• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International Conference on Computer Vision, Image and Deep Learning >Research on weather radar nowcasting extrapolation
【24h】

Research on weather radar nowcasting extrapolation

机译：天气雷达诺卡化推断研究

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The purpose of heavy rainfall forecast is to predict the distribution of local rainfall intensity in the next 0-2 hours, and the accurate extrapolation image can provide accurate spatial-temporal data reference for the nowcasting of heavy rainfall forecast. Although the accuracy of the radar extrapolation results of the deep learning model based on the recurrent network has been greatly improved compared with the traditional extrapolation model in recent two years, it still needs to be further improved in many aspects of the model. Based on the analysis of the existing ConvSLTM model and TrajGRU model of radar extrapolation, this paper improves them by increasing the number of radar layers and the weight of prediction results. Experiments are carried out with open competition data and real radar data as samples. The experimental results show that the modified network model can better capture spatiotemporal correlation and has more accurate extrapolation effect.
机译：暴雨预测的目的是预测未来0-2小时内局部降雨强度的分布，准确的外推图像可以为大雨预测的州播种预测提供准确的空间数据参考。近两年的传统推断模型相比，基于经常性网络的深度学习模型的雷达外推结果的准确性得到了大大提高，但在模型的许多方面，它仍然需要进一步改善。基于现有ConvslTM模型的分析和雷达外推的Trajgru模型，本文通过增加雷达层的数量和预测结果的重量来改善它们。实验与打开竞争数据和真实雷达数据作为样品进行。实验结果表明，改进的网络模型可以更好地捕获时空相关性并具有更准确的外推效果。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号