为了实现在近红外波段测量介质中颗粒的粒径，提出了一种基于消光光谱结合数据库技术的颗粒粒径反演方法。首先通过全散射消光的Lambert-Beer定律计算生成与颗粒粒径大小及分布相关的消光光谱数据库，然后用数据库处理算法比较数据库中数据与实测颗粒消光光谱数据，两者最佳吻合所对应的数据库中颗粒粒度及分布参数即为所求值。分别对颗粒的单峰和双峰粒度分布进行数值仿真，仿真结果的体积频率分布的相对均方根误差小于10%。利用3种不同粒度的国家标准颗粒进行实验验证，实测得到粒径D50的相对误差均在10%以内，仿真及实验对所提算法的验证得到了满意的结果。研究表明，将数据库技术及算法应用于近红外窄波段消光法颗粒粒度的测量是可行的和有效的。

In order to measure the particle size in the medium using near-infrared light source, a particle size measurement method based on light extinction method combined with database technology is proposed. The extinction spectrum database related to different particle size and distribution is first calculated and generated by the total scattering extinction Lambert-Beer law. The database processing algorithm is employed to compare the data in the database with the measured particle extinction spectrum data. Then the particle size and distribution parameters of the measured samples are determined from the calculated values of the best match in the database. Next, numerical simulations of both the unimodal and bimodal particle size distribution are carried out, and the relative root mean square error of the volume frequency distribution of the simulation results is less than 10%. Three kinds of national standard particles with different particle sizes are used for experimental verification, and the relative errors of the measured particle size D50 are all less than 10%. Thus the validity of the proposed method and algorithm is demonstrated through simulations and experiments with satisfactory results. The research in this paper shows that the application of database technology and algorithm to the measurement of particle size by near-infrared narrow-band extinction method is feasible and effective.