Application of multispectral imaging to determine quality attributes and ripeness stage in strawberry fruit

PLoS One. 2014 Feb 4;9(2):e87818. doi: 10.1371/journal.pone.0087818. eCollection 2014.

Abstract

Multispectral imaging with 19 wavelengths in the range of 405-970 nm has been evaluated for nondestructive determination of firmness, total soluble solids (TSS) content and ripeness stage in strawberry fruit. Several analysis approaches, including partial least squares (PLS), support vector machine (SVM) and back propagation neural network (BPNN), were applied to develop theoretical models for predicting the firmness and TSS of intact strawberry fruit. Compared with PLS and SVM, BPNN considerably improved the performance of multispectral imaging for predicting firmness and total soluble solids content with the correlation coefficient (r) of 0.94 and 0.83, SEP of 0.375 and 0.573, and bias of 0.035 and 0.056, respectively. Subsequently, the ability of multispectral imaging technology to classify fruit based on ripeness stage was tested using SVM and principal component analysis-back propagation neural network (PCA-BPNN) models. The higher classification accuracy of 100% was achieved using SVM model. Moreover, the results of all these models demonstrated that the VIS parts of the spectra were the main contributor to the determination of firmness, TSS content estimation and classification of ripeness stage in strawberry fruit. These results suggest that multispectral imaging, together with suitable analysis model, is a promising technology for rapid estimation of quality attributes and classification of ripeness stage in strawberry fruit.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Food Inspection* / methods
  • Food Quality*
  • Fragaria*
  • Fruit* / chemistry
  • Fruit* / standards
  • Humans

Grants and funding

This 269 study is supported by the Specialized Research Fund for the Doctoral Program of 270 Higher Education (20120111110024), the Fundamental Research Funds for the Central Universities (2012HGCX0003), (2012HGZY0021), the National Key Technologies R&D Programme (2012BAD07B01), and the Funds for Huangshan Professorship of Hefei University of Technology. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.