EUVIMEDEUVIMED entdecken
Crossref

HARNN ‐ IoT : IoT ‐Enabled Hybrid Deep Learning Model for Cotton Plant Disease Detection and Classification

Bhushan V. Patil, Pravin Sahebrao Patil, Punam R. Patil

Journal of Phytopathology · 2025-11 · 173 (6)

DOI: 10.1111/jph.70213 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Phytopathology, online erschienen am 29.12.2025. Verfasst von Bhushan V. Patil und 2 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 41 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

One of the most widely grown cultivated natural fibre crops is cotton in the world, but it is extremely susceptible to diverse diseases and pests, which significantly minimise harvests and quality. Existing cotton disease detection system offers challenges due to noisy images and limited feature generalisation, which leads to minimise detection accuracy and real‐time feasibility. To address these problems, this paper proposes an Internet of Things (IoT) enabled deep learning (DL) model for automatic cotton plant disease detection and classification. A Raspberry Pi system was used to collect the inputs in the proposed study. The proposed method uses a hybrid attention‐enabled residual neural network (HARNN) model for cotton plant disease classification. Initially, the input images are pre‐processed with adaptive weighted mean filtering (AWMF) to remove unwanted noise. Handcrafted features are extracted using a histogram of oriented gradients (HOG), a uniform local binary pattern (uLBP) and a grey level co‐occurrence matrix (GLCM); statistical features are extracted using mean, standard deviation, skewness, kurtosis, minimum and maximum value and deep features are extracted using the Inception V3 model. The extracted features are fused using the Fusenet model. The concatenated feature is fed into the proposed HARNN classifier to classify varied types of cotton plant disease. HARNN model's parameters are fine‐tuned with the chaotic Lyre Bird (CLBOA) optimization method. The simulation results show that the proposed research obtained superior results in terms of accuracy of 99.7%, precision of 99.3%, recall of 99.5%, f ‐score of 99.8% and specificity of 99.9%, respectively.

Zitieren

APA

Patil, B. V., Patil, P. S., & Patil, P. R. (2025). HARNN ‐ IoT : IoT ‐Enabled Hybrid Deep Learning Model for Cotton Plant Disease Detection and Classification. Journal of Phytopathology, 173(6). https://doi.org/10.1111/jph.70213

Vancouver

Patil BV, Patil PS, Patil PR. HARNN ‐ IoT : IoT ‐Enabled Hybrid Deep Learning Model for Cotton Plant Disease Detection and Classification. Journal of Phytopathology. 2025;173(6). doi:10.1111/jph.70213

BibTeX
@article{Patil2025,
  author = {Patil, Bhushan V. and Patil, Pravin Sahebrao and Patil, Punam R.},
  title = {HARNN ‐ IoT : IoT ‐Enabled Hybrid Deep Learning Model for Cotton Plant Disease Detection and Classification},
  journal = {Journal of Phytopathology},
  year = {2025},
  volume = {173},
  number = {6},
  doi = {10.1111/jph.70213}
}
Kostenlos starten →