基于深度学习的毫米波雷达检测、分割、跟踪等下游任务的各类论文、资料整理

持续更新中！

交流可加v:naca_yu，加入毫米波雷达相关交流群（非广告、盈利），请备注：csdn+名称，谢谢！

一个人更新不易，欢迎大家补充，帮助到你的话，感谢你能点个Star！

https://github.com/nacayu/awsome-deeplearning-based-radar-perception

---

近年毫米波雷达相关论文 & 多模态融合相关论文

论文名称	任务类别	传感器	源码地址	已更新解读	备注
MmWave Radar and Vision Fusion for Object Detection in Autonomous Driving: A Review	综述	radar&camera			自动驾驶中radar相关的多传感器融合综述
Towards Deep Radar Perception for Autonomous Driving: Datasets, Methods, and Challenges	综述	radar			面向自动驾驶的深度雷达感知:数据集、方法和挑战综述
Radar-PointGNN: Graph Based Object Recognition for Unstructured Radar Point-cloud Data	3D检测	radar		https://zhuanlan.zhihu.com/p/549641548	基于GNN
2D Car Detection in Radar Data with PointNets	3D检测	radar		https://zhuanlan.zhihu.com/p/568160922	改进PointNets
Bridging the View Disparity of Radar and Camera Features for Multi-modal Fusion 3D Object Detection	3D检测	camera&radar		https://zhuanlan.zhihu.com/p/568160922
RadSegNet: A Reliable Approach to Radar Camera Fusion	分割	camera&radar		https://zhuanlan.zhihu.com/p/568160922
Depth Estimation From Monocular Images and Sparse Radar Using Deep Ordinal Regression Network	深度估计	camera&radar	https://github.com/lochenchou/DORN_radar	https://zhuanlan.zhihu.com/p/568160922	序数回归，改进自DORN
A Simple Baseline for BEV Perception Without LiDAR（2022， MIT）	分割	camera&radar	https://github.com/aharley/simple_bev	https://zhuanlan.zhihu.com/p/568160922	对比LSS…
See Through Smoke: Robust Indoor Mapping with Low-cost mmWave Radar	密集点云生成GAN	lidar & radar		https://zhuanlan.zhihu.com/p/568160922
Radar Occupancy Prediction With Lidar Supervision While Preserving Long-Range Sensing and Penetrating Capabilities	可行区域生成：freespace	lidar & radar		https://zhuanlan.zhihu.com/p/568160922
RADIANT: Radar-Image Association Network for 3D Object Detection	3D检测	camera&radar	https://github.com/longyunf/radiant	https://zhuanlan.zhihu.com/p/597739906	一种全新的毫米波雷达图像关联网络用于3D目标检测
CRFNet for Object Detection (Camera and Radar Fusion Network)	2D检测	camera&radar	https://github.com/nacayu/CRFNet_Tensorflow2.4.1	https://zhuanlan.zhihu.com/p/112578232	基于yolov3
A frustum proposal-based 3D object detection network for multi-stage fusion in autonomous driving	3D检测	camera&radar	https://github.com/brandesjj/centerfusionpp	https://zhuanlan.zhihu.com/p/603398636	基于centerfusion改进的下一代毫米波雷达与视觉融合方案
CRAFT: Camera-Radar 3D Object Detection with Spatio-Contextual Fusion Transformer	3D检测	camera&radar		https://zhuanlan.zhihu.com/p/581055339	一种基于空间-语义信息互补的毫米波雷达与相机融合3D检测方法
SAF-FCOS: Spatial Attention Fusion for Obstacle Detection using MmWave Radar and Vision Sensor	2D检测	camera&radar	https://github.com/Singingkettle/SAF-FCOS	即将更新	基于FCOS
NVRadarNet: Real-Time Radar Obstacle and Free Space Detection for Autonomous Driving	可行区域生成：Freespace	lidar&radar		https://zhuanlan.zhihu.com/p/575385783	实时(1.5ms)BEV多任务
CenterFusion: Center-based Radar and Camera Fusion for 3D Object Detection	3D检测	camera&radar	https://github.com/mrnabati/CenterFusion	https://zhuanlan.zhihu.com/p/508905129	基于CenterNet
Multi-Modal Fusion Transformer for End-to-End Autonomous Driving	自动驾驶车辆路径预测	camera&radar	https://github.com/autonomousvision/transfuser	https://zhuanlan.zhihu.com/p/508898376	基于transformer 语义信息注意力关联
K-Radar: 4D Radar Object Detection for Autonomous Driving in Various Weather Conditions	3D检测	camera&4d radar	https://github.com/kaist-avelab/k-radar	即将更新	4D雷达
Improved Orientation Estimation and Detection with Hybrid Object Detection Networks for Automotive Radar	2D检测 in BEV	radar		即将更新	结合基于网格和基于点的处理方法
CramNet: Camera-Radar Fusion with Ray-Constrained Cross-Attention for Robust 3D Object Detection	3D检测	radar&camera		即将更新	射线约束交叉注意机制 考虑传感器短时间失灵
DeepFusion: A Robust and Modular 3D Object Detector for Lidars, Cameras and Radars	3D检测	lidar&camera&radar		https://zhuanlan.zhihu.com/p/578655032	特征提取模块化设计 每个传感器模块设计auxiliary loss
GRIF Net: Gated Region of Interest Fusion Network for Robust 3D Object Detection from Radar Point Cloud and Monocular Image	3D检测	radar&camera		即将更新	设置模态融合阈值，自适应地选择较优输入，调节最终贡献 二阶段检测
RCDPT: RADAR-CAMERA FUSION DENSE PREDICTION TRANSFORMER	单目深度估计	radar&camera		即将更新
Radar Voxel Fusion for 3D Object Detection	3D检测	lidar&camera&radar		即将更新
CFTrack: Center-based Radar and Camera Fusion for 3D Multi-Object Tracking	3D目标跟踪	radar&camera		即将更新	end-to-end跟踪方法 基于centerfusion

数据集

数据集名称	传感器	解读	地址	备注
		原始数据
CRUW	radar	https://zhuanlan.zhihu.com/p/508896576	https://www.cruwdataset.org/	原始数据的自动驾驶数据集 支持tracking
CARRADA	camera radar		https://github.com/valeoai/carrada_dataset	支持tracking,segmentation
RADDet	radar		https://github.com/ZhangAoCanada/RADDet#Dataset
		radar点云数据集
nuscenes	radar camera lidar	https://zhuanlan.zhihu.com/p/508912923
RadarScenes	radar camera	https://zhuanlan.zhihu.com/p/424535286	https://radar-scenes.com/	提供point-wise点级别细粒度的标注
RADIATE	radar camera lidar	https://blog.csdn.net/CCChester/article/details/127565100	http://pro.hw.ac.uk/radiate/	支持tracking
Pointillism	radar camera lidar		https://github.com/Kshitizbansal/pointillism-multi-radar-data
Zendar	radar camera lidar		https://storage.googleapis.com/www.archive.zendar.io/dataset.html	支持tracking
		4D雷达数据集
TJ4DRadSet	radar camera lidar	https://blog.csdn.net/u013019296/article/details/127219137	https://github.com/TJRadarLab/TJ4DRadSet	支持tracking
K-Radar Dataset	camera images, Lidar point cloud, RTK-GPS, and Radar tensor		https://github.com/kaist-avelab/K-Radar

其他资料