北京大学学报自然科学版 ›› 2020, Vol. 56 ›› Issue (5): 796-804.DOI: 10.13209/j.0479-8023.2020.065

上一篇    下一篇

基于U-net神经网络模型的PM2.5逐小时浓度值预测模型

李燚航1, 翟卫欣2,3,†, 颜寒祺3, 朱道也3, 童晓冲4, 程承旗5   

  1. 1. 北京大学城市与环境学院, 北京 100871 2. 中国农业大学信息与电气工程学院, 北京 100083 3. 北京大学前沿交叉学科研究院, 北京 100871 4. 信息工程大学地理空间信息学院, 郑州450052 5. 北京大学工学院空天信息工程研究中心, 北京 100871
  • 收稿日期:2019-09-11 修回日期:2020-03-16 出版日期:2020-09-20 发布日期:2020-09-20
  • 通讯作者: 翟卫欣, E-mail: pkuzhaiweixin(at)gmail.com
  • 基金资助:
    国家重点研发计划项目(2018YFB0505300, 2017YFB0503703)、广西科技重大专项项目(桂科 AA18118025)、国防科技创新特区项目和中国博士后科学基金(2020M670024)资助

Prediction of PM2.5 Hour Concentration Based on U-net Neural Network

LI Yihang1, ZHAI Weixin2,3,†, YAN Hanqi3, ZHU Daoye3, TONG Xiaochong4, CHENG Chengqi5   

  1. 1. College of Urban and Environmental Sciences, Peking University, Beijing 100871 2. College of Information and Electrical Engineering, China Agricultural University, Beijing 100083 3. Academy for Advanced Interdisciplinary Studies, Peking University, Beijing 100871 4. Institute of Geospatial Information, Information Engineering University, Zhengzhou 450052 5. Aerospace Information Engineering Research Center, Peking University, Beijing 100871
  • Received:2019-09-11 Revised:2020-03-16 Online:2020-09-20 Published:2020-09-20
  • Contact: ZHAI Weixin, E-mail: pkuzhaiweixin(at)gmail.com

PDF

摘要：

针对目前多数PM2.5预测模型泛化能力较差的问题, 提出基于U-net神经网络模型的PM2.5逐小时浓度值预测模型。该模型通过引入历史风场数据, 将离散的监测站点PM2.5浓度值插值为PM2.5网格图; 然后将U-net神经网络作为预测模型, 基于实验区域的10小时内的PM2.5网格图, 预测下一时刻的PM2.5网格图。该模型可以利用历史不同时刻提取的PM2.5浓度值网格图, 在预测区域内所有位置PM2.5浓度值的同时, 还可以提升预测的准确性以及对PM2.5浓度值突变情况的适应性。实验结果表明, 所提方法在PM2.5浓度值短时间突变情况下, 预测精度比传统方法有10%左右的提升。

关键词: PM2.5预测, 突变, 基于历史风速插值, 网格图, 神经网络

Abstract:

Most of the previous PM2.5 prediction models present unsatisfactory performance in several aspects, including predicting accuracy and generalization ability, especially in case of the sudden change in the value of PM2.5 situation. Therefore, we propose a method based on the U-net neural network to predict the hourly PM2.5 concentration value on the research area, attempting to improve the prediction performance. The proposed model includes two major steps. First, based on the inverse distance interpolation of historical wind field data, discrete station PM2.5 values are interpolated into a PM2.5 grid map; second, the U-net neural network is applied to train the prepared spatiotemporal grid data and make predictions. The model can use the PM2.5 concentration values of the grid map extracted at different time stamps for the PM2.5 prediction. The PM2.5 concentration values at all locations in the research region can be achieved. Specifically, the prediction accuracy and the generalization ability of the model in case of sudden changes are revealed. Experimental results indicate that the proposed method has a 10% improvement in the prediction accuracy of PM2.5 concentration values in the case of sudden change.

Key words: PM2.5 prediction, abrupt scenarios, interpolation of historical wind speed, grid graph, neural network

版权所有 © 《北京大学学报（自然科学版）》编辑部
技术支持：北京玛格泰克科技发展有限公司