小波变换和连续投影算法在猪肉嫩度光谱法分析中的应用

The Application of Wavelet Transform and Successive Projections Algorithm in the Spectral Analysis of the tenderness of Pork

文章将小波变换与连续投影算法用于猪肉嫩度的近红外光谱分析中。先利用小波变换对64个猪肉样本的原始光谱进行降噪预处理,然后由连续投影算法提取了猪肉嫩度的18个特征波长,结果表明:小波变换可以降低猪肉光谱光谱数据的噪声干扰,基于连续投影算法提取的18个特征波长建立的猪肉嫩度模型与全光谱模型性能相当,但建模波长数量仅为全光谱建模的3.51%,提取的猪肉嫩度特征波长为开发便携式猪肉嫩度检测仪提供了参考。

In this paper wavelet transform and successive projections algorithm were applied together in the NIRS （near infrared spectoroseopy） of detection of pork tenderness . firstly,the primitive spectra of 64 pork samples were be pretreated through denoising by wavelet transform and then 18 characteristic wavelengths of pork tenderness were extracted by successive projections algorithm. The result showed wavelet transform can reduce the noise interference . And the performance of the model of pork tenderness built by the 18 characteristic wavelengths extracted based on successive projections algorithm was comparable to that of full spectra, while the number of modelling wavelengths was only 3.51% of full spectra modelling wavelengths . The extracted characteristic wavelengths of pork tenderness can be a reference for designing portable detector of pork tenderness.