CHEN Xia,WANG Haijiang,ZHOU Shuyue,et al.A Hail Weather Recognition Method based on Artificial Intelligence[J].Journal of Chengdu University of Information Technology,2021,36(05):512-517.[doi:10.16836/j.cnki.jcuit.2021.05.007]
基于人工智能的冰雹天气识别方法研究
- Title:
- A Hail Weather Recognition Method based on Artificial Intelligence
- 文章编号:
- 2096-1618(2021)05-0512-06
- Keywords:
- hail weather; artificial intelligence; tracking
- 分类号:
- TP183
- 文献标志码:
- A
- 摘要:
- 冰雹天气是强对流天气系统产生的一种强天气现象,具有历时短、局地性强、破坏力大等特点,是主要的灾害性天气之一,因此对冰雹天气的识别一直是气象领域一个重要的研究课题。提出一种基于人工智能的冰雹天气识别方法,通过使用Faster R-CNN深度学习对冰雹天气区域进行识别,在测试集上的识别率和准确率分别达到87%和97%。通过理论分析,结合实际降雹情况及与传统冰雹天气识别方法对比验证的方式,验证了识别结果的准确性及可靠性。实验结果对冰雹天气这种灾害性天气的识别与预警具有重要的参考意义。
- Abstract:
- Hail weather is a strong weather phenomenon produced by the system of severe convective weather, which has the characteristics of short duration, strong locality and great destructive power. Therefore, the detection of hail weather has always been a significant research topic in the meteorological field. This paper presents a hail weather recognition method based on artificial intelligence. In this method, the Faster R-CNN deep learning is applied to detect hail, and accuracy and precision on the test set are 87% and 97% respectively. The reliability and accuracy of the method are verified by adopting theoretical analysis and comparing with the actual situation and the traditional hail weather identification algorithm. The experimental results have significant reference for the identification and early warning of hail weather.
参考文献/References:
[1] 慕建利.陕西关中强暴雨中尺度对流系统研究[D]. 江苏:南京信息工程大学, 2009.
[2] 赵泽皓, 孔誉旋. 人工智能应用于气象业务的现状及发展趋势[J]. 河南科技, 2019,690(28):152-154.
[3] Hinton G E.Reducing the Dimensionality of Data with Neural Networks[J]. Science, 2006, 37(6):104-112.
[4] P Sermanet, D Eigen, X Zhang et al. OverFeat: Integrated Recognition, Localization and Detection using Convolutional Networks[J]. arXiv preprint arXiv,2013,1312:6229.
[5] Girshick R, Donahue J, Darrell T, et al. Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation[C]. 2014 IEEE Conference on Computer Vision and Pattern Recognition(CVPR). IEEE Computer Society, 2014.
[6] Girshick R. Fast R-CNN[J]. Computer Science, 2015, 34(6):904-917.
[7] Ren Shaoqing, He Kaiming, Girshick Ross, et al. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks[J]. IEEE transactions on pattern analysis and machine intelligence, 2017, 39(6).
[8] Dai J, Li Y, He K, et al. R-FCN: Object Detection via Region-based Fully Convolutional Networks[J]. Advances in Neural Information Processing Systems, 2016, 30(2):190-200.
[9] Simonyan K, Zisserman A. Very Deep Convolutional Networks for Large-Scale Image Recognition[J]. Computer Science, 2015, 34(3):186-196.
[10] Jiang B, Luo R, Mao J, et al. Acquisition of Localization Confidence for Accurate Object Detection[C].European Conference on Computer Vision(ECCV),2018.
[11] 曹俊武,刘黎平.双线偏振多普勒天气雷达识别冰雹区方法研究[J].气象,2006(6):13-19.
相似文献/References:
[1]赵鑫宁,喻 歆,吴 锡.一种基于混合概率选择算子的改进遗传算法[J].成都信息工程大学学报,2016,(03):247.
ZHAO Xin-ning,YU Xin,WU Xi.Improving Genetic Algorithm based on A Hybrid Probabilistic Selector[J].Journal of Chengdu University of Information Technology,2016,(05):247.
备注/Memo
收稿日期:2021-09-08
基金项目:国家自然科学基金资助项目(U1733103)