Nondestructive discrimination of advanced clones and cultivars of strawberry using an innovative approach involving image analysis and machine learning
Acta Scientiarum Polonorum Hortorum Cultus · 30.04.2025 · 24 (2) · 3-14
DOI: 10.24326/asphc.2025.5370 ↗Kurz gefasst
Fachartikel in der Zeitschrift Acta Scientiarum Polonorum Hortorum Cultus, online erschienen am 30.04.2025. Verfasst von Ewa Ropelewska und 1 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Different clones and cultivars of strawberry can differ in morphological and chemical properties, as well as productivity, adaptation to cultivation conditions, and post-harvest quality during storage and processing. Due to differences in the quality of raw materials and final products depending on the strawberry clone/cultivar, correct distinguishing clones and cultivars is important for growers, consumers and processors. This study was aimed at distinguishing advanced clones and cultivars of strawberry using an innovative approach involving image processing and artificial intelligence. The raw material included the advanced clones and cultivars of strawberry, such as clone with the breeding code T-201457-16 (Grandarosa × Elsanta), clone T-201536-06 (Clery × Grandarosa), clone T-201567-01 (Patty × Panvik), as well as the cultivars Fibion, Grandarosa, and Markat. The fruit image acquisition was performed using a digital camera. As many as 2172 image parameters were extracted from the image of each fruit converted to different color channels R, G, B, L, a, b, X, Y, Z, U, V, and S and textures with the highest discriminative power were selected to develop models using various machine learning algorithms, such as Multilayer Perceptron, MultiClass Classifier, IBk, and LMT, Linear Discriminant, Quadratic SVM, Subspace Discriminant, and Wide Neural Network. The most accurate classifications were obtained for a model built using Subspace Discriminant (96.30%) and Multilayer Perceptron (95.83%). For the model developed using Subspace Discriminant, clone T-201567-01 and cultivar Markat were completely correctly classified with the highest accuracy of 100%. Whereas in the case of the model built using Multilayer Perceptron clone T-201567-01 was characterized by the highest classification metrics, such as Precision and F-measure equal to 0.983, MCC of 0.980, PRC Area and ROC Area of 1.000. The developed approach can be used in practice to discriminate advanced clones and cultivars of strawberry in an objective and nondestructive manner.
Zitieren
APA
Ropelewska, E., & Masny, A. (2025). Nondestructive discrimination of advanced clones and cultivars of strawberry using an innovative approach involving image analysis and machine learning. Acta Scientiarum Polonorum Hortorum Cultus, 24(2), 3-14. https://doi.org/10.24326/asphc.2025.5370
Vancouver
Ropelewska E, Masny A. Nondestructive discrimination of advanced clones and cultivars of strawberry using an innovative approach involving image analysis and machine learning. Acta Scientiarum Polonorum Hortorum Cultus. 2025;24(2):3-14. doi:10.24326/asphc.2025.5370
BibTeX
@article{Ropelewska2025,
author = {Ropelewska, Ewa and Masny, Agnieszka},
title = {Nondestructive discrimination of advanced clones and cultivars of strawberry using an innovative approach involving image analysis and machine learning},
journal = {Acta Scientiarum Polonorum Hortorum Cultus},
year = {2025},
volume = {24},
number = {2},
pages = {3--14},
doi = {10.24326/asphc.2025.5370}
}
EUVIMED