Related Experiment Video
Updated: Nov 25, 2025
08:20
Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images
Published on: October 27, 2023
2.1K
Real-time plant health assessment via implementing cloud-based scalable transfer learning on AWS DeepLens
Asim Khan1, Umair Nawaz2, Anwaar Ulhaq1,3
1The Institute for Sustainable Industries and Liveable Cities (ISILC), College of Engineering and Science, Victoria University, Melbourne, Australia.
Plos One
|December 17, 2020
Summary
This study introduces a DeepLens Classification and Detection Model (DCDM) for automated plant leaf disease identification. The model achieves 98.78% accuracy in real-time diagnosis, improving agricultural efficiency and economic outcomes.
Area of Science:
- Agricultural Science
- Computer Science
- Machine Learning
Background:
- Plant leaf diseases significantly impact crop yield, quality, and national economies.
- Existing automated disease detection methods often suffer from hardware complexity, scalability issues, and practical inefficiencies.
Purpose of the Study:
- To develop a user-friendly, scalable, and efficient automated system for detecting and classifying plant leaf diseases.
- To address the limitations of current Machine Learning (ML) models in real-world agricultural applications.
Main Methods:
- Implemented a DeepLens Classification and Detection Model (DCDM) using scalable transfer learning on Amazon Web Services (AWS) SageMaker.
- Integrated the model with AWS DeepLens for real-time, on-device disease diagnosis.
- Trained the deep learning model on 40,000 images and evaluated it on 10,000 images.
Main Results:
- Achieved a high accuracy of 98.78% in identifying diseases across various fruit and vegetable plants.
- Demonstrated real-time disease diagnosis capability, with an average image testing time of 0.349 seconds using AWS DeepLens.
- Provided disease information to users in under one second.
Conclusions:
- The DCDM offers a scalable and accessible solution for automated plant leaf disease control.
- The model's high accuracy and real-time performance enhance usability and reduce economic losses in agriculture.
- Cloud integration ensures ubiquitous access and efficient disease management for farmers.
