Citation: Pu, J.; Zhang, W. Electric
Vehicle Fire Trace Recognition Based
on Multi-Task Semantic
Segmentation. Electronics 2022, 11,
1738. https://doi.org/10.3390/
electronics11111738
Academic Editors: Luis
Hernández-Callejo, Sergio
Nesmachnow and Sara Gallardo
Saavedra
Received: 5 May 2022
Accepted: 27 May 2022
Published: 30 May 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
Electric Vehicle Fire Trace Recognition Based on Multi-Task
Semantic Segmentation
Jiankun Pu and Wei Zhang *
School of Microelectronics, Tianjin University, Tianjin 300072, China; 3015204014@tju.edu.cn
* Correspondence: tjuzhangwei@tju.edu.cn
Abstract:
Conflagration is the major safety issue of electric vehicles (EVs). Due to their well-kept
appearance and structure, which demonstrate salient visual changes after combustion, EV bodies are
recognized as an important basis for on-spot inspection of burnt EVs and make application using
semantic segmentation possible. The combination of deep learning-based semantic segmentation
and recognition of visual traces of burnt EVs would provide preliminary analytical results of fire
spread trends and output status descriptions of burnt EVs for further investigation. In this paper,
a dataset of image traces of burnt EVs was built, and a two-branch network structure that splits
the whole task into two sub-tasks separately concentrated on foreground extraction and severity
segmentation is proposed. The proposed network is trained on the dataset via the transfer learning
method and is tested using 5-fold cross validation. The foreground extraction branch achieved a
mean intersection over union (mIoU) of 95.16% in the burnt EV foreground extraction task, and
the burnt severity branch achieved a mIoU of 66.96% for the severity segmentation task. By jointly
training two branches and applying a foreground mask to 3-class severity output, the mIoU was
improved to 68.92%.
Keywords: deep learning; semantic segmentation; electric vehicle fire
1. Introduction
Vehicles are necessities in human life and are extensively utilized in logistics, trans-
portation and travel. The termination of the production of traditional internal combustion
engine vehicles (ICEVs) is being gradually implemented worldwide under the pressure of
the global energy shortage and environment pollution issues, and electric vehicles (EVs)
are recognized ideal alternatives in this situation. Partially or fully driven by Li-ion bat-
teries, EVs have presented the potential hazard of fire, which heavily affects the safety of
passengers under various scenarios, e.g., parking, charging and driving. Fire incidents in
EVs and plug-in hybrid electric vehicles (PHEVs) mostly begin in the battery power system.
Compared with gasoline-caused vehicle fires, battery-caused vehicle fires contain more
energy, extremely high temperatures, and the release of combustible and toxic gas, thus
leading to higher risks and difficulty in extinguishing the fire [1,2].
In order to eliminate potential fire hazards and improve the manufacturing safety of
EVs, correlative research should not only focus on prevention of combustion, but also on
analysis and research of existing cases of burnt EVs. Recently, the on-spot investigation
of burnt EVs has become an important method for analysis and research. Fire or damage
traces remaining on the body panels and vehicle frames are frequently used to locate the
origin of fire [
3
]. When the vehicle is not burnt extensively, traces with salient appearances,
e.g., burnt-off paint and rusted metal, can provide reliable clues for the determination of
fire origin [
4
]. Due to the similarity of material and paint utilized in EVs and conventional
vehicles, fire traces of bodies of burnt EVs are also applicable and credible for investigation.
Moreover, fire traces can be conveniently captured as digital images, which also provides
possibilities for using a computer vision method for recognition.
Electronics 2022, 11, 1738. https://doi.org/10.3390/electronics11111738 https://www.mdpi.com/journal/electronics
