基于数字光处理技术的梨可溶性固形物含量的无损速测研究

Research on the rapid and non-destructive determination of soluble solid content of pears based on digital light procession

中文关键词:  数字光处理技术  近红外光谱  可溶性固形物含量  梨

英文关键词:digital light procession  near-infrared spectroscopy  soluble solid content  pear

中文摘要:

      目的 为降低近红外光谱仪器制造成本, 将近红外技术推广到农业生产一线, 检验自主集成水果品质无损快速分析仪实验样机性能。方法 以北京大兴产黄金梨、园黄梨为例, 利用基于数字光处理技术内核的实验样机采集数据, 采用偏最小二乘回归结合全交互验证算法分别建立黄金梨、园黄梨以及两种梨的可溶性固形物含量定量校正模型, 并采用外部验证集对模型预测性能做进一步验证。结果 黄金梨、园黄梨以及两种梨的可溶性固形物含量模型的测定系数R2分别为0.6136、0.6576、0.5105, RMSEC分别为0.71、0.79、0.87; 交互验证测定系数R2CV分别为0.5332、0.5076、0.4193, RMSECV分别为0.78、0.96、0.95; 外部验证集相关系数r分别为0.7239、0.6825、0.6550, RMSEP分别为0.83、1.03、0.94。结论 基于数字光处理技术内核自主集成的水果品质无损快速分析仪器在梨可溶性固形物含量的无损速测以及降低仪器制造成本方面具有一定的应用潜力。

英文摘要:

      Objective In order to decrease the manufacturing cost of the near-infrared (NIR) spectrometer to extend NIR technology to the frontline of agricultural production, and validate the autonomous integrating experimental prototype of the rapid and nondestructive analyzer for fruit quality. Methods Taking the Huangjin pear and Yuanhuang pear as the experimental objectives, based on the data collected by experimental prototype with the kernel of digital light procession (DLP) technology, the calibration models of soluble solid content (SSC) for Huangjin pear, Yuanhuang pear and the both, were developed by the partial least square (PLS) regression combined with full cross validation algorithm. Furthermore, the external validation sets were collected to validate the prediction performance of the models developed. Results For Huangjin pear, Yuanhuang pear and the both, the determination coefficients of calibration (R2) were 0.6136, 0.6576 and 0.5105 respectively; the RMSEC were 0.71, 0.79 and 0.87 respectively; the determination coefficients of full cross validation (R2CV) were 0.5332, 0.5076 and 0.4193 respectively; the RMSECV were 0.78, 0.96 and 0.95 respectively. Meanwhile, for the external validation set, the correlation coefficients (r) were 0.7239, 0.6835 and 0.6550 respectively; the RMSEP were 0.83, 1.03 and 0.94 respectively. Conclusion The autonomous integrating experimental prototype of the rapid and nondestructive analyzer for fruit quality based on DLP technology has certain application potential in the rapid and nondestructive determination of SSC for pears and the decreasement of the manufacturing cost of the instrument.