Study Of LED Spectra Wavelet De-noising Based On MATLAB

The spectroscopic study of LED is important for analyzing the luminescence efficiency and aging mechanism of LED. Because of environmental factors and itself introducing to the spectral signal a lot of noise in measurements of the spectrum, making a large impact on the results of spectrum, it makes pretreatment of spectra de-noising as a necessary element of spectrum. In recent years, wavelet de-noising in signal processing is hot and cutting-edge, Wavelet toolbox in MATLAB software provides a wealth of function for all types of wavelet de-noising, however, the selection of different parameters of the function for signal denoising have a great influence on the de-noising results. The software uses wavelet threshold de-noising of the signal spectrum, assess the effect of de-noising in different parameters or parameter combinations by using mean square error, signal to ratio, spectral angle as the evaluation criteria.This paper measures the spectra of white LED in different operating current and UV LED in different junction temperature firstly, by using a WGD-8/8A type sectional and multi-function grating spectroscope. This paper designs a de-noising evaluation software based on'wden'function in graphical user interface development environment of MATLAB. Simulated results for white and UV LED spectrum signal using this software give corresponding optimal parameters spectra denoising. The results has some value in the pre-processing stage of the LED spectrum.