The invention relates to a vehicle target detection method based on thunder view data credibility analysis. The method comprises the steps of millimeter wave radar data acquisition, preprocessing and vehicle target detection. Collecting and processing video data and detecting a vehicle target; carrying out space-time alignment on the thundersight vehicle target; carrying out adaptive fusion on the credibility analysis of the thunder view data and a vehicle target; and vehicle target visualization and protocol transmission. The invention also discloses a vehicle target detection device based on the thunder view data credibility analysis for realizing the method. The method can be better applied to a complex traffic scene, and a video detection result is emphatically utilized in a scene with limited radar perception by counting a radar data credibility score and a video data credibility score; in a scene with limited video perception, a radar data result is emphasized to be utilized; the radar data and the video data are fully fused under the condition that the perception of the radar data and the video data is not limited, and accurate and stable vehicle target detection data are provided.
本发明是一种基于雷视数据可信度分析的车辆目标检测方法,所述方法包括毫米波雷达数据采集、预处理及车辆目标检测;视频数据采集、处理及车辆目标检测;雷视车辆目标的时空对齐;雷视数据可信度分析与车辆目标的自适应融合;车辆目标的可视化与协议传输。本发明还公开了实现该方法的基于雷视数据可信度分析的车辆目标检测装置。本发明能更好应用于复杂的交通场景,通过统计雷达数据可信度分数及视频数据可信度分数,在雷达感知受限的场景下,偏重利用视频检测结果;在视频感知受限的场景下,偏重利用雷达数据结果;在二者感知不受限的情况下,充分融合雷达数据与视频数据,提供准确、稳定的车辆目标检测数据。
Vehicle target detection method and device based on thundersight data credibility analysis
基于雷视数据可信度分析的车辆目标检测方法与装置
2024-02-13
Patent
Electronic Resource
Chinese
Expressway truck intrusion detection method based on thundersight trajectory fusion
European Patent Office | 2024
|METHOD FOR SHIFT TORQUE CREDIBILITY GAP COMPENSATION OF HYBRID VEHICLE
European Patent Office | 2016
|