C2L3-Fusion: An Integrated 3D Object Detection Method for Autonomous Vehicles

Accurate 3D object detection is crucial for autonomous vehicles (AVs) to navigate safely in complex environments. This paper introduces a novel fusion framework that integrates Camera image-based <b>2D object detection using YOLOv8</b> and LiDAR data-based <b>3D object detection us...

Full description

Saved in:
Bibliographic Details
Main Authors: Thanh Binh Ngo, Long Ngo, Anh Vu Phi, Trung Thị Hoa Trang Nguyen, Andy Nguyen, Jason Brown, Asanka Perera
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/9/2688
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850279177788325888
author Thanh Binh Ngo
Long Ngo
Anh Vu Phi
Trung Thị Hoa Trang Nguyen
Andy Nguyen
Jason Brown
Asanka Perera
author_facet Thanh Binh Ngo
Long Ngo
Anh Vu Phi
Trung Thị Hoa Trang Nguyen
Andy Nguyen
Jason Brown
Asanka Perera
author_sort Thanh Binh Ngo
collection DOAJ
description Accurate 3D object detection is crucial for autonomous vehicles (AVs) to navigate safely in complex environments. This paper introduces a novel fusion framework that integrates Camera image-based <b>2D object detection using YOLOv8</b> and LiDAR data-based <b>3D object detection using PointPillars, hence named C2L3-Fusion</b>. Unlike conventional fusion approaches, which often struggle with feature misalignment, <b>C2L3-Fusion</b> enhances spatial consistency and multi-level feature aggregation, significantly improving detection accuracy. Our method outperforms state-of-the-art approaches such as YoPi-CLOCs Fusion Network, standalone YOLOv8, and standalone PointPillars, achieving mean Average Precision (mAP) scores of <b>89.91% (easy), 79.26% (moderate), and 78.01% (hard)</b> on the KITTI dataset. Successfully implemented on the Nvidia Jetson AGX Xavier embedded platform, <b>C2L3-Fusion</b> maintains real-time performance while enhancing robustness, making it highly suitable for self-driving vehicles. This paper details the methodology, mathematical formulations, algorithmic advancements, and real-world testing of C2L3-Fusion, offering a comprehensive solution for 3D object detection in autonomous navigation.
format Article
id doaj-art-99ac59cbebbb456a9ce66a892ffec4f6
institution OA Journals
issn 1424-8220
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-99ac59cbebbb456a9ce66a892ffec4f62025-08-20T01:49:11ZengMDPI AGSensors1424-82202025-04-01259268810.3390/s25092688C2L3-Fusion: An Integrated 3D Object Detection Method for Autonomous VehiclesThanh Binh Ngo0Long Ngo1Anh Vu Phi2Trung Thị Hoa Trang Nguyen3Andy Nguyen4Jason Brown5Asanka Perera6Department of Electrical and Electronic Engineering, University of Transport and Communications, Hanoi 100000, VietnamSoftware and Service Development Department, Mobifone Digital Services, Hanoi 100000, VietnamComputer Science Department, College of Engineering, Michigan State University, East Lansing, MI 48823, USADepartment of Electrical and Electronic Engineering, University of Transport and Communications, Hanoi 100000, VietnamSchool of Engineering, University of Southern Queensland, Springfield, QLD 4300, AustraliaSchool of Engineering, University of Southern Queensland, Springfield, QLD 4300, AustraliaSchool of Engineering, University of Southern Queensland, Springfield, QLD 4300, AustraliaAccurate 3D object detection is crucial for autonomous vehicles (AVs) to navigate safely in complex environments. This paper introduces a novel fusion framework that integrates Camera image-based <b>2D object detection using YOLOv8</b> and LiDAR data-based <b>3D object detection using PointPillars, hence named C2L3-Fusion</b>. Unlike conventional fusion approaches, which often struggle with feature misalignment, <b>C2L3-Fusion</b> enhances spatial consistency and multi-level feature aggregation, significantly improving detection accuracy. Our method outperforms state-of-the-art approaches such as YoPi-CLOCs Fusion Network, standalone YOLOv8, and standalone PointPillars, achieving mean Average Precision (mAP) scores of <b>89.91% (easy), 79.26% (moderate), and 78.01% (hard)</b> on the KITTI dataset. Successfully implemented on the Nvidia Jetson AGX Xavier embedded platform, <b>C2L3-Fusion</b> maintains real-time performance while enhancing robustness, making it highly suitable for self-driving vehicles. This paper details the methodology, mathematical formulations, algorithmic advancements, and real-world testing of C2L3-Fusion, offering a comprehensive solution for 3D object detection in autonomous navigation.https://www.mdpi.com/1424-8220/25/9/2688AIdeep learningC2L3-Fusion2D detection3D detectionautonomous vehicle
spellingShingle Thanh Binh Ngo
Long Ngo
Anh Vu Phi
Trung Thị Hoa Trang Nguyen
Andy Nguyen
Jason Brown
Asanka Perera
C2L3-Fusion: An Integrated 3D Object Detection Method for Autonomous Vehicles
Sensors
AI
deep learning
C2L3-Fusion
2D detection
3D detection
autonomous vehicle
title C2L3-Fusion: An Integrated 3D Object Detection Method for Autonomous Vehicles
title_full C2L3-Fusion: An Integrated 3D Object Detection Method for Autonomous Vehicles
title_fullStr C2L3-Fusion: An Integrated 3D Object Detection Method for Autonomous Vehicles
title_full_unstemmed C2L3-Fusion: An Integrated 3D Object Detection Method for Autonomous Vehicles
title_short C2L3-Fusion: An Integrated 3D Object Detection Method for Autonomous Vehicles
title_sort c2l3 fusion an integrated 3d object detection method for autonomous vehicles
topic AI
deep learning
C2L3-Fusion
2D detection
3D detection
autonomous vehicle
url https://www.mdpi.com/1424-8220/25/9/2688
work_keys_str_mv AT thanhbinhngo c2l3fusionanintegrated3dobjectdetectionmethodforautonomousvehicles
AT longngo c2l3fusionanintegrated3dobjectdetectionmethodforautonomousvehicles
AT anhvuphi c2l3fusionanintegrated3dobjectdetectionmethodforautonomousvehicles
AT trungthihoatrangnguyen c2l3fusionanintegrated3dobjectdetectionmethodforautonomousvehicles
AT andynguyen c2l3fusionanintegrated3dobjectdetectionmethodforautonomousvehicles
AT jasonbrown c2l3fusionanintegrated3dobjectdetectionmethodforautonomousvehicles
AT asankaperera c2l3fusionanintegrated3dobjectdetectionmethodforautonomousvehicles