微型光谱仪的数据噪声处理  

Signal Processing Research of Noisy Data in Micro Spectrometer

作　　者：谢印忠[1] 

机构地区：[1]临沂大学信息学院,山东省临沂市双岭路中段276005

出　　处：《光谱实验室》2011年第6期3067-3072,共6页Chinese Journal of Spectroscopy Laboratory

摘　　要：原始信号的噪声处理是微型光谱仪器数据处理的重要组成部分,为了提高在弱光照射下便携式微型光谱仪的工作性能,在分析了线阵CCD探测器的各种噪声来源基础上,指出依靠硬件电路去除噪声在微型仪器设备应用的局限性。虽然判断一段离散数据曲线上噪声大小可以采用解析几何或者傅里叶变换得到特征频谱来判断,但是运算复杂,不利于在单片机应用中编程。通过对Freeman方向链码的分析,提出利用Freeman方向链码判断一段曲线上的噪声程度,量化后作为平滑窗口尺寸选择的依据。在处理线阵CCD输出信号中将该方法与固定窗口曲线平滑或者FFT数字滤波器进行了比较,证明了其可行性。Freeman链码的应用简化了在单片机编程中使用数字滤波器的大模块的编程,降低了编程难度,同时使信号噪声的处理仅仅在数学运算阶段中,不引入其他的电路噪声,从而达到比前期增加电子元件处理的方法能更好地抑制噪声的目的。The original signal noise treatment is an important part of data processing in miniature spectral instrument.In order to improve the performance of the portable spectrometer at low light levels,and the linear CCD detectors were analyzed based on all kinds of noise sources,and pointed out the application limitations that relied on hardware circuit for removing noise in the micro instrument equipment.Although noise level of a discrete data curve base on characteristics from analytic geometry or Fourier transform were judged,but the operations are too complex to make program with single chip computer.Based on the analysis of the Freeman chain direction code,a Freeman direction chain code for a curve of judgment was proposed,quantization noise level as smooth window after size selection basis.In dealing with linear CCD output signal,the new method and fixed window curve smooth or FFT digital filters were compared,and feasibility was proved.The application of Freeman chain code simplified the use of digital filter programming of big modules,and reduced the programming difficulty in single chip computer,and also made the signal noise treatment only in the mathematical operation stage,no longer imported other circuit noise,thus reached better noise restraining purpose than previous increasing electronic components in signal processing.

关 键 词：光谱学 信号处理 电荷耦合器件 FREEMAN链码 

分 类 号：TH744.1[机械工程—光学工程]