Технічні науки English Відкритий доступ
Neural network-based detection of leaf defects in trees using TensorFlow
Анотація
This study addresses the development of an automated methodology for the identification of tree foliage and associated morphological defects utilizing digital imagery. The primary objective is to augment the reliability of defect detection through advanced image analysis techniques. To this end, a neural network architecture was engineered using Python and the TensorFlow framework to classify defects and segment compromised tissue. This system facilitates the early recognition of pathologies, thereby enabling prophylactic interventions to maintain arboreal health. Beyond its diagnostic utility, the proposed framework streamlines biological and botanical research by automating data acquisition and analysis, thus assisting in the investigation of environmental factors influencing tree development. In the agricultural sector, the system offers significant value by enabling the rapid, precise diagnosis of diseases and pest infestations in deciduous species. Furthermore, it aids in the identification of macronutrient deficiencies (nitrogen, phosphorus, potassium), allowing agronomists to implement precision agriculture strategies that optimize plant nutrition and physiological vigor. The MobileNet-based solution demonstrated a 92.8% accuracy on the validation dataset and 88.8% in real-world field tests, outperforming baseline models such as ResNet-50 and SVM.
Ключові слова: artificial neural networks; precision agriculture; leaf disease identification; automated diagnosis; machine learning; computer vision; plant pathology
Як цитувати
Prykhodko T. Neural network-based detection of leaf defects in trees using TensorFlow. International scientific journal "Internauka". 2026. № 8. https://doi.org/10.25313/2520-2057-2026-8-12115
@article{internauka12115,
author = {Prykhodko, Taisiіa},
title = {Neural network-based detection of leaf defects in trees using TensorFlow},
journal = {Міжнародний науковий журнал «Інтернаука»},
year = {2026},
number = {8},
doi = {10.25313/2520-2057-2026-8-12115},
url = {https://www.inter-nauka.com/issues/2026/8/12115}
}
TY - JOUR AU - Prykhodko, Taisiіa TI - Neural network-based detection of leaf defects in trees using TensorFlow JO - Міжнародний науковий журнал «Інтернаука» PY - 2026 IS - 8 SN - 2520-2057 DO - 10.25313/2520-2057-2026-8-12115 UR - https://www.inter-nauka.com/issues/2026/8/12115 LA - en ER -