小波包熵的复杂体系近红外光谱信息提取

发布时间：2018-09-14 20:13

【摘要】：有用信息提取是复杂体系近红外检测的重点和难点之一。由于复杂体系光谱中存在各种噪声、基线漂移、谱带重叠及复杂背景的干扰,常规方法不能准确地从光谱中获得有用信息。为此,将小波包变换(DWPT)和信息熵理论相结合——小波包熵(EWPIE)提取复杂体系光谱中的有用信息。思路是采用小波包变换对光谱信号进行多频带分解,根据有用信号与噪声的频带分布特点,基于信息熵理论滤除干扰的频率分量,采用正交校正法(OSC)剔除与被测组分无关的信息,然后对处理后的频率分量进行重构,从而实现复杂体系有用信息的准确提取。通过对复杂体系光谱数据建立多元校正模型来验证该方法的效果。采用牛奶的近红外光谱数据,以牛奶中脂肪和蛋白质浓度为研究对象,建立了偏最小二乘法(PLS)模型。结果显示,牛奶中脂肪和蛋白质的预测均方根误差(RMSEP)分别为0.132%和0.121%,与单纯的DWPT和OSC相比,EWPIE能够有效地提取有用信息,避免了无用信息的干扰,明显提高了模型的预测精度,对复杂体系的准确检测具有一定的理论意义和实际应用价值。
[Abstract]:The extraction of useful information is one of the most important and difficult problems in NIR detection of complex systems. Because of the noise, baseline drift, spectral band overlap and complex background interference in the spectrum of complex systems, the conventional methods can not accurately obtain useful information from the spectrum. For this reason, wavelet packet transform (DWPT) and information entropy theory (EWPIE) are combined to extract useful information from the spectra of complex systems. The idea is to use wavelet packet transform to decompose the spectral signal in multi-band. According to the characteristics of frequency distribution of useful signal and noise, the frequency component of interference is filtered based on information entropy theory. The orthogonal correction method (OSC) is used to remove the information independent of the measured components, and then reconstruct the frequency components after processing, so that the useful information of complex systems can be extracted accurately. The multivariate calibration model of complex system spectral data was established to verify the effectiveness of the method. Based on the near infrared spectrum data of milk and the concentration of fat and protein in milk, a partial least square (PLS) model was established. The results showed that the root mean square error (RMSEP) of fat and protein prediction in milk was 0.132% and 0.121%, respectively. Compared with DWPT and OSC alone, EWPIE could extract useful information effectively, avoid the interference of useless information, and improve the prediction accuracy of the model. It has certain theoretical significance and practical application value for accurate detection of complex system.
【作者单位】：河南工业大学粮油食品学院;
【基金】：国家自然科学基金项目(31601537) 河南省教育厅科学技术研究重点项目(14A550016)资助
【分类号】：O657.33

【相似文献】