Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Determination of shape features for apricot based on machine vision

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  2018 ASABE Annual International Meeting  1800739.(doi:10.13031/aim.201800739)
Authors:   Xi Yang, Ruoyu Zhang, Kunlin Zou
Keywords:   apricot; fruit shape; machine vision; machine learning

Abstract. Apricot is the third most important stone fruit in the world. Its shape is one significant element to describe characteristics. However, there are no common methods or standards to describe fruit shape in the world. According to the rules of species and the standards of market detection, there is still need a more precise uniform standard to describe the fruit shape qualitatively and quantitatively. In this study, the shape features of four apricot species were investigated. To simplify geometric attributes, line items were used to describe the fruit shape. Based on the properties of shape, apricot species relationship was studied by cluster analysis. The identification of apricot species was tried by three machine learning methods consisting of partial least squares discriminant analysis (PLS_DA) , support vector machine (SVM), and back-propagation neural network (BP_NN). Accuracy of PLS_DA, SVM and BP_NN were 95.5, 99 and 96.5 %. Thus, quantitative research could be done in this way.

(Download PDF)    (Export to EndNotes)