提出了一种将优化雷达点迹关联和雷达信息与相机图像信息融合决策的方法。使用微波雷达和可变焦相机,从带有噪声的原始雷达数据中提取出有效的目标信号,利用齐次坐标表示投影方程,实现雷达-视觉数据空间对齐,利用雷达点投影影响视觉ROI检测框的位置。实验结果表明,相对于传统方法,对于复杂环境中的目标融合,本文提出的方法融合检测成功率提升了1.5％～5.7％,在获取更精确目标属性的同时消除了环境杂波干扰。
Abstract
In the paper,a method of optimizing radar plot correlation and fusion decision of radar information and camera image information is proposed.Microwave radar and zoom camera are used to extract effective target signals from raw radar data with noise.Projection equations are expressed in homogeneous coordinates to achieve radar visual data space alignment.Radar point projection is used to affect the position of visual ROI detection frame.The experiment results show that,for target fusion in complex environment,the fusion detection method proposed in this paper improves the fusion detection success rate by 1.5%~5.7% compared with traditional methods,and can obtain more accurate target attributes while reducing more environmental clutter interference.
关键词
微波雷达 /
数据融合 /
目标检测 /
单目相机
Key words
microwave radar /
data fusion /
target detection /
monocular camera
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 范邓楠.基于微波雷达与视频融合技术的车辆检测系统研究[D].桂林:桂林电子科技大学,2020.
[2] Wei Z,Zhang F,Chang S,et al.MmWave Radar and Vision Fusion based Object Detection for Autonomous Driving:A Survey[J].2021.
[3] Michaelis C,Mitzkus B,Geirhos R.Benchmarking robustness in object detection: Autonomous driving when winter is coming.arXiv,2019.
[4] 朱玉成.智能驾驶汽车毫米波雷达应用分析[J].中国新通信,2021,23(2):2.
[5] NAGASAKU T,KOGO K,SHINODA H,ET AL.77GHZ LOW-COST SINGLE-CHIP RADAR SENSOR FOR AUTOMOTIVE GROUND SPEED DETECTION[C]//IEEE COMPOUND SEMICONDUCTOR INTEGRATED CIRCUITS SYMPOSIUM.IEEE,2008.
[6] 袁赛柏,金胜,朱天林.MIMO 雷达发展技术综述[J].现代雷达,2017,39(8):5-8.
[7] 孙扬,杨贺.无人驾驶车辆智能水平等级划分[J].科技导报,2017,35(17):82-85.
[8] 刘任.机场周界监视雷达入侵目标检测与跟踪技术研究[D].成都:电子科技大学,2018.
[9] 万仁杰.毫米波线性调频连续波汽车雷达关键信号处理技术研究[D].南京:南京理工大学,2020.
[10] 黄海.视频与雷达数据融合在周界入侵报警的应用探讨[J].智能建筑与智慧城市,2019(6):37-39.
[11] Guo X,Du J,Gao J,et al,Pedestrian Detection Based on Fusion of Millimeter Wave Radar and Vision[C]//In Proceedings ofthe 2018 International Conference on Artificial Intelligence and Pattern Recognition,Beijing,China,2018:38-42.
[12] 冯勋勋.车辆目标跟踪系统设计与交互行为预测方法研究[D].西安:西安理工大学,2021.
基金
*国家自然科学基金(No.62122070)。
PDF(2392 KB)
