多模态数据集论文阅读 | WRDI: A Multimodal Dataset of mmWave Radar Data and Image, 2021 IEEE Big Data

本文链接：https://blog.csdn.net/qazwsxrx/article/details/126676616

本文介绍了WRDI，一个包含5000帧的毫米波雷达和图像多模态数据集，旨在解决安全监控领域的目标检测和跟踪问题。WRDI弥补了现有数据集在复杂环境和场景多样性上的不足，提供室内、夜间、雨雾条件下的数据，支持对行人和车辆的精确标注，为开发更稳健的检测和跟踪模型奠定了基础。

摘要由CSDN通过智能技术生成

原文链接：https://mp.weixin.qq.com/s/nLPfErOciUnxhzXx9aI-yA

论文阅读：[1] Z. Wang, C. Jin, L. Sun, and Y. Li, “WRDI: A Multimodal Dataset of mmWave Radar Data and Image,” in 2021 IEEE International Conference on Big Data (Big Data), 2021, pp. 3209–3214. doi: 10.1109/BigData52589.2021.9671545.

Abstract

mmWave radar and image data fusion 在 Security monitoring field 作用很大
- real-time
- high-sensitivity
- 全天候全天时
- effectively filter out fake objects
挑战：lack of a comprehensive mmWave radar and image multimodal dataset
- 已成为 a bottleneck in developing robust object detection and tracking models
本文贡献： WRDI
- WRDI: a mmWave radar data and image multimodal dataset with 5000 frames
- 意义：the first comprehensive mmWave radar and image multimodal dataset for security monitoring field.

1 Introduction

安防( intrusion events )领域：mmWave Radar and Camera can complement each other

Camera
- main terminal device
- 优点：capture fine texture and rich color information of the object
- 缺点：difficult to receive enough information to under bad condition
mmWave Radar
- 优点: wide sensing range + strong obstacle penetration ability, 不受 weathers and lighting conditions 影响
- 缺点：sparse and lack rich scene information
$\Rightarrow$ fuse mmWave radar and optical sensor data in security monitoring system

数据融合 (camera + mmWave radar) 研究现状

现有研究
- [1] radar + lidar + camera: detection and tracking
- [2] fuse in feature extraction
- [3] radar + image
- [4] radar + image
- [5] radar + image
缺陷
- 不适用 ：most of the radar and image multimodal datasets are collected in urban traffic roads