线阵光学降水现象自动测量系统

Optical precipitation auto-measurement system based on linear image sensor

采用平行光成像的方式,研制了基于线阵图像传感器的降水现象自动测量系统,用于揭示大气降水物理特征及本质。首先,根据降水粒子微物理特征和降水现象业务观测的需求,设计了光学系统及高速线阵图像数据采集处理平台。然后,针对降水观测系统连续运行条件下线阵实时采集数据量大、不便于传输及处理的特点,利用现场可编程门阵列(FPGA)芯片实现片内线阵数据实时二值化处理及编码传输。为消除平行光源性能漂移及传感器光强响应非均匀性等对成像阈值设置的影响,保证长期运行测量精度和一致性,设计了基于微处理器软核MicroBlaze的自适应阈值设置方案。在室内定标实验的基础上,与翻斗雨量计进行初步观测对比试验。实验室定标结果表明,系统最小分辨率优于0.1mm,实际同步观测累加降水量误差在15%左右,两者变化趋势大致相同,有较好的线性相关性。提出的系统满足降水现象自动测量对稳定可靠、自动化、精细化的要求。

Abstract. An optical precipitation auto-measurement system based on a linear image sensor was designed to reveal the physical mechanism and essence of the precipitation. First, according to physical characteristics of precipitation particles and the demand for precipitation observation, an optical unit and a data acquisition and processing platform for the high speed linear image sensor were designed. As the mass acquisition data were inconvenient for transmission and processing under a continuous operation, the real-time binarization and encoding processing of the linear image were carried out by Field Programmable Gate Array （FPGA）. To eliminate the influence of performance excursion of parallel light source and pixel photoelectric response nonuniformity on the setting thresholds and to ensure the measurement precision and coherence, an adaptive threshold setting scheme was proposed and realized based on a MicroBlaze soft-core processor in the FPGA. The comparison experiment with tipping bucket pluviometer based on calibration in a laboratory was carried out. Experimental result indicates that the resolution of precipitation auto-measurement system can achieve 0.1 mm, and the error of cumulative precipitation amount in synchronization observation is about 15 %. Obtained results can satisfy the requirements of precipitation auto-measurement for stabilization and high precision.