Study On Signal Acquisition & Proessing Method Of The Micro Spectrometer

Spectrometer is a kind of precision optics analysis instrument analyzing chemical components and its content of substances of various kinds. It is extensively applied in such fields as modern scientific experiments, biological research, medical science & medicine study, industrial and agricultural production, national defense, astronomy observation, etc. As an important part in spectrometer design, signal acquisition and software processing play a role of great significance here.As a sub-project of " Micro-spectrometer Analysis System" sponsored by National Science Foundation, it analyzes present situation for micro-spectrometer research, utilizes digital-signal processing techniques, software engineering theory and Windows visible programming techniques to develop then spectrum signal acquisition and processing software aiming at laying a proper foundation for micro-spectrum application The Main tasks as follows:1. Develop spectrum signal acquisition software in a discontinuous way and achieve real time of spectrum signal acquisition and simultaneity of signal processing after analyzing hardware characteristics of spectrum acquisition system and investigating programming techniques and multithreading software design methods under Window protection.2. Design the signal pre-treatment model and make use of such methods as Fourier transfer and wavelet analysis, etc to perform filtering treatment for spectrum signals. The experiment results indicate that changing filtering parameters can not only eliminate high-frequency noise and low-frequency disturbance but also keep detailed signals of spectrum untouched. derivative deconvolution is utilized to separate overlapped spectra so as to improve resolving power.3. Program to get the peak by evaluation zero of the first-order derivative; the space coordinate of the spectrum signal can be transform into wavelength one by using solve the regression equation.4.The dissertation also designs data structure in spectrum data system and proposes that linear discriminant function can be used to determine the nature of a specific substance.