Vehicle Detection Based on Deep Neural Network Combined with Radar Attention Mechanism 2020-01-5171
In the autonomous driving perception task, the accuracy of target detection is an essential evaluation, especially for small targets. In this work, we propose a multi-sensor fusion neural network that combines radar and image data to improve the confidence level of the camera when detecting targets and the accuracy of the prediction box regression. The fusion network is based on the basic structure of single-shot multi-box detection (SSD). Inspired by the attention mechanism in image processing, our work incorporates the a priori knowledge of radar detection in the convolutional block attention module (CBAM), which forms a new attention mechanism module called radar convolutional block attention module (RCBAM). We add the RCBAM into the SSD target detection network to build a deep neural network fusing millimeter-wave radar and camera. The traditional SSD network and multi-sensor fusion network in this paper are trained simultaneously by using the recently introduced nuTonomy scenes (nuScenes) dataset that contains both radar and image data. The test results demonstrate that our algorithm improves the mean average precision (mAP) from 38.3% to 43.7%. And the results obtained from the dataset and realistic scenarios show that the RCBAM-based multi-sensor fusion network in this paper performs better in detecting small targets.
Citation: Bai, J., Zhang, Y., Huang, L., and Li, S., "Vehicle Detection Based on Deep Neural Network Combined with Radar Attention Mechanism," SAE Technical Paper 2020-01-5171, 2020, https://doi.org/10.4271/2020-01-5171. Download Citation
Author(s):
Jie Bai, Yifan Zhang, Libo Huang, Sen Li
Affiliated:
Tongji University
Pages: 7
Event:
Automotive Technical Papers
ISSN:
0148-7191
e-ISSN:
2688-3627
Related Topics:
Neural networks
Autonomous vehicles
Radar
Sensors and actuators
Machine learning
SAE MOBILUS
Subscribers can view annotate, and download all of SAE's content.
Learn More »