Sumi Yasushi, Kim Bong Keun, Ogure Takuya, Kodama Masato, Sakai Naoki, Kobayashi Masami
National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba 305-8568, Japan.
Altech Corporation, Yokohama 220-6218, Japan.
Sensors (Basel). 2024 Apr 24;24(9):2713. doi: 10.3390/s24092713.
This study comprehensively investigates how rain and drizzle affect the object-detection performance of non-contact safety sensors, which are essential for the operation of unmanned aerial vehicles and ground vehicles in adverse weather conditions. In contrast to conventional sensor-performance evaluation based on the amount of precipitation, this paper proposes spatial transmittance and particle density as more appropriate metrics for rain environments. Through detailed experiments conducted under a variety of precipitation conditions, it is shown that sensor performance is significantly affected by the density of small raindrops rather than the total amount of precipitation. This finding challenges traditional sensor-evaluation metrics in rainfall environments and suggests a paradigm shift toward the use of spatial transmittance as a universal metric for evaluating sensor performance in rain, drizzle, and potentially other adverse weather scenarios.
本研究全面调查了降雨和毛毛雨如何影响非接触式安全传感器的目标检测性能，这种传感器对于无人机和地面车辆在恶劣天气条件下的运行至关重要。与基于降水量的传统传感器性能评估不同，本文提出空间透过率和粒子密度作为更适用于降雨环境的指标。通过在各种降水条件下进行的详细实验表明，传感器性能受小雨滴密度的影响显著，而非降水量的总量。这一发现挑战了降雨环境下传统的传感器评估指标，并表明了一种范式转变，即转向使用空间透过率作为评估降雨、毛毛雨以及潜在其他恶劣天气场景中传感器性能的通用指标。
