多品种苹果可溶性固形物近红外无损检测通用模型研究
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

基金项目:


General near-infrared model of soluble solids content in multi-variety apples
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    苹果可溶性固形物含量(soluble solid content,SSC)是影响果实质量的重要因素,利用近红外光谱(NIR)技术则可以实现对苹果SSC的无损检测。为获得稳健的多品种苹果无损检测通用模型,本研究将模型更新方法与变量筛选方法相结合,对红富士(Red Fuji)、青苹果(Green apple)、黄元帅(Golden Delicious)、红玫瑰(Rose)和乐淇(Lokit)等5个品种苹果的SSC进行无损检测。结果表明,更新后的新模型可以实现对5个品种苹果的SSC的高精度预测,此时模型预测均方根误差(RMSEP)为0.698%,预测相关系数(Rp)为0.904,预测偏差为0.074%,范围误差比(RPD)可达2.340。为识别和提取光谱的重要信息波段,还采用竞争性自适应重加权算法(CARS)、连续投影算法(SPA)和无信息变量消除算法(UVE)3种波段筛选方法优化模型。结果表明,CARS更能有效地选取出有效变量,建立的模型对新品种苹果的预测性能有明显改善,RMSEP为0.587%,Rp为0.928,预测偏差减少到-0.052%,RPD=2.684。

    Abstract:

    China is a big country of apple producers and exporters,so it is essential to control the quality of exported apples. The soluble solids content (SSC) in apples is an essential factor affecting the quality of apples. Non-destructive inspection of SSC in apples can be realized by near-infrared spectroscopy (NIR). The model update method combined with the variable screening method was used to perform nondestructive inspection on the SSC in five apple varieties including Red Fuji,Green apple,Golden Delicious,Rose and Lokit to obtain a general robust multi-variety apple model. The results showed that the updated new model achieved high-precision prediction of the SSC in five varieties of apples. At this time,the root mean square error of prediction (RMSEP) of the model was 0.698%,the correlation coefficient of prediction (Rp) was 0.904,the prediction deviation was 0.074%,and the ratio of performance to standard deviate (RPD) reached 2.340. The model was optimized with three waveband selection methods including competitive adaptive reweighted sampling (CARS),successive projections algorithm (SPA) and uninformative variable elimination (UVE) to identify and extract important information bands of the spectrum. The results showed that CARS selected effective wavebands more effectively. The established model significantly improved the analytical performance in new apple varieties,with RMSEP of 0.587%,Rp of 0.928,and prediction bias reduced to -0.052%,with RPD of 2.684. Therefore,the strategy of combining model updating with waveband selection has great application potential in studying and developing the general model of portable near-infrared detector for multi-variety fruits.

    参考文献
    相似文献
    引证文献
引用本文

刘燕德,黎丽莎,李斌,宋烨,朱向荣,姜延泉.多品种苹果可溶性固形物近红外无损检测通用模型研究[J].华中农业大学学报,2022,41(2):237-244

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2022-04-02
  • 出版日期: