Open Access

ARTICLE

Research on Thunderstorm Identification Based on Discrete Wavelet Transform

Xiaopeng Li¹, Ziyuan Xu^3,4, Jin Han^1,*, Xingming Sun^1,2, Yi Cao⁵

1 Nanjing University of Information Science & Technology, Nanjing, 210044, China
2 Engineering Research Center of Digital Forensics, Ministry of Education, Nanjing, 201144, China
3 Nanjing University (Suzhou) high and New Technology Research Institute, Suzhou, 215123, China
4 Jiangsu Union Technical Institute, Wuxi, 214145, China
5 Department of Electrical and Computer Engineering, University of Windsor, Windsor, N9B 3P4, Canada

* Corresponding Author: Jin Han. Email:

Intelligent Automation & Soft Computing 2022, 33(2), 1153-1166. https://doi.org/10.32604/iasc.2022.023261

Received 01 September 2021; Accepted 28 October 2021; Issue published 08 February 2022

Abstract

Lightning has been one of the most talked-about natural disasters worldwide in recent years, as it poses a great threat to all industries and can cause huge economic losses. Thunderstorms are often accompanied by natural phenomena such as lightning strikes and lightning, and many scholars have studied deeply the regulations of thunderstorm generation, movement and dissipation to reduce the risk of lightning damage. Most of the current methods for studying thunderstorms focus on using more complex algorithms based on radar or lightning data, which increases the computational burden and reduces the computational efficiency to some extent. This paper proposes a raster-based DWT (discrete wavelet transform) method for thunderstorm identification, this method uses DWT, CFSFD (clustering algorithm for fast search and finding density peaks) algorithm and ADTD (active divectory topology diagrammer) lightning location data for thunderstorm identification. The advantage of this method is that it supports different spatial resolutions and can identify any shape and number of thunderstorms at the same time and in the same area. It is effective in eliminating some of cluttered, scattered lightning data and extracting dense areas of thunderstorms. Furthermore, the method has a time complexity of O(n), and the computational efficiency is significantly better than the current TITAN (thunderstorm identification, tracking, analysis, and nowcasting) algorithm, which provides a good basis for subsequent extrapolation studies of thunderstorms.

---

Keywords

---