摘要
透射式能见度仪的校准结果无法溯源到世界气象组织对气象光学视程(MOR)的定义,如何实
现 MOR 定义中 2700 K 白炽灯的光谱模拟已成为透射式能见度仪定标技术亟待解决的关键
问题,故对透射式能见度仪定标光源的光谱模拟方法进行了研究。首先,从理论上分析了光
谱不匹配对透射式能见度仪定标的影响,建立了满足透射式能见度仪定标精度要求的光源光
谱分布判别依据,得出透射式能见度仪定标光源的光谱模拟误差应优于±5.5%的结论。然后,
基于一种数字微镜器件(DMD)的透射式能见度仪定标光源系统,在遗传算法基础上,提出了一
种基于无余均分机制的光谱模拟方法,以均分 800 列 DMD 微镜阵列面为前提,通过不断均匀
划分 DMD 阵列面的方式增加光谱拟合单元数量,实现了 2700 K 绝对色温的光谱模拟。最
后,测试了 20,50,80 个基础光谱拟合单元对应的定标光源的光谱分布,并进行了不确定度分
析。测试结果表明,80 个基础光谱拟合单元对应的定标光源光谱模拟误差为±5.3%,扩展不
确定度为 4.04%。
Abstract
The calibration results of the transmissometers cannot be traced to the definition of
meteorological optical range (MOR) of the World Meteorological Organization. How to
realize the spectrum simulation of the 2700 K incandescent lamp in the MOR definition
has become a key issue that needs to be solved urgently in the calibration technology of
the transmissometers, so the spectral simulation method of calibration light source of the
transmissometers is studied. First, the influence of spectral mismatch on the calibration
of the transmissometers is analyzed theoretically, and the basis for discriminating the
spectral distribution of the light source that meets the calibration accuracy requirements
of the transmissometers is established. It is concluded that the spectral simulation errors
of the calibration light sources of the transmissometers should be better than ±5.5%.
Then, based on a transmissometer calibration light source system with digital micromirror
device (DMD), and on the basis of genetic algorithm, a spectral simulation method based
on a uniform division mechanism without residues is proposed. On the premise of
dividing uniformly the array surface of DMD micro mirrors into 800 columns, the number
of fitting units is increased by uniformly dividing the DMD array surface, and and the
spectral simulation of absolute color temperature of 2700 K is realized. Finally, the
spectral distribution of the calibration light source corresponding to 20, 50, and 80 basic
spectral fitting units is tested, and the uncertainty is analyzed. The test result show that
the spectral simulation error of the calibration light source corresponding to the 80 basic
spectral fitting units is ±5.3%, and the expanded uncertainty is 4.04%.