Vol. 54, Issue 3, pp. 353-363 (2024)

Abstract

In order to minimize errors during the LED spectrum fitting process and achieve a closer match between the fitted spectrum and the target spectrum, a method employing double Gaussian functions for fitting is proposed. This approach, in comparison to the widely applied modified Gaussian model, avoids issues such as sidelobe uplift when dealing with narrow-band LED spectra. It demonstrates effective fitting for narrow-band LED spectra, exhibiting higher fitting accuracy at the bottom of the spectrum compared to the simple Gaussian model. Moreover, this fitting method introduces no additional fitting variables. The model is constructed by combining two Gaussian functions with weighted summation, and optimizing the weighted coefficients of the two Gaussian functions achieves the best fitting results. Finally, the proposed model is applied to fit various LED spectra, yielding satisfactory results. The outcomes indicate that the overall fitting performance of the double Gaussian model surpasses that of the Gaussian model and the modified Gaussian model.