Detection of mango soluble solid content using hyperspectral imaging technology

被引：17

作者：

Tian, Pan ^{[1
]}

Meng, Qinghua ^{[1
,2
]}

Wu, Zhefeng ^{[1
,2
]}

Lin, Jiaojiao ^{[1
]}

Huang, Xin ^{[1
]}

Zhu, Hui ^{[1
]}

Zhou, Xulin ^{[1
]}

Qiu, Zouquan ^{[1
]}

Huang, Yuqing ^{[3
,4
]}

Li, Yu ^{[5
]}

机构：

[1] Nanning Normal Univ, Sch Phys & Elect, Nanning 530001, Peoples R China

[2] Nanning Normal Univ, Key Lab New Elect Funct Mat Guangxi Coll & Univ, Nanning 530001, Peoples R China

[3] Nanning Normal Univ, Key Lab Environm Evolut & Resource Utilizat Beibu, Minist Educ, Nanning 530001, Peoples R China

[4] Nanning Normal Univ, Guangxi Key Lab Earth Surface Proc & Intelligent S, Nanning 530001, Peoples R China

[5] Guangxi Tech Instruct Off Fruit, Nanning 530022, Peoples R China

来源：

INFRARED PHYSICS & TECHNOLOGY | 2023年 / 129卷

基金：

中国国家自然科学基金;

关键词：

Hyperspectral imaging; Mango; Nondestructive detection; SSC; Partial least squares regression; PREDICTION; SELECTION; FIRMNESS; INJURY;

D O I：

10.1016/j.infrared.2023.104576

中图分类号：

TH7 [仪器、仪表];

学科分类号：

0804 ; 080401 ; 081102 ;

摘要：

Soluble solid content (SSC) is an important indicator for evaluating mango quality. The main task of this study is to develop a partial least squares (PLS) regression model for SSC by combinating the visible and near infrared (400-1000 nm) hyperspectral imaging. The PLSR model can be used to assess the quality grading of mangoes. By comparing the performance of five preprocessing full-band models, the standard normal variable transformation (SNV) and multiplicative scatter correction algorithm (MSC) are selected for this study. Otherwise, three variable selection methods, including successive projection algorithm (SPA), competitive adaptive reweighted sampling (CARS) method, and the genetic algorithm (GA) are used for the identification of the characteristic wavelengths. The screened feature bands are used to build PLS regression models.The SNV-CARS-PLS model is found to show the best prediction performance. The correlation coefficient for the predicted value for the mango SSC and its root mean square error are determined to be 0.9001 and 0.6162, respectively. These results suggest that the SNVCARS-PLS model is an effective method for predicting mango SSC.

引用

页数：9