中文摘要：

为了更有效地消除岩心高光谱数据中的噪声，提出基于奇异值分解的岩心高光谱数据降噪方法，引入奇异值下降率的概念，利用奇异值下降率单调性的突变点来确定表征信号有用奇异值的个数。用该方法对地物光谱仪ASDField.Spec.4实地采集到的岩心高光谱数据进行降噪处理，并与依据奇异值相对强度确定奇异值突变点的降噪方法进行对比，利用均方根误差（RMSE）、信噪比（SNR）两项指标对降噪效果进行评价。实验结果表明，该方法更能提高信噪比，降低均方根误差，更能有效保持原始岩心高光谱曲线的吸收特征，消除高光谱曲线上的毛噪现象。

英文摘要：

In order to eliminate the noise in the core hyperspectral data more effectively,a new method about core hyper?spectral data denoising based on singular value decomposition is proposed,in which the concept of singular value decline rate isbrought. The number of useful singular value of the characterization signal is determined by the abrupt change point of the singu?lar value decline rate. This method is used to denoise the core hyperspectral data collected by ASD FieldSpec? 4,and comparedwith the denoising method that relies on the singular value relative strength to determine singular value mutation point. Its denois?ing effect was evaluated with the root mean square error(RMSE)and the signal?to?noise ratio(SNR). The results show that themethod can improve SNR,reduce RMSE,keep the absorption characteristics of original core hyperspectral curve more effective?ly,and eliminate the frizz phenomenon of core hyperspectral curve.