Open AccessMWNet: object detection network applicable for different weather conditions
Author(s) -
Pei Liu,
Yuan Xue,
Dai XueRui
Publication year - 2019
Publication title -
iet intelligent transport systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.579
H-Index - 45
eISSN - 1751-9578
pISSN - 1751-956X
DOI - 10.1049/iet-its.2019.0023
Subject(s) - subnet , computer science , convolutional neural network , object detection , real time computing , object (grammar) , encoder , extreme weather , artificial intelligence , pattern recognition (psychology) , computer network , operating system , ecology , climate change , biology
The existing vehicle environment perception systems remain limited with regard to the ability to detect objects under complex weather conditions. This study proposes a novel network architecture named multi‐weather network (MWNet), which can improve the performance of the on‐board object detection system under extreme weather conditions. It consists of an encoder and a decoder. The encoder is comprised of shared convolutional layers used to extract features, while the decoder consists of three subnets, namely weather classification subnet, bad weather detection subnet, and fair weather detection subnet. Moreover, the results are satisfactory even for images photographed under different weather and illumination conditions.