改进U-Net的风云四号卫星降水估计算法研究

doi:10.3778/j.issn.1002-8331.2202-0255

摘要/Abstract

摘要： 针对强对流天气条件下利用卫星图像进行降水量估计时精度不高、时空分辨率低的问题，提出了一种改进U-Net的降水量估计算法。将U-Net模型的编码器通过残差模块与解码器相结合，使得模型参数可以共享，避免深层网络模型梯度消失的情况。在此结构基础上引入了空间金字塔模块进行多尺度特征提取，保留更多的图像特征，加强对细小降水云团信息的特征提取能力；引入了注意力机制模块，提取重要降水特征信息。实验结果表明，该算法在命中率、虚警率、相关系数分别为0.84、0.48、0.59；均方根误差、平均绝对误差分别为1.354?mm/h、0.432?mm/h。与PERSIANN-CNN、U-Net算法相比，有效地提升了降水量估计精度。与其他降水产品对比，能更好地识别出降水区。该算法可以实现近实时的降水估计，能够有效地提升降水估计精度，对低时间分辨率的降水估计研究具有一定的价值。

关键词: 降水量估计, 风云四号卫星, 卷积神经网络, 语义分割

Abstract: Aiming at the problems of low accuracy and low spatial-temporal resolution of precipitation estimation using satellite images under the condition of severe convective weather, an improved U-Net precipitation estimation algorithm is proposed. Firstly the encoder of the U-Net model is combined with the decoder through the residual module, so that the model parameters can be shared to avoid the disappearance of deep network model gradients. Based on this structure, the spatial pyramid module is introduced for multi-scale feature extraction to retain more image features and strengthen the feature extraction ability of small precipitation cloud information. The attention mechanism module is added to extract important precipitation feature information. The experimental results show that probability of detection, false alarm ration, critical success index of the proposed algorithm are 0.84, 0.48 and 0.59, respectively. The root mean square error and mean absolute error are 1.354?mm/h and 0.432?mm/h respectively. Compared with PERSIANN-CNN and U-Net, the proposed algorithm effectively improves the accuracy of precipitation estimation. Compared with other precipitation products, it also has certain advantages. Therefore, the algorithm can achieve near-real-time results and effectively improve the accuracy of precipitation estimation, which is valuable for the research of precipitation estimation with low tempotal resolution.

Key words: quantitative precipitation estimation, Fengyun-4 satellite, convolutional neural network, semantic segmentation

黄杰, 张永宏, 马光义, 朱灵龙, 田伟. 改进U-Net的风云四号卫星降水估计算法研究[J]. 计算机工程与应用, 2023, 59(11): 285-293.

HUANG Jie, ZHANG Yonghong, MA Guangyi, ZHU Linglong, TIAN Wei. Research on Precipitation Estimation Algorithm from Fengyun-4 Satellite Based on Improved U-Net[J]. Computer Engineering and Applications, 2023, 59(11): 285-293.

参考文献

[1] KIDD C，LEVIZZANI V.Status of satellite precipitation retrievals[J].Hydrology and Earth System Sciences，2011，15（5）：1109-1116.
[2] NGUYEN P，SELLARS S，THORSTENSEN A，et al.Satellites track precipitation of super Typhoon Haiyan[J].EOS Transactions American Geophysical Union，2014，95（16）：133-135.
[3] ALI M，AMIR A K.Capabilities of satellite precipitation datasets to estimate heavy precipitation rates at different temporal accumulations[J].Hydrological Processes，2014，28（4）：2262-2270.
[4] KUHNLEIN M，APPELHANS T，THIES B，et al.Precipitation estimates from MSG SEVIRI daytime，nighttime，and twilight data with random forests[J].Journal of Applied Meteorology and Climatology，2014，53（11）：2457-2480.
[5] HABIB E，HAILE A T，TIAN Y，et al.Evaluation of the high-resolution CMORPH satellite rainfall product using dense rain gauge observations and radar-based estimates[J].Journal of Hydrometeorology，2012，13（6）：1784-1798.
[6] 黄安明，冷谦，袁博，等.郴州天气雷达估测降水量误差分析[J].气象科技，2020，48（2）：178-184.
HUANG A M，LENG Q，YUAN B，et al.Analysis of precipitation estimation deviation of synoptic radar in Chenzhou[J].Meteorological Science and Technology，2020，48（2）：178-184.
[7] JING R，YONG H，LI G，et al.Two different integration methods for weather radar-based quantitative precipitation estimation[J].Advances in Meteorology，2017：1-12.
[8] 延昊，吴门新，刘桂青，等.利用MODIS的多通道信息估计中纬度降雨[J].气象学报，2006，64（2）：205-210.
YAN H，WU M X，LIU G Q，et al.Overland rainfall estimates in midlatitude by using MODIS multi-channels[J].ACTA Meteorologica Sinica，2006，64（2）：205-210.
[9] EBERT E.Comparison of near-real-time precipitation estimates from satellite observations and numerical models[J].Bulletin of the American Meteorological Society，2010，88（1）：47-64.
[10] TAO Y，GAO X，IHLER A，et al.Precipitation identification with bispectral satellite information using deep learning approaches[J].Journal of Hydrometeorology，2017，18（5）：1271-1283.
[11] TAO Y，HSU K，IHLER A，et al.A two-stage deep neural network framework for precipitation estimation from bispectral satellite information[J].Journal of Hydrometeorology，2018，19（2）：393-408.
[12] SADEGHI M，ASANJAN A A，FARIDZAD M，et al.PERSIANN-CNN：precipitation estimation from remotely sensed information using artificial neural networks- convolutional neural networks[J].Journal of Hydrometeorology，2019，20（12）：2273-2289.
[13] MORAUX A，DEWITTE S，CORNELIS B，et al.Deep learning for precipitation estimation from satellite and rain gauges measurements[J].Remote Sensing，2019，11（21）：2463.
[14] XUE M，HANG R L，LIU Q S，et al.CNN-based near-real-time precipitation estimation from Fengyun-2 satellite over Xinjiang，China[J].Atmospheric Research，2021，250：105337.
[15] YANG J，ZHANG Z，WEI C，et al.Introducing the new generation of Chinese geostationary weather satellites，Fengyun-4[J].Bulletin of the American Meteorological Society，2017，98（8）：1637-1658.
[16] WANG C，XU J，TANG G，et al.Infrared precipitation estimation using convolutional neural network[J].IEEE Transactions on Geoscience and Remote Sensing，2020，58（12）：8612-8625.
[17] LU X Y，TANG G Q，WEI M，et al.Evaluation of multi-satellite precipitation products in Xinjiang，China[J].International Journal of Remote Sensing，2018，39（21）：7437-7462.
[18] YANG X，SUN P，ZHANG F，et al.Multi-task collaboration deep learning framework for infrared precipitation estimation[J].Remote Sensing，2021，13（12）：2310.
[19] MAYOR Y G，TERESHCHENKO I，FONSECA-HERNáNDEZ M，et al.Evaluation of error in IMERG precipitation estimates under different topographic conditions and temporal scales over Mexico[J].Remote Sensing，2017，9（5）：503.
[20] WANG P，CHEN P，YUAN Y，et al.Understanding convolution for semantic segmentation[J].arXiv：1702.08502，2017.
[21] 张永宏，严斌，田伟，等.基于PPMU-net的多特征高分辨率遥感道路提取[J].计算机工程与应用，2021，57（1）：200-206.
ZHANG Y H，YAN B，TIAN W，et al.Multi-feature high-resolution remote sensing road extraction based on PPMU-net[J].Computer Engineering and Applications，2021，57（1）：200-206.
[22] LIU Q，LI Y，YU M Z，et al.Daytime rainy cloud detection and convective precipitation delineation based on a deep neural network method using GOES-16 ABI Images[J].Remote Sensing，2019，11（21）：2555.
[23] MS A，PN A，KH A，et al.Improving near real-time precipitation estimation using a U-Net convolutional neural network and geographical information[J].Environmental Modelling & Software，2020，134：1-14.