle J H, Meng X, Wildani A, Waller L A, Strickland M J and Liu Y. 2017. Estimating PM2.5 concentrations in the conterminous United States using the random forest approach. Environmental Science and Technology, 51(12): 6936-6944 [DOI: 10.1021/acs.est.7b01210http://dx.doi.org/10.1021/acs.est.7b01210]
Just A C, Wright R O, Schwartz J, Coull B A, Baccarelli A A, Tellez-Rojo M M, Moody E, Wang Y J, Lyapustin A and Kloog I. 2015. Using high-resolution satellite aerosol optical depth to estimate daily PM2.5 geographical distribution in Mexico city. Environmental Science and Technology, 49(14): 8576-8584 [DOI: 10.1021/acs.est.5b00859http://dx.doi.org/10.1021/acs.est.5b00859]
Kim S W, Yoon S C, Kim J and Kim S Y. 2007. Seasonal and monthly variations of columnar aerosol optical properties over East Asia determined from multi-year MODIS, LIDAR, and AERONET Sun/sky radiometer measurements. Atmospheric Environment, 41(8): 1634-1651 [DOI: 10.1016/j.atmosenv.2006.10.044http://dx.doi.org/10.1016/j.atmosenv.2006.10.044]
LeCun Y, Bengio Y and Hinton G. 2015. Deep learning. Nature, 521(7553): 436-444 [DOI: 10.1038/nature14539http://dx.doi.org/10.1038/nature14539]
Li S Z, Xing X F, Fan W Q, Cai B L, Fordson P and Xu X M. 2021. Spatiotemporal and frequential cascaded attention networks for speech emotion recognition. Neurocomputing, 448: 238-248 [DOI: 10.1016/j.neucom.2021.02.094http://dx.doi.org/10.1016/j.neucom.2021.02.094]
Li Z Q, Zhang Y, Shao J, Li B S, Hong J, Liu D, Li D H, Wei P, Li W, Li L, Zhang F X, Guo J, Deng Q, Wang B X, Cui C L, Zhang W C, Wang Z Z, Lv Y, Xu H, Chen X F, Li L and Qie L. 2016. Remote sensing of atmospheric particulate mass of dry PM2.5 near the ground: method validation using ground-based measurements. Remote Sensing of Environment, 173: 59-68 [DOI: 10.1016/j.rse.2015.11.019http://dx.doi.org/10.1016/j.rse.2015.11.019]
Lin C Q, Li Y, Yuan Z B, Lau A K H, Li C C and Fung J C H. 2015. Using satellite remote sensing data to estimate the high-resolution distribution of ground-level PM2.5. Remote Sensing of Environment, 156: 117-128 [DOI: 10.1016/j.rse.2014.09.015http://dx.doi.org/10.1016/j.rse.2014.09.015]
Liu Y, Franklin M, Kahn R and Koutrakis P. 2007. Using aerosol optical thickness to predict ground-level PM2.5 concentrations in the St. Louis area: a comparison between MISR and MODIS. Remote Sensing of Environment, 107(1/2): 33-44 [DOI: 10.1016/j.rse.2006.05.022http://dx.doi.org/10.1016/j.rse.2006.05.022]
Lyapustin A, Wang Y J, Korkin S and Huang D. 2018. MODIS collection 6 MAIAC algorithm. Atmospheric Measurement Techniques, 11(10): 5741-5765 [DOI: 10.5194/amt-11-5741-2018http://dx.doi.org/10.5194/amt-11-5741-2018]
Ma Z W. 2015. Study on Spatiotemporal Distributions of PM2.5 in China Using Satellite Remote Sensing. Nanjing: Nanjing University
马宗伟. 2015. 基于卫星遥感的我国PM2.5时空分布研究. 南京: 南京大学
Mou L C and Zhu X X. 2020. Learning to pay attention on spectral domain: a spectral attention module-based convolutional network for hyperspectral image classification. IEEE Transactions on Geoscience and Remote Sensing, 58(1): 110-122 [DOI: 10.1109/TGRS.2019.2933609http://dx.doi.org/10.1109/TGRS.2019.2933609]
Nair V and Hinton G E. 2010. Rectified linear units improve restricted boltzmann machines//Proceedings of the 27th International Conference on International Conference on Machine Learning. Haifa: Omnipress: 807-814
Reddy R V K, Rao B S and Raju K P. 2018. Handwritten hindi digits recognition using convolutional neural network with RMSprop optimization//2018 Second International Conference on Intelligent Computing and Control Systems (ICICCS). Madurai: IEEE: 45-51 [DOI: 10.1109/ICCONS.2018.8662969http://dx.doi.org/10.1109/ICCONS.2018.8662969]
Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez A N, Kaiser Ł and Polosukhin I. 2017. Attention is all you need//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, California : Curran Associates Inc.: 6000-6010
Wang X M. 2020. Analysis of Spatial and Temporal Variation of PM2.5 Meteorological Influencing Factors in Anhui Province. Hefei: Anhui Agricultural University
王学梅. 2020. 安徽省PM2.5时空变化及其气象影响因素分析. 合肥: 安徽农业大学
Wei J, Li Z Q, Cribb M, Huang W, Xue W H, Sun L, Guo J P, Peng Y R, Li J, Lyapustin A, Liu L, Wu H and Song Y M. 2020. Improved 1 km resolution PM2.5 estimates across China using enhanced space–time extremely randomized trees. Atmospheric Chemistry and Physics, 20(6): 3273-3289 [DOI: 10.5194/acp-20-3273-2020http://dx.doi.org/10.5194/acp-20-3273-2020]
Xiao Q Y, Wang Y J, Chang H H, Meng X, Geng G N, Lyapustin A and Liu Y. 2017. Full-coverage high-resolution daily PM2.5 estimation using MAIAC AOD in the Yangtze River Delta of China. Remote Sensing of Environment, 199: 437-446 [DOI: 10.1016/j.rse.2017.07.023http://dx.doi.org/10.1016/j.rse.2017.07.023]
Xie Y Y, Wang Y X, Zhang K, Dong W H, Lv B L and Bai Y Q. 2015. Daily estimation of ground-Level PM2.5 concentrations over Beijing using 3 km resolution MODIS AOD. Environmental Science and Technology, 49(20): 12280-12288 [DOI: 10.1021/acs.est.5b01413http://dx.doi.org/10.1021/acs.est.5b01413]
Yang L J, Xu H Q and Jin Z F. 2018. Estimation of ground-level PM2.5 concentrations using MODIS satellite data in Fuzhou, China. Journal of Remote Sensing, 22(1): 64-75
杨立娟, 徐涵秋, 金致凡. 2018. MODIS卫星遥感估计福州地区近地面PM2.5浓度. 遥感学报, 22(1): 64-75 [DOI: 10.11834/jrs.20186501http://dx.doi.org/10.11834/jrs.20186501]
Ye Z R, Fu Y Y, Gan M Y, Deng J S, Comber A and Wang K. 2019. Building extraction from very high resolution aerial imagery using joint attention deep neural network. Remote Sensing, 11(24): 2970 [DOI: 10.3390/rs11242970http://dx.doi.org/10.3390/rs11242970]
Zang L, Mao F Y, Guo J P, Wang W, Pan Z X, Shen H F, Zhu B and Wang Z M. 2019. Estimation of spatiotemporal PM1.0 distributions in China by combining PM2.5 observations with satellite aerosol optical depth. Science of the Total Environment, 658: 1256-1264 [DOI: 10.1016/j.scitotenv.2018.12.297http://dx.doi.org/10.1016/j.scitotenv.2018.12.297]
Zhang Y and Li Z Q. 2013. Estimation of PM2.5 from fine-mode aerosol optical depth. Journal of Remote Sensing, 17(4): 929-943
张莹, 李正强. 2013. 利用细模态气溶胶光学厚度估计PM2.5. 遥感学报, 17(4): 929-943 [DOI: 10.11834/jrs.20133063http://dx.doi.org/10.11834/jrs.20133063]
Zhu J A. 2020. Research on Visual Object Tracking Algorithm based on Deep Learning. Changchun: University of Chinese Academy of Sciences (Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences)
朱均安. 2020. 基于深度学习的视觉目标跟踪算法研究. 长春: 中国科学院大学(中国科学院长春光学精密机械与物理研究所)
相关作者
相关机构
京公网安备11010802024621
