EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Computer Vision-Based Disease Recognition in Tomato Leaves Using SURF–CNN Feature Fusion and SVM Classification

V. A. Hegde, A. R. Kasetwar

AFRICAN JOURNAL OF APPLIED RESEARCH · 04.07.2026 · 12 (5) · 1-25

DOI: 10.26437/wsctv062 ↗

Kurz gefasst

Fachartikel in der Zeitschrift AFRICAN JOURNAL OF APPLIED RESEARCH, online erschienen am 04.07.2026. Verfasst von V. A. Hegde und 1 weiteren Autor:innen. Die Arbeit nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Purpose

To design and implement an efficient and accurate Artificial Intelligence (AI)- based Deep Learning (DL) model for early identification of tomato leaf diseases, with an aim of increasing crop growth and maintaining the good health of plants. Design/Methodology/Approach: This research work proposes a Convolutional Neural Network (CNN)-based model that integrates basic image processing with DL methods. Initially, preprocessing is performed using a bilateral filter to preserve edges, followed by segmentation using the OTSU thresholding method. Speeded-Up Robust Features (SURF) is used to extract relevant features from the image, which are then fused with the CNN features. The final classification is conducted using a Support Vector Machine (SVM) classifier. Research Limitation: The work mainly centred on tomato leaves and relies on a controlled dataset, which may limit the generalisation of images across varying light conditions, leaf occlusions, and image capture methods. So, validation is required with a large number of real-time datasets.

Findings

The experimental evaluation demonstrated that the proposed hybrid CNN model achieved the highest accuracy of 98.15%, surpassing other conventional DL models. For real-time applications, the model has also demonstrated low computational complexity and good speed. These findings have highlighted the accuracy of the hybrid model for tomato leaf disease detection with feature fusion methodology. Practical Implication: This efficient, less complex hybrid model enables real-time crop monitoring and the design of low-cost hardware. Also assist farmers and agricultural experts with diagnosis and interventions to achieve good crop yields. Social Implication: Early and accurate detection of tomato leaf diseases contributes to the agricultural sector by increasing food and crop security and reducing pesticide use, benefiting the environment and farmers. Originality / Value: The study has introduced a hybrid framework that integrates CNN-based DL, the SURF feature fusion method, and an SVM classifier.

Zitieren

APA

Hegde, V. A., & Kasetwar, A. R. (2026). Computer Vision-Based Disease Recognition in Tomato Leaves Using SURF–CNN Feature Fusion and SVM Classification. AFRICAN JOURNAL OF APPLIED RESEARCH, 12(5), 1-25. https://doi.org/10.26437/wsctv062

Vancouver

Hegde VA, Kasetwar AR. Computer Vision-Based Disease Recognition in Tomato Leaves Using SURF–CNN Feature Fusion and SVM Classification. AFRICAN JOURNAL OF APPLIED RESEARCH. 2026;12(5):1-25. doi:10.26437/wsctv062

BibTeX
@article{Hegde2026,
  author = {Hegde, V. A. and Kasetwar, A. R.},
  title = {Computer Vision-Based Disease Recognition in Tomato Leaves Using SURF–CNN Feature Fusion and SVM Classification},
  journal = {AFRICAN JOURNAL OF APPLIED RESEARCH},
  year = {2026},
  volume = {12},
  number = {5},
  pages = {1--25},
  doi = {10.26437/wsctv062}
}
Kostenlos starten →