《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (9): 2904-2909.DOI: 10.11772/j.issn.1001-9081.2022091360

• 多媒体计算与计算机仿真 • 上一篇    下一篇

基于事件相机的雨滴检测算法

杨君宇1, 董岩1, 龙镇南1, 杨新2, 韩斌1()   

  1. 1.华中科技大学 机械科学与工程学院,武汉 430074
    2.广东省智能机器人研究院,广东 东莞 523000
  • 收稿日期:2022-09-20 修回日期:2022-10-31 接受日期:2022-11-02 发布日期:2023-01-16 出版日期:2023-09-10
  • 通讯作者: 韩斌
  • 作者简介:杨君宇(1997—),男,湖北荆门人,硕士,主要研究方向:机器视觉、图像处理
    董岩(1996—),男,河北石家庄人,博士研究生,主要研究方向:计算机视觉、3D感知
    龙镇南(1999—),男,江苏苏州人,硕士研究生,主要研究方向:无人机避障
    杨新(1970—),男,上海人,工程师,博士,主要研究方向:机器人路径规划;
  • 基金资助:
    数字制造装备与技术国家重点实验室自主课题基金资助项目(DMETZZ2022108)

Rain detection algorithm based on event camera

Junyu YANG1, Yan DONG1, Zhennan LONG1, Xin YANG2, Bin HAN1()   

  1. 1.School of Mechanical Science and Engineering,Huazhong University of Science and Technology,Wuhan Hubei 430074,China
    2.Guangdong Intelligent Robotics Research Institute,Dongguan Guangdong 523000,China
  • Received:2022-09-20 Revised:2022-10-31 Accepted:2022-11-02 Online:2023-01-16 Published:2023-09-10
  • Contact: Bin HAN
  • About author:YANG Junyu, born in 1997, M. S. His research interests include machine vision, image processing.
    DONG Yan, born in 1996, Ph. D. candidate. His research interests include computer vision, 3D sensing.
    LONG Zhennan, born in 1999, M. S. candidate. His research interests include drone obstacle avoidance.
    YANG Xin, born in 1970, Ph. D., engineer. His research interests include robot path planning.
  • Supported by:
    Independent Project Fund of State Key Laboratory of Digital Manufacturing Equipment and Technology(DMETZZ2022108)

摘要:

图像除雨算法一般对单帧图像或视频流中的雨滴进行去除,以降低雨滴对视觉任务的不良影响。然而,由于雨滴下落速度极快,基于帧的相机无法获取雨滴在时间上的连续性,且相机的曝光时间和运动模糊进一步降低了图像中雨滴的清晰度,导致传统图像的除雨算法无法准确检出雨滴覆盖区域。为探究图像除雨的新思路,利用事件相机极高采样率、无运动模糊的特性,分析并建立了雨滴事件生成模型,并提出了基于时空关联性的事件相机雨滴检测算法。该算法通过分析事件相机记录下的每个事件与相邻事件之间的时空关系来对每个事件产生自雨滴运动的概率进行计算,从而实现雨滴检测。在三种降雨场景上的实验结果表明,在相机静止不动时,所提算法的雨滴检测正确率可达95%以上,误检率低于5%;当相机处于运动状态时,所提算法仍可达到95%以上的正确率与不超过20%的误检率。说明所提算法可有效检出雨滴。

关键词: 事件相机, 图像处理, 雨滴识别, 目标检测, 时空关联性

Abstract:

To reduce the harmful effects of rain for visual tasks, rain removal algorithms are commonly utilized on single frame images or video streams to remove rain. However, since rain falls extremely fast, frame-based cameras cannot capture the temporal continuity of rain, and the fixed exposure time and motion blur further reduce the sharpness of the rain on images, as a result, the traditional image rain removal algorithms cannot detect rain coverage areas accurately. In order to explore the new idea of image rain removal, a rain event generation model was constructed and a rain detection algorithm for event camera based on spatial-temporal relevance was proposed by using the characteristics of event camera: extremely high sampling rate and no motion blur. In this algorithm, the probability of each event generated by rain movement was calculated by analyzing the spatial-temporal relationship between each event recorded by the event camera and adjacent events, so as to achieve rain detection. Experimental results on three rainfall scenes show that when the camera is static, the proposed algorithm can reach more than 95% rain detection true positive rate, and the false positive rate less than 5%, and when the camera moves, the proposed algorithm can still reach more than 95% true positive rate and no more than 20% false positive rate. The above shows that the rain can be detected effectively by the proposed algorithm.

Key words: event camera, image processing, rain detection, object detection, spatial-temporal relevance

中图分类号: