Real-Time 3D Multi-Object Detection and Localization Based on Deep Learning for Road and Railway Smart Mobility
For smart mobility, autonomous vehicles, and advanced driver-assistance systems (ADASs), perception of the environment is an important task in scene analysis and understanding. Better perception of the environment allows for enhanced decision making, which, in turn, enables very high-precision actio...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2021-08-01
|
Series: | Journal of Imaging |
Subjects: | |
Online Access: | https://www.mdpi.com/2313-433X/7/8/145 |
_version_ | 1797523342142472192 |
---|---|
author | Antoine Mauri Redouane Khemmar Benoit Decoux Madjid Haddad Rémi Boutteau |
author_facet | Antoine Mauri Redouane Khemmar Benoit Decoux Madjid Haddad Rémi Boutteau |
author_sort | Antoine Mauri |
collection | DOAJ |
description | For smart mobility, autonomous vehicles, and advanced driver-assistance systems (ADASs), perception of the environment is an important task in scene analysis and understanding. Better perception of the environment allows for enhanced decision making, which, in turn, enables very high-precision actions. To this end, we introduce in this work a new real-time deep learning approach for 3D multi-object detection for smart mobility not only on roads, but also on railways. To obtain the 3D bounding boxes of the objects, we modified a proven real-time 2D detector, YOLOv3, to predict 3D object localization, object dimensions, and object orientation. Our method has been evaluated on KITTI’s road dataset as well as on our own hybrid virtual road/rail dataset acquired from the video game Grand Theft Auto (GTA) V. The evaluation of our method on these two datasets shows good accuracy, but more importantly that it can be used in real-time conditions, in road and rail traffic environments. Through our experimental results, we also show the importance of the accuracy of prediction of the regions of interest (RoIs) used in the estimation of 3D bounding box parameters. |
first_indexed | 2024-03-10T08:41:35Z |
format | Article |
id | doaj.art-065d796043354852b6ef0247d72c2319 |
institution | Directory Open Access Journal |
issn | 2313-433X |
language | English |
last_indexed | 2024-03-10T08:41:35Z |
publishDate | 2021-08-01 |
publisher | MDPI AG |
record_format | Article |
series | Journal of Imaging |
spelling | doaj.art-065d796043354852b6ef0247d72c23192023-11-22T08:14:02ZengMDPI AGJournal of Imaging2313-433X2021-08-017814510.3390/jimaging7080145Real-Time 3D Multi-Object Detection and Localization Based on Deep Learning for Road and Railway Smart MobilityAntoine Mauri0Redouane Khemmar1Benoit Decoux2Madjid Haddad3Rémi Boutteau4Normandie Univ, UNIROUEN, ESIGELEC, IRSEEM, 76000 Rouen, FranceNormandie Univ, UNIROUEN, ESIGELEC, IRSEEM, 76000 Rouen, FranceNormandie Univ, UNIROUEN, ESIGELEC, IRSEEM, 76000 Rouen, FranceHaddad is with SEGULA Technologies, 19 rue d’Arras, 92000 Nanterre, FranceNormandie Univ, UNIROUEN, UNILEHAVRE, INSA Rouen, LITIS, 76000 Rouen, FranceFor smart mobility, autonomous vehicles, and advanced driver-assistance systems (ADASs), perception of the environment is an important task in scene analysis and understanding. Better perception of the environment allows for enhanced decision making, which, in turn, enables very high-precision actions. To this end, we introduce in this work a new real-time deep learning approach for 3D multi-object detection for smart mobility not only on roads, but also on railways. To obtain the 3D bounding boxes of the objects, we modified a proven real-time 2D detector, YOLOv3, to predict 3D object localization, object dimensions, and object orientation. Our method has been evaluated on KITTI’s road dataset as well as on our own hybrid virtual road/rail dataset acquired from the video game Grand Theft Auto (GTA) V. The evaluation of our method on these two datasets shows good accuracy, but more importantly that it can be used in real-time conditions, in road and rail traffic environments. Through our experimental results, we also show the importance of the accuracy of prediction of the regions of interest (RoIs) used in the estimation of 3D bounding box parameters.https://www.mdpi.com/2313-433X/7/8/145object detectionlocalizationdistance estimationobject dimensionsobject orientation3D bounding box estimation |
spellingShingle | Antoine Mauri Redouane Khemmar Benoit Decoux Madjid Haddad Rémi Boutteau Real-Time 3D Multi-Object Detection and Localization Based on Deep Learning for Road and Railway Smart Mobility Journal of Imaging object detection localization distance estimation object dimensions object orientation 3D bounding box estimation |
title | Real-Time 3D Multi-Object Detection and Localization Based on Deep Learning for Road and Railway Smart Mobility |
title_full | Real-Time 3D Multi-Object Detection and Localization Based on Deep Learning for Road and Railway Smart Mobility |
title_fullStr | Real-Time 3D Multi-Object Detection and Localization Based on Deep Learning for Road and Railway Smart Mobility |
title_full_unstemmed | Real-Time 3D Multi-Object Detection and Localization Based on Deep Learning for Road and Railway Smart Mobility |
title_short | Real-Time 3D Multi-Object Detection and Localization Based on Deep Learning for Road and Railway Smart Mobility |
title_sort | real time 3d multi object detection and localization based on deep learning for road and railway smart mobility |
topic | object detection localization distance estimation object dimensions object orientation 3D bounding box estimation |
url | https://www.mdpi.com/2313-433X/7/8/145 |
work_keys_str_mv | AT antoinemauri realtime3dmultiobjectdetectionandlocalizationbasedondeeplearningforroadandrailwaysmartmobility AT redouanekhemmar realtime3dmultiobjectdetectionandlocalizationbasedondeeplearningforroadandrailwaysmartmobility AT benoitdecoux realtime3dmultiobjectdetectionandlocalizationbasedondeeplearningforroadandrailwaysmartmobility AT madjidhaddad realtime3dmultiobjectdetectionandlocalizationbasedondeeplearningforroadandrailwaysmartmobility AT remiboutteau realtime3dmultiobjectdetectionandlocalizationbasedondeeplearningforroadandrailwaysmartmobility |