通过近红外漫反射光谱技术对甜柿的货架期进行了定性判别研究,运用不同光谱预处理方法和不同波段选择,发现在1 100~2 400 nm内,采用二阶导数结合标准正常化和去散射(SNVD)处理的光谱预处理方法最好。判别模型的正确分类率达到97.8%~100%,预测准确率达到88.9%~100%。因此,近红外光谱技术对甜柿货架期的检测具有应用价值。

The near-infrared spectroscopy was used in qualitative determination of sweet persimmon shelf-life. The calibration results for shelf-life of sweet persimmon were compared by different spectra pretreatment and different spectral ranges. The second derivative with standard normal variation and detrend spectral pretreatment method were proved to be the best within the range between 1 100 nm to 2 500 nm. The correct identification rate of discriminate model is 97. 8% to 100%,predict accuracy is 88. 9% to 100%. Therefore,the qualitative identification of sweet persimmon shelf-life by near infrared spectroscopy is applicable.