A Lightweight Traffic Lights Detection and Recognition Method for Mobile Platform
Traffic lights detection and recognition (TLDR) is one of the necessary abilities of multi-type intelligent mobile platforms such as drones. Although previous TLDR methods have strong robustness in their recognition results, the feasibility of deployment of these methods is limited by their large mo...
Main Authors: | , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-04-01
|
Series: | Drones |
Subjects: | |
Online Access: | https://www.mdpi.com/2504-446X/7/5/293 |
_version_ | 1797600380762193920 |
---|---|
author | Xiaoyuan Wang Junyan Han Hui Xiang Bin Wang Gang Wang Huili Shi Longfei Chen Quanzheng Wang |
author_facet | Xiaoyuan Wang Junyan Han Hui Xiang Bin Wang Gang Wang Huili Shi Longfei Chen Quanzheng Wang |
author_sort | Xiaoyuan Wang |
collection | DOAJ |
description | Traffic lights detection and recognition (TLDR) is one of the necessary abilities of multi-type intelligent mobile platforms such as drones. Although previous TLDR methods have strong robustness in their recognition results, the feasibility of deployment of these methods is limited by their large model size and high requirements of computing power. In this paper, a novel lightweight TLDR method is proposed to improve its feasibility to be deployed on mobile platforms. The proposed method is a two-stage approach. In the detection stage, a novel lightweight YOLOv5s model is constructed to locate and extract the region of interest (ROI). In the recognition stage, the HSV color space is employed along with an extended twin support vector machines (TWSVMs) model to achieve the recognition of multi-type traffic lights including the arrow shapes. The dataset, collected in naturalistic driving experiments with an instrument vehicle, is utilized to train, verify, and evaluate the proposed method. The results suggest that compared with the previous YOLOv5s-based TLDR methods, the model size of the proposed lightweight TLDR method is reduced by 73.3%, and the computing power consumption of it is reduced by 79.21%. Meanwhile, the satisfied reasoning speed and recognition robustness are also achieved. The feasibility of the proposed method to be deployed on mobile platforms is verified with the Nvidia Jetson NANO platform. |
first_indexed | 2024-03-11T03:47:19Z |
format | Article |
id | doaj.art-01bf0189c9ee40c29976a9852c6a181c |
institution | Directory Open Access Journal |
issn | 2504-446X |
language | English |
last_indexed | 2024-03-11T03:47:19Z |
publishDate | 2023-04-01 |
publisher | MDPI AG |
record_format | Article |
series | Drones |
spelling | doaj.art-01bf0189c9ee40c29976a9852c6a181c2023-11-18T01:06:53ZengMDPI AGDrones2504-446X2023-04-017529310.3390/drones7050293A Lightweight Traffic Lights Detection and Recognition Method for Mobile PlatformXiaoyuan Wang0Junyan Han1Hui Xiang2Bin Wang3Gang Wang4Huili Shi5Longfei Chen6Quanzheng Wang7College of Electromechanical Engineering, Qingdao University of Science & Technology, Qingdao 266100, ChinaCollege of Electromechanical Engineering, Qingdao University of Science & Technology, Qingdao 266100, ChinaCollege of Electromechanical Engineering, Qingdao University of Science & Technology, Qingdao 266100, ChinaCollege of Electromechanical Engineering, Qingdao University of Science & Technology, Qingdao 266100, ChinaCollege of Electromechanical Engineering, Qingdao University of Science & Technology, Qingdao 266100, ChinaCollege of Electromechanical Engineering, Qingdao University of Science & Technology, Qingdao 266100, ChinaCollege of Electromechanical Engineering, Qingdao University of Science & Technology, Qingdao 266100, ChinaCollege of Electromechanical Engineering, Qingdao University of Science & Technology, Qingdao 266100, ChinaTraffic lights detection and recognition (TLDR) is one of the necessary abilities of multi-type intelligent mobile platforms such as drones. Although previous TLDR methods have strong robustness in their recognition results, the feasibility of deployment of these methods is limited by their large model size and high requirements of computing power. In this paper, a novel lightweight TLDR method is proposed to improve its feasibility to be deployed on mobile platforms. The proposed method is a two-stage approach. In the detection stage, a novel lightweight YOLOv5s model is constructed to locate and extract the region of interest (ROI). In the recognition stage, the HSV color space is employed along with an extended twin support vector machines (TWSVMs) model to achieve the recognition of multi-type traffic lights including the arrow shapes. The dataset, collected in naturalistic driving experiments with an instrument vehicle, is utilized to train, verify, and evaluate the proposed method. The results suggest that compared with the previous YOLOv5s-based TLDR methods, the model size of the proposed lightweight TLDR method is reduced by 73.3%, and the computing power consumption of it is reduced by 79.21%. Meanwhile, the satisfied reasoning speed and recognition robustness are also achieved. The feasibility of the proposed method to be deployed on mobile platforms is verified with the Nvidia Jetson NANO platform.https://www.mdpi.com/2504-446X/7/5/293mobile platformconnected and automated vehiclesvisual sensingtraffic lights detection and recognitionlightweight model |
spellingShingle | Xiaoyuan Wang Junyan Han Hui Xiang Bin Wang Gang Wang Huili Shi Longfei Chen Quanzheng Wang A Lightweight Traffic Lights Detection and Recognition Method for Mobile Platform Drones mobile platform connected and automated vehicles visual sensing traffic lights detection and recognition lightweight model |
title | A Lightweight Traffic Lights Detection and Recognition Method for Mobile Platform |
title_full | A Lightweight Traffic Lights Detection and Recognition Method for Mobile Platform |
title_fullStr | A Lightweight Traffic Lights Detection and Recognition Method for Mobile Platform |
title_full_unstemmed | A Lightweight Traffic Lights Detection and Recognition Method for Mobile Platform |
title_short | A Lightweight Traffic Lights Detection and Recognition Method for Mobile Platform |
title_sort | lightweight traffic lights detection and recognition method for mobile platform |
topic | mobile platform connected and automated vehicles visual sensing traffic lights detection and recognition lightweight model |
url | https://www.mdpi.com/2504-446X/7/5/293 |
work_keys_str_mv | AT xiaoyuanwang alightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT junyanhan alightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT huixiang alightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT binwang alightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT gangwang alightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT huilishi alightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT longfeichen alightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT quanzhengwang alightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT xiaoyuanwang lightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT junyanhan lightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT huixiang lightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT binwang lightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT gangwang lightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT huilishi lightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT longfeichen lightweighttrafficlightsdetectionandrecognitionmethodformobileplatform AT quanzhengwang lightweighttrafficlightsdetectionandrecognitionmethodformobileplatform |