基于光谱和水分补偿方法的鲜枣内部品质检测
发布时间：2018-09-10 08:18
【摘要】：为了建立稳定可靠的鲜枣品质检测模型,利用光谱和水分补偿方法进行鲜枣内部品质的检测。首先,针对鲜枣各品质指标(水分含量、可溶性固形物含量、维生素C含量、蛋白质含量、硬度值),采用回归系数法(RC)提取特征波段并建立最小二乘支持向量机(LS-SVM)检测模型,预测集的决定系数(R2P)均在0.8261以上,预测均方根误差(RMSEP)均在3.324 9以下。在提取各项品质指标特征波段的基础上,剔除其他四项单一品质特征波段中与水分特征波段(包含利用RC法所提取到的水分特征波长和鲜枣中具有明显水分特征的吸收峰)重叠或接近的波段,并与鲜枣水分含量值进行数据融合建立了各项指标的水分补偿模型。结果表明,硬度值的水分补偿模型精度有一定提高,R2P和RMSEP分别为0.830 5和0.055 3;可溶性固形物含量、维生素C含量、蛋白质含量的水分补偿模型精度均有所下降,R2P分别为0.804 1,0.878 2和0.837 8,RMSEP分别为1.347 3,0.638 0和3.503 2。然后,分析各品质指标间的相关性,结果表明,水分含量在0.05水平上与硬度值呈现显著的相关性,在0.01的水平上与其余三项品质指标之间存在极显著的相关性,相关性强弱与水分补偿模型的建模结果相互支持。研究表明,水分补偿法所建的预测模型可用于鲜枣内部品质的检测,水分含量与其他四项品质指标之间有相互作用并影响其他品质指标所建立的预测模型。该研究为进一步探讨光谱检测中各内部品质指标间交互作用的解耦提供了新思路。
[Abstract]:In order to establish a stable and reliable quality detection model of fresh jujube, the internal quality of fresh jujube was detected by spectrum and moisture compensation method. Firstly, aiming at the quality indexes (water content, soluble solids content, vitamin C content, protein content, hardness value) of fresh jujube, the characteristic bands were extracted by regression coefficient method (RC) and the detection model of least square support vector machine (LS-SVM) was established. The coefficient of determination (R2P) of prediction set is above 0.8261, and the root mean square error (RMSEP) of prediction is below 3.324 9. On the basis of extracting characteristic bands of each quality index, The other four single quality bands overlap or approach the water characteristic bands (including the water characteristic wavelengths extracted by RC method and the absorption peaks with obvious water characteristics in fresh jujube). The moisture compensation model of each index was established by data fusion with water content of fresh jujube. The results showed that the precision of water compensation model of hardness value was improved to some extent, the values of R2P and RMSEP were 0.830 5 and 0.055 3, respectively, the content of soluble solids and vitamin C were increased, The precision of water compensation model for protein content decreased to a certain extent. The R2P values of R2P were 0.804 ~ 0.878 2 and 0.837 ~ 8 ~ (-1) RMSEP of 1.347 ~ 3 ~ 0 ~ 0.638 0 and 3.503 ~ 2, respectively. Then, the correlation between each quality index was analyzed. The results showed that there was a significant correlation between water content at 0.05 level and hardness value, and a very significant correlation between water content and the other three quality indexes at 0.01 level. The modeling results of correlation and moisture compensation model support each other. The results show that the prediction model established by water compensation method can be used to detect the internal quality of fresh jujube, and there is interaction between water content and the other four quality indexes. This study provides a new idea for further exploring the decoupling of internal quality indexes in spectral detection.
【作者单位】： 山西农业大学工学院;
【基金】：国家自然科学基金项目(31271973)资助
【分类号】：O657.3;TS255.7
，
本文编号：2233887
[Abstract]:In order to establish a stable and reliable quality detection model of fresh jujube, the internal quality of fresh jujube was detected by spectrum and moisture compensation method. Firstly, aiming at the quality indexes (water content, soluble solids content, vitamin C content, protein content, hardness value) of fresh jujube, the characteristic bands were extracted by regression coefficient method (RC) and the detection model of least square support vector machine (LS-SVM) was established. The coefficient of determination (R2P) of prediction set is above 0.8261, and the root mean square error (RMSEP) of prediction is below 3.324 9. On the basis of extracting characteristic bands of each quality index, The other four single quality bands overlap or approach the water characteristic bands (including the water characteristic wavelengths extracted by RC method and the absorption peaks with obvious water characteristics in fresh jujube). The moisture compensation model of each index was established by data fusion with water content of fresh jujube. The results showed that the precision of water compensation model of hardness value was improved to some extent, the values of R2P and RMSEP were 0.830 5 and 0.055 3, respectively, the content of soluble solids and vitamin C were increased, The precision of water compensation model for protein content decreased to a certain extent. The R2P values of R2P were 0.804 ~ 0.878 2 and 0.837 ~ 8 ~ (-1) RMSEP of 1.347 ~ 3 ~ 0 ~ 0.638 0 and 3.503 ~ 2, respectively. Then, the correlation between each quality index was analyzed. The results showed that there was a significant correlation between water content at 0.05 level and hardness value, and a very significant correlation between water content and the other three quality indexes at 0.01 level. The modeling results of correlation and moisture compensation model support each other. The results show that the prediction model established by water compensation method can be used to detect the internal quality of fresh jujube, and there is interaction between water content and the other four quality indexes. This study provides a new idea for further exploring the decoupling of internal quality indexes in spectral detection.
【作者单位】： 山西农业大学工学院;
【基金】：国家自然科学基金项目(31271973)资助
【分类号】：O657.3;TS255.7
，
本文编号：2233887
本文链接：https://www.wllwen.com/kejilunwen/huaxue/2233887.html
教材专著
