您好,请登录, 注册


当前位置: 首页 > 技术文献 > 解决方案 > 近红外光谱无损预测苹果可溶性固形物和干物质含量


来源: Yiyi Zhang et al  发布日期: 2020-10-09  访问量: 190

标签: 苹果,可溶性固形物,干物质,气调保鲜

Non-destructive prediction of soluble solids and dry matter concentrations in apples using near-infrared spectroscopy

Yiyi Zhang, Jacqueline F. Nock, Yosef Al Shoffe, Christopher B. Watkins⁎
Horticulture Section, School of Integrative Plant Science, College of Agriculture and Life Sciences, Cornell University, Ithaca, NY, 14853, USA

Soluble solids content (SSC) is an important factor for assessing quality of apples as it is linked to consumer taste preferences. Fruit dry matter content (DMC) is dominated by soluble sugar and starch concentrations at harvest, and therefore the DMC at the time of harvest can be strongly correlated with the post-storage SSC. The objective of this study was to develop models based on near-infrared (NIR) spectroscopy using a commercially available handheld instrument to predict SSC and DMC of fruit at harvest and after storage. ‘Gala’, ‘Honeycrisp’, ‘McIntosh’, ‘Jonagold’, ‘NY1′, ‘NY2′, ‘Red Delicious’ and ‘Fuji’ apples were tested. Partial least square regression was used to build calibration models for prediction of SSC and DMC. Models were also built for individual and multiple cultivars. Internal and external validations were applied to test the accuracy and precision of both models. In general, the individual- and multi-cultivar models have similar calibration performance. In internal validations, R2 and RMSE from multi-cultivar and individual-cultivar models were similar, but the slope values were higher in individual-cultivar than multi-cultivar models, indicating that the prediction using individual-cultivar model was more accurate. However, for individual-cultivar models, data-overfitting and the reference values distribution may lead to poor prediction in external validation. Overall the results support use of a portable NIR-based instrument to predict SSC and DMC, but to obtain precision and accurate predictions, calibration models should be built based on individual cultivars and the variability from seasonal and regional effects have to be taken into consideration.

可溶性固形物含量(SSC)与消费者的口味偏好有关,是评价苹果品质的重要因素。果实干物质含量(DMC)主要由可溶性糖和淀粉含量决定,因此采收时的DMC与贮藏后的SSC密切相关。本研究的目的是建立基于近红外(NIR)光谱的模型,使用商用手持器预测水果收获和贮藏后的SSC和DMC。对“Gala”、“Honeycrisp”、“McIntosh”、“Jonagold”、“NY1”、“NY2”、“Red Delicious”和“Fuji”苹果进行了测试。采用偏最小二乘回归法建立了SSC和DMC预测的校正模型。同时建立了单个和多个品种的模型。通过内部和外部验证来检验两种模型的准确性和精度。一般来说,单个品种和多品种模型具有相似的校准性能。在内部验证中,多品种模型和单个品种模型的R2和RMSE相似,但单个品种的斜率值高于多品种模型,说明用单个品种模型预测更准确。然而,对于单个品种模型,数据的过度拟合和参考值的分布可能导致外部验证的预测性较差。总的来说,结果支持使用基于近红外的便携式仪器来预测SSC和DMC,但是为了获得精确和准确的预测,应基于单个品种建立校准模型,并且必须考虑季节和区域效应的变化。