Open AccessIdentification system for beeswax on Malang apple using VNIR imaging
Author(s) -
Naufal Praditya,
Adhi Harmoko Saputro
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1528/1/012037
Subject(s) - vnir , beeswax , confusion matrix , artificial intelligence , wax , support vector machine , identification (biology) , computer science , pattern recognition (psychology) , remote sensing , materials science , hyperspectral imaging , botany , geology , biology , composite material
Wax coating identification on fruits is very difficult without a non-destructive method. In general, destructive methods were used to identify wax or coatings by soaking the fruit in hot water or using a mixture of vinegar and water. There are also destructive systems that was used such as gas chromatography linked with mass spectrometry, but this method takes much time and difficult to operate. Visible Near Infrared Imaging (VNIR) becomes the alternate solution to identify wax on the surface of the fruit without spoiling the quality of the fruit. In this study, identification system for beeswax application on apples has been made successfully. The process starts through image acquisition, image correction, object detection, window averaging, classification model, and the coating status. The VNIR image was acquired on a wavelength range from 400 to 1000 nm. The data was divided for training and testing the classification model using cross-validation method, then the model was evaluated using confusion matrix. Several classification models were used to compare the result and to conclude which model gives the best accuracy for identification and classification problems. The accuracy of the three models were 72.92% for PCA-SVM model, 81.25% for DT model, and 91.67% for RF model.