大气与环境光学学报, 2009, 4 (1): 30, 网络出版: 2010-05-26
气溶胶粒子等效复折射率的BP神经网络建模
BP Neural Network Model of the Equivalent Complex Refractive Index of Aerosol
大气光学 BP神经网络 等效复折射率 散射系数 吸收系数 atmospheric optics BP neural network equivalent refractive index scattering coefficient absorption coefficient
摘要
利用改进的BP算法,结合Mie散射理论, 建立了波 长、散射系数、吸收系数与气溶胶粒子等效复折射率之间的非线性映射关系,实现 了反演实际大气非均匀气溶胶粒子等效复折射率的BP神经网络模型的建模。利用该 模型对气溶胶粒子等效复折射率进行了反演,根据反演得到的等效复折射率计算了 其光学特征量,计算结果与实测结果基本吻合。
Abstract
Based on the improved BP algorithm and Mie scattering theory, the nonlinear relation of wavelength, scattering coefficient, absorption coefficient and equivalent complex refractive index is developed. The BP neural network model of the equivalent refractive index of aerosol is also developed. The optical properties of aerosol obtained from inversion refractive index, such as scattering coefficient, absorption coefficient, are in good agreement with measurement results.
亓秀永, 黄印博, 方帅, 饶瑞中. 气溶胶粒子等效复折射率的BP神经网络建模[J]. 大气与环境光学学报, 2009, 4(1): 30. QI Xiu-yong, HUANG Yin-bo, FANG Shuai, RAO Rui-zhong. BP Neural Network Model of the Equivalent Complex Refractive Index of Aerosol[J]. Journal of Atmospheric and Environmental Optics, 2009, 4(1): 30.