Real-Time Object Detection for the Running Train Based on the Improved YOLO V4 Neural Network

Matching the detection speed and accuracy in practical applications is considered to improve the speed of video object detection in front of trains. Lightweight convolutional neural network MobileNet and clustering ideas are combined to improve the object detection algorithm, and the MYOLO-lite mode...

Full description

Saved in:
Bibliographic Details
Main Authors: Yang Liu, Mengfei Gao, Humin Zong, Xinping Wang, Jinshuang Li
Format: Article
Language:English
Published: Wiley 2022-01-01
Series:Journal of Advanced Transportation
Online Access:http://dx.doi.org/10.1155/2022/4377953
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849304674760392704
author Yang Liu
Mengfei Gao
Humin Zong
Xinping Wang
Jinshuang Li
author_facet Yang Liu
Mengfei Gao
Humin Zong
Xinping Wang
Jinshuang Li
author_sort Yang Liu
collection DOAJ
description Matching the detection speed and accuracy in practical applications is considered to improve the speed of video object detection in front of trains. Lightweight convolutional neural network MobileNet and clustering ideas are combined to improve the object detection algorithm, and the MYOLO-lite model object detection algorithm is designed. The self-made object dataset of a forward-moving train is combined with the K-means clustering idea. The transcendental frame is redesigned to enhance scale adaptability. The nonmaximum suppression and loss function improvement methods of the MYOLO-lite network model are proposed for occluded tracks in front of the train, obstruction of trains, low detection accuracy of large object coincidence, and uneven distribution of positive and negative samples. The mean average precision value of the experimentally designed MYOLO-lite model object detection algorithm reaches 95.74%, with 42.04 frames per second. Detecting a picture only takes 0.024 s.
format Article
id doaj-art-41eca192223b48669bf99b8a74925eb9
institution Kabale University
issn 2042-3195
language English
publishDate 2022-01-01
publisher Wiley
record_format Article
series Journal of Advanced Transportation
spelling doaj-art-41eca192223b48669bf99b8a74925eb92025-08-20T03:55:40ZengWileyJournal of Advanced Transportation2042-31952022-01-01202210.1155/2022/4377953Real-Time Object Detection for the Running Train Based on the Improved YOLO V4 Neural NetworkYang Liu0Mengfei Gao1Humin Zong2Xinping Wang3Jinshuang Li4School of Automation and Electrical EngineeringSchool of Automation and Electrical EngineeringNational Marine Environmental Monitoring CenterSchool of Automation and Electrical EngineeringSchool of Automation and Electrical EngineeringMatching the detection speed and accuracy in practical applications is considered to improve the speed of video object detection in front of trains. Lightweight convolutional neural network MobileNet and clustering ideas are combined to improve the object detection algorithm, and the MYOLO-lite model object detection algorithm is designed. The self-made object dataset of a forward-moving train is combined with the K-means clustering idea. The transcendental frame is redesigned to enhance scale adaptability. The nonmaximum suppression and loss function improvement methods of the MYOLO-lite network model are proposed for occluded tracks in front of the train, obstruction of trains, low detection accuracy of large object coincidence, and uneven distribution of positive and negative samples. The mean average precision value of the experimentally designed MYOLO-lite model object detection algorithm reaches 95.74%, with 42.04 frames per second. Detecting a picture only takes 0.024 s.http://dx.doi.org/10.1155/2022/4377953
spellingShingle Yang Liu
Mengfei Gao
Humin Zong
Xinping Wang
Jinshuang Li
Real-Time Object Detection for the Running Train Based on the Improved YOLO V4 Neural Network
Journal of Advanced Transportation
title Real-Time Object Detection for the Running Train Based on the Improved YOLO V4 Neural Network
title_full Real-Time Object Detection for the Running Train Based on the Improved YOLO V4 Neural Network
title_fullStr Real-Time Object Detection for the Running Train Based on the Improved YOLO V4 Neural Network
title_full_unstemmed Real-Time Object Detection for the Running Train Based on the Improved YOLO V4 Neural Network
title_short Real-Time Object Detection for the Running Train Based on the Improved YOLO V4 Neural Network
title_sort real time object detection for the running train based on the improved yolo v4 neural network
url http://dx.doi.org/10.1155/2022/4377953
work_keys_str_mv AT yangliu realtimeobjectdetectionfortherunningtrainbasedontheimprovedyolov4neuralnetwork
AT mengfeigao realtimeobjectdetectionfortherunningtrainbasedontheimprovedyolov4neuralnetwork
AT huminzong realtimeobjectdetectionfortherunningtrainbasedontheimprovedyolov4neuralnetwork
AT xinpingwang realtimeobjectdetectionfortherunningtrainbasedontheimprovedyolov4neuralnetwork
AT jinshuangli realtimeobjectdetectionfortherunningtrainbasedontheimprovedyolov4neuralnetwork