Data Acquisition And Digital Signal Processing Of CCD Spectrometer

CCD(Charge Coupled Device) image detector is widely used in new spectrometers for its ability to detect a wide spectrum range in a single exposure.In the first part of our research we design a high-sensitivity high-resolution high-speed real-time and USB2.0-based linear CCD data acquisition system for spectral measurement. We choose a right CCD for this system according to the requirement of spectral-measureing application, choose a right A/D devise according to the real effective dynamic range of CCD. We achive the real-time transport of the spectrum data based on non-external-buffer structure and USB2.0. We use a CPLD(Complex Programable Logic Device) to drive the CCD and control the whole system, which is programed in Verilog Hardware Description language. The system is optimized on access timing and EMC to achieve the least noise and the best system stability. The system requires only a USB cable be connected when working, and can use it immediately after the connection cable is connected to the computer. The RS232 and Jteg interface enable the system to updata the USB firmware or programe the CPLD in system.In the second part of our research we do some tests on the spectrum digital signal processing, which include denoise of the fixed patten noise of CCD and the real effective dynamic range of CCD after this process, denoise of the random noise from two kinds of lighting source, with continuous spectrum or with striation spectrum. We try different denoise methods, which include FFT(Fast Fourier Transform Algorithm), Savitzky-Golay, Boxcar and Wavelet. A guidence of choosing different denoise method for different spectrum kind of lighting source is given below. Boxcar and Wavelet have the best porformence in denoising the contimuous spectrum signal, while Savitzky-Golay and Wavelet are doing well in denoising the striation spectrum signal, here the Wavelet-based denoise method suits both. The FFT does no get a high score when denoising neither of these two kinds of spectrum.