Multiattention Mechanism 3D Object Detection Algorithm Based on RGB and LiDAR Fusion for Intelligent Driving
Open Access
Abstract
This paper proposes a multimodal fusion 3D target detection algorithm based on the attention mechanism to improve the performance of 3D target detection. The algorithm utilizes point cloud data and information from the camera. For image feature extraction, the ResNet50 + FPN architecture extracts features at four levels. Point cloud feature extraction employs the voxel method and FCN to extract point and voxel features. The fusion of image and point cloud features is achieved through regional point fusion and voxel fusion methods. After information fusion, the Coordinate and SimAM attention mechanisms extract fusion features at a deep level. The algorithm’s performance is evaluated using the DAIR-V2X dataset. The results show that compared to the Part-A2 algorithm; the proposed algorithm improves the mAP value by 7.9% in the BEV view and 7.8% in the 3D view at IOU = 0.5 (cars) and IOU = 0.25 (pedestrians and cyclists). At IOU = 0.7 (cars) and IOU = 0.5 (pedestrians and cyclists), the mAP value of the SECOND algorithm is improved by 5.4% in the BEV view and 4.3% in the 3D view, compared to other comparison algorithms.Keywords
Funding Information
- research and development of intelligent vehicle key technologies and industrialization projects based on new-energy vehicles (TC210H02S, 20220301012GX)
- quantitative development and measurement technology research of expected functional safety based on vehicle–cloud collaboration (TC210H02S, 20220301012GX)
This publication has 45 references indexed in Scilit:
- Vehicle Detection from 3D Lidar Using Fully Convolutional Network2016
- Fast R-CNN2015
- Learning Spatiotemporal Features with 3D Convolutional Networks2015
- Data-driven 3D Voxel Patterns for object category recognition2015
- Multiview random forest of local experts combining RGB and LIDAR data for pedestrian detection2015
- Pedestrian detection combining RGB and dense LIDAR data2014
- Single image 3D object detection and pose estimation for grasping2014
- Multi-view traffic sign detection, recognition, and 3D localisationMachine Vision and Applications, 2011
- Least Squares Support Vector Machine ClassifiersNeural Processing Letters, 1999
- COSMOS-A representation scheme for 3D free-form objectsIEEE Transactions on Pattern Analysis and Machine Intelligence, 1997