EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Advancing real-time plant disease detection: A lightweight deep learning approach and novel dataset for pigeon pea crop

Sandesh Bhagat, Manesh Kokare, Vineet Haswani, Praful Hambarde, Trupti Taori, P.H. Ghante, D.K. Patil

Smart Agricultural Technology · 2024-03 · 7 · 100408

DOI: 10.1016/j.atech.2024.100408 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Smart Agricultural Technology, 2024-03. Verfasst von Sandesh Bhagat und 6 weiteren Autor:innen. Die Arbeit wurde bisher 64-mal zitiert (Crossref) und nennt 56 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministry of Education, India, All India Council for Technical Education, Nevada Division of Forestry.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Plant disease detection and early disease treatment are essential for sustainable crop production. Computer vision for crop science is overgrowing with the advancement in deep learning. However, plant disease detection is challenging due to random disease spread throughout the plant body, environmental challenges, and lack of actual field datasets. The proposed work systematically addresses these issues with(a) The construction of the first pigeon pea image dataset from the agriculture field in collaboration with 20 Agricultural Research Centers (ARS) and government agencies across 18 states in India. (b) The design of lightweight and high-performance models for realtime plant disease detection in resource-constrained devices. The proposed lightweight Lite-MDC architecture uses the Multi-kernel Depthwise separable Convolutions (MDConv), depthwise convolution with 3 × 3 and 5 × 5 kernels help in extracting multiscale features of plant diseases. This structure reduces the computational complexity and uses 2.2 M parameters, which is 62% less than standard VGG16 architecture. The proposed method outperforms the state-of-the-art techniques such as InceptionV3, VGG16, ResNet50, DenseNet, MobileNet, MobileNetV3, NASNet, and EfficieNetB0 on the proposed pigeon pea dataset and obtained 94.14% accuracy with 34 Frame Per Second (FPS). The performance of the proposed method is validated using publicly available datasets; the plant village dataset, Cassava, and apple leaf datasets, achieved 99.78%, 86.4%, and 97.2% accuracy, respectively. The Lite-MDC model can be used for plant disease detection in real-time with resource-constrained devices and webbased cloud platforms.

Förderung

  • Ministry of Education, India
  • All India Council for Technical Education
  • Nevada Division of Forestry

Zitieren

APA

Bhagat, S., Kokare, M., Haswani, V., Hambarde, P., Taori, T., Ghante, P., & Patil, D. (2024). Advancing real-time plant disease detection: A lightweight deep learning approach and novel dataset for pigeon pea crop. Smart Agricultural Technology, 7, 100408. https://doi.org/10.1016/j.atech.2024.100408

Vancouver

Bhagat S, Kokare M, Haswani V, Hambarde P, Taori T, Ghante P, et al. Advancing real-time plant disease detection: A lightweight deep learning approach and novel dataset for pigeon pea crop. Smart Agricultural Technology. 2024;7:100408. doi:10.1016/j.atech.2024.100408

BibTeX
@article{Bhagat2024,
  author = {Bhagat, Sandesh and Kokare, Manesh and Haswani, Vineet and Hambarde, Praful and Taori, Trupti and Ghante, P.H. and Patil, D.K.},
  title = {Advancing real-time plant disease detection: A lightweight deep learning approach and novel dataset for pigeon pea crop},
  journal = {Smart Agricultural Technology},
  year = {2024},
  volume = {7},
  pages = {100408},
  doi = {10.1016/j.atech.2024.100408}
}
Kostenlos starten →