内有参考文献、matlab编程代码、解释说明 利用近红外光谱技术对某海产品的水分含量进行无损检测。通过420～1023nm近红外光谱采集了266个海产品样本的光谱信息。分别采用多元散射校正(MSC)、标准正态变量交换(SNV)、归一化(Normalize)、数据中心化(Mean centering)、标准化(Autoscales)、移动窗口平滑、Savitzky-Golay卷积平滑法、一阶导数(FD)、二阶导数(SD)等方法对光谱进行预处理，并采用PCA主成分分析 (principal component analysis)结合马氏距离法对近红外校正样品集中的异常样品进行剔除。剔除样本后使用偏最小二乘法(PLS)建立模型对样本进行定量分析。比较各种光谱预处理的方法以及剔除异常值的权重阈值的选取，获取最佳PLS模型便可对待测样本进行水分预测。

光谱预处理，基准线校准，Spectral preprocessing baseline calibration

在网上下载的一些关于PLS和光谱预处理的matlab程序，程序比较多，都放在一起的，有需要的可以看看

The data signal obtained by using the near-infrared spectroscopy detector includes the information of the component to be tested, including the noise of various instruments, such as high-frequency random noise, baseline drift, spurious information, background of the sample.Therefore, before data analysis, the specific signal measurement and sample system should be processed reasonably, and the influence of various non-target factors on the detection signal information should be weakened or even eliminated, which lays a foundation for establishing a stable and reliable mathematical model. Commonly used data preprocessing methods include data normalization processing (mean centering, normalization, standard normal transformation, etc.), high frequency noise filtering (convolution smoothing, Fourier transform, wavelet transform, etc.), differential derivation of signals, and Baseline correction, etc.