| 张湃,王丽侠,任丽棉.基于VMD-SVD-MSR模型的樱桃品质信息检测[J].唐山学院学报,2022,35(3):18-24,99 |
| 基于VMD-SVD-MSR模型的樱桃品质信息检测 |
| Cherry Quality Information Detection Based On VMD-SVD-MSR Model |
| |
| DOI：10.16160/j.cnki.tsxyxb.2022.03.003 |
| 中文关键词: 樱桃 VMD-SVD-MSR模型 信息检测 |
| 英文关键词: cherry VMD-SVD-MSR model information detection |
| 基金项目: |
|
| 摘要点击次数: 6118 |
| 全文下载次数: 0 |
| 中文摘要: |
| 樱桃的可溶性固形物含量(SSC)和含水率是衡量其品质的重要指标。鉴于传统的检测方法均为破坏性生化检测,文章以山海关樱桃样品为研究对象,提出一种基于近红外光谱信息融合的樱桃SSC和含水率无损检测方法。该方法首先利用变分模式分解(VMD)对近红外反射光谱进行多模态分解,得到各个固有模态(IMF)并分别求解各IMF层与SSC及含水率之间的相关系数,相关系数值越大说明对应的IMF层越适合特征提取;然后再进一步采用SiPLS波长筛选方法提取IMF层光谱的最佳波段,利用奇异值分解(SVD)求得奇异熵,建立多元逐步回归(MSR)预测模型(简称VMD-SVD-MSR模型)。为了验证该模型的有效性,引入连续投影算法和竞争性自适应重加权等变量优选方法进行特征波段筛选,输入多元逐步回归预测模型进行比较,结果表明,VMD-SVD-MSR模型通过一次光谱提取,能够同时实现樱桃的SSC和含水率的无损检测且预测能力较强。 |
| 英文摘要: |
| The cherry's soluble solids content (SSC) and moisture content are important indexes to measure its quality. However, the traditional detection methods are destructive biochemical detection. With the Shanhaiguan cherry sample as the research object, this paper proposes a nondestructive detection method for SSC and moisture content based on near infrared spectrum information fusion. Firstly, the near infrared reflection spectrum is decomposed with multi-modal by variational mode decomposition (VMD)to obtain each inherent mode function (IMF), and the correlation coefficients between each IMF layer,SSC and moisture content are calculated. The larger the coefficient is, the more suitable the corresponding IMF layer is for feature extraction; Then, the best band of IMF layer spectrum is extracted by SiPLS wavelength screening method, and the singular value decomposition (SVD) is used to obtain the singular entropy, and the multiple stepwise regression (MSR) prediction model(VMD-SVD-MSR model) is established. In order to verify the effectiveness of this model, continuous projection algorithm and competitive adaptive reweighting and other variable optimization methods are introduced to screen the characteristic bands, and multiple stepwise regression prediction models are input for comparison. The results show that the VMD-SVD-MSR model can realize the nondestructive detection of SSC and moisture content of the cherry at the same time through one-time spectral extraction, and the prediction ability of this model is strong. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|