Citation: Shi, M.; He, P.; Shi, Y.
Detecting Extratropical Cyclones of
the Northern Hemisphere with
Single Shot Detector. Remote Sens.
2022, 14, 254. https://doi.org/
10.3390/rs14020254
Academic Editor: Nunzio Cennamo
Received: 12 November 2021
Accepted: 1 January 2022
Published: 6 January 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Detecting Extratropical Cyclones of the Northern Hemisphere
with Single Shot Detector
Minjing Shi
1,2
, Pengfei He
1
and Yuli Shi
1,
*
1
School of Remote Sensing and Geomatic Engineering, Nanjing University of Information & Science
Technology, Nanjing 210044, China; miningshi@gmail.com or mishi@ucsd.edu (M.S.);
gopfhe@nuist.edu.cn (P.H.)
2
Department of Computer Science and Engineering, University of California San Diego,
La Jolla, CA 92093, USA
* Correspondence: ylshi.nuist@gmail.com or ylshi@nuist.edu.cn
Abstract:
In this paper, we propose a deep learning-based model to detect extratropical cyclones
(ETCs) of the northern hemisphere, while developing a novel workflow of processing images and
generating labels for ETCs. We first labeled the cyclone center by adapting an approach from Bonfanti
et al. in 2017 and set up criteria of labeling ETCs of three categories: developing, mature, and
declining stages. We then gave a framework of labeling and preprocessing the images in our dataset.
Once the images and labels were ready to serve as inputs, an object detection model was built with
Single Shot Detector (SSD) and adjusted to fit the format of the dataset. We trained and evaluated our
model with our labeled dataset on two settings (binary and multiclass classifications), while keeping
a record of the results. We found that the model achieves relatively high performance with detecting
ETCs of mature stage (mean Average Precision is 86.64%), and an acceptable result for detecting
ETCs of all three categories (mean Average Precision 79.34%). The single-shot detector model can
succeed in detecting ETCs of different stages, and it has demonstrated great potential in the future
applications of ETC detection in other relevant settings.
Keywords: extratropical cyclone; SSD; deep learning; cyclone detection; front cloud system
1. Introduction
1.1. Background
Extratropical cyclones (ETCs) are common large-scale meteorological systems that are
often formed in the middle to high latitudes for each hemisphere and feature intense varia-
tion in several horizontal and vertical physical quantities, such as temperature, air pressure,
wind speed, etc. [
1
]. Because of the ETCs’ huge influence on the ecological environment in
terms of weather disasters such as low temperatures or storms in a wide span of regions, it
is important for researchers and forecasters to find systems that automatically detect and
identify them from satellite imagery, so that they can closely examine their characteristics
since the images can bring instant visual observations of the ETCs.
However, the automatic detection and identification of ETCs from satellite images
have been a difficult problem for a long time due to various unique features of the cyclones.
Firstly, different cyclones can bear varying degrees of mobility based on their size and
geographic location. This makes it difficult to track down the cyclones by purely capturing
the place at which the cyclones formed. Secondly, ETC systems bear large amounts of
internal variations between individual samples. Some ETCs are small and regular-looking,
while other ETCs can be gigantic in size and holding irregular shapes and features. Third,
analyzing a sizable amount of ETC systems from a database generated by satellites requires
a large amount of computing power, while also calling for a model complex enough
to handle large variations between different samples. Finally, while the current several
meteorological databases provide a plethora of ETC examples within the satellite images,
Remote Sens. 2022, 14, 254. https://doi.org/10.3390/rs14020254 https://www.mdpi.com/journal/remotesensing
