New Method for Single-Site Cloud-to-Ground Lightning Location Based on Tri-Pre Processing

The single-site lightning detection system can provide timely and effective information on lightning activity in areas where a multi-site lightning network cannot be built. Using deep learning, the single-site lightning detection achieves better performance than traditional methods, but it is highly...

Full description

Saved in:
Bibliographic Details
Main Authors: Bingzhe Dai, Qilin Zhang, Jie Li, Yi Liu, Minhong Zhao
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/17/10/1766
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850126943879430144
author Bingzhe Dai
Qilin Zhang
Jie Li
Yi Liu
Minhong Zhao
author_facet Bingzhe Dai
Qilin Zhang
Jie Li
Yi Liu
Minhong Zhao
author_sort Bingzhe Dai
collection DOAJ
description The single-site lightning detection system can provide timely and effective information on lightning activity in areas where a multi-site lightning network cannot be built. Using deep learning, the single-site lightning detection achieves better performance than traditional methods, but it is highly dependent on the quality of the training dataset. To address this, this paper proposes a method called Tri-Pre to improve dataset quality and thereby enhance the performance of single-site cloud-to-ground lightning detection based on deep learning. After using the Tri-Pre method, the location model’s distance estimation error decreases by 36.08%. For lightning with propagation distances greater than 1000 km, the average relative error of the results from the built model based on the Tri-Pre method is 3.78%. When verified using additional measured data, the model also shows satisfactory accuracy, particularly for lightning with propagation distances beyond 1000 km. Specifically, for lightning with propagation distances between 1500 and 1600 km, the average relative location error is approximately 5.46%.
format Article
id doaj-art-d0a2af9ca6d44218b8bb0d36b5374021
institution OA Journals
issn 2072-4292
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-d0a2af9ca6d44218b8bb0d36b53740212025-08-20T02:33:48ZengMDPI AGRemote Sensing2072-42922025-05-011710176610.3390/rs17101766New Method for Single-Site Cloud-to-Ground Lightning Location Based on Tri-Pre ProcessingBingzhe Dai0Qilin Zhang1Jie Li2Yi Liu3Minhong Zhao4Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters (CIC-FEMD), Key Laboratory of Meteorological Disaster, Ministry of Education (KLME), Research Institute of Intelligent-Sensing and Disaster Prevention for Extreme Weather, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaCollaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters (CIC-FEMD), Key Laboratory of Meteorological Disaster, Ministry of Education (KLME), Research Institute of Intelligent-Sensing and Disaster Prevention for Extreme Weather, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaCollaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters (CIC-FEMD), Key Laboratory of Meteorological Disaster, Ministry of Education (KLME), Research Institute of Intelligent-Sensing and Disaster Prevention for Extreme Weather, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaJiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology (CICAEET), Nanjing University of Information Science & Technology, Nanjing 210044, ChinaCollaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters (CIC-FEMD), Key Laboratory of Meteorological Disaster, Ministry of Education (KLME), Research Institute of Intelligent-Sensing and Disaster Prevention for Extreme Weather, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaThe single-site lightning detection system can provide timely and effective information on lightning activity in areas where a multi-site lightning network cannot be built. Using deep learning, the single-site lightning detection achieves better performance than traditional methods, but it is highly dependent on the quality of the training dataset. To address this, this paper proposes a method called Tri-Pre to improve dataset quality and thereby enhance the performance of single-site cloud-to-ground lightning detection based on deep learning. After using the Tri-Pre method, the location model’s distance estimation error decreases by 36.08%. For lightning with propagation distances greater than 1000 km, the average relative error of the results from the built model based on the Tri-Pre method is 3.78%. When verified using additional measured data, the model also shows satisfactory accuracy, particularly for lightning with propagation distances beyond 1000 km. Specifically, for lightning with propagation distances between 1500 and 1600 km, the average relative location error is approximately 5.46%.https://www.mdpi.com/2072-4292/17/10/1766single-site lightning locationpreprocessing methodlightning waveformdeep learning
spellingShingle Bingzhe Dai
Qilin Zhang
Jie Li
Yi Liu
Minhong Zhao
New Method for Single-Site Cloud-to-Ground Lightning Location Based on Tri-Pre Processing
Remote Sensing
single-site lightning location
preprocessing method
lightning waveform
deep learning
title New Method for Single-Site Cloud-to-Ground Lightning Location Based on Tri-Pre Processing
title_full New Method for Single-Site Cloud-to-Ground Lightning Location Based on Tri-Pre Processing
title_fullStr New Method for Single-Site Cloud-to-Ground Lightning Location Based on Tri-Pre Processing
title_full_unstemmed New Method for Single-Site Cloud-to-Ground Lightning Location Based on Tri-Pre Processing
title_short New Method for Single-Site Cloud-to-Ground Lightning Location Based on Tri-Pre Processing
title_sort new method for single site cloud to ground lightning location based on tri pre processing
topic single-site lightning location
preprocessing method
lightning waveform
deep learning
url https://www.mdpi.com/2072-4292/17/10/1766
work_keys_str_mv AT bingzhedai newmethodforsinglesitecloudtogroundlightninglocationbasedontripreprocessing
AT qilinzhang newmethodforsinglesitecloudtogroundlightninglocationbasedontripreprocessing
AT jieli newmethodforsinglesitecloudtogroundlightninglocationbasedontripreprocessing
AT yiliu newmethodforsinglesitecloudtogroundlightninglocationbasedontripreprocessing
AT minhongzhao newmethodforsinglesitecloudtogroundlightninglocationbasedontripreprocessing