Abstract—The export of agricultural products is expanding in developed agricultural countries like Vietnam. The dragon fruit is a fruit that accounts for a large percentage of exports in order to meet the import standards of nations throughout the world, dragon fruit is sorting according to each importing country’s standard. The paper describes a dragon fruit classification system automatically that solves the pricing and accuracy difficulties compared with the manual one. Using a combination of models such as KNN to identify dragon fruit, CNN to extract dragon fruit features, and ANN to calculate dragon fruit classification based on the data provided, proposed a model self-training model to improve the accuracy of the system due to insufficient or not various data for training. The dragon fruit database was collected from 1287 dragon fruit supplied from dragon fruit farms, or from export dragon fruit packaging and sorting facilities. Dragon fruit is divided into 3 groups of G1, G2, G3 with different standard in length, width, weight, defects outside the dragon fruit skin. The automatic dragon fruit classification system proposed after the test achieved 98.5% accuracy 6 times more yield than the manual classification.
Index Terms—classifying, sorting, grading, CNN, ANN, KNN, machine learning, AI
Cite: Nguyen Minh Trieu and Nguyen Truong Thinh, "A Study of Combining KNN and ANN for Classifying Dragon Fruits Automatically," Journal of Image and Graphics, Vol. 10, No. 1, pp. 28-35, March 2022.
Copyright © 2022 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.
Copyright © 2012-2023 Journal of Image and Graphics, All Rights Reserved