A Lightweight Fine-Grained VRU Detection Model for Roadside Units

Jian Shi, Dongxian Sun, Haodong Zhang, Haiqiu Tan, Yaoguang Hu, Wuhong Wang*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Object detection of vulnerable road users (VRU) under low computing resources of roadside units is one of the key technologies to achieve vehicle-infrastructure cooperative perception. In this paper, a lightweight fine-grained VRU detection model is proposed. Analyzing the existing complex traffic environment, the traditional definition of VRU is no longer applicable. Our work includes two parts: One is to redefine the fine-grained VRU and construct a new dataset. This task makes the perceptual information obtained by detection more comprehensive and accurate. Another is to optimize YOLOv4 by using the channel pruning method in model compression. The optimized model is 60% lighter than the original model. Under the limitation of low computing resources at the roadside units, the real-time detection of VRU is realized while ensuring a certain detection accuracy.

Original languageEnglish
Title of host publicationGreen Transportation and Low Carbon Mobility Safety - Proceedings of the 12th International Conference on Green Intelligent Transportation Systems and Safety
EditorsWuhong Wang, Jianping Wu, Ruimin Li, Xiaobei Jiang, Haodong Zhang
PublisherSpringer Science and Business Media Deutschland GmbH
Pages301-309
Number of pages9
ISBN (Print)9789811956140
DOIs
Publication statusPublished - 2023
Event12th International Conference on Green Intelligent Transportation Systems and Safety, 2021 - Beijing, China
Duration: 17 Nov 202119 Nov 2021

Publication series

NameLecture Notes in Electrical Engineering
Volume944
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Conference

Conference12th International Conference on Green Intelligent Transportation Systems and Safety, 2021
Country/TerritoryChina
CityBeijing
Period17/11/2119/11/21

Fingerprint

Dive into the research topics of 'A Lightweight Fine-Grained VRU Detection Model for Roadside Units'. Together they form a unique fingerprint.

Cite this