Page 191 - AI for Good-Innovate for Impact Final Report 2024
P. 191
AI for Good-Innovate for Impact
Use case – 44: Intelligent UAV-Assisted Plant Disease Detection in
Rock Melon Greenhouses 44 - UTM
Country: Malaysia
Organization: Universiti Teknologi Malaysia
Contact person: Norulhusna Ahmad, norulhusna.kl@ utm .my, +60192777681
44�1� Use case summary table
Domain Smart Agriculture
The problem to be addressed • Decrease crop losses
• Inconsistencies in plant disease detection
• Decrease human errors
Key aspects of the solution Plant disease detection
Technology keywords Image, agriculture, plant detection.
Data availability Currently Private.
Metadata (type of data) • NPK sensors
• Soil moisture sensors
• Drone camera
Model Training and fine-tuning • Reinforcement learning
• Categories of plant diseases
• One class detection and five class classification
• YOLOv9, IoT, Edge Computing
Testbeds or pilot deployments http:// dx .doi .org/ 10 .14569/ IJACSA .2024 .01501119
44�2� Use-case description
44�2�1� Description
The Intelligent UAV-Assisted Plant Disease Detection project in Rock Melon Greenhouses,
supported by the ASEAN IVO grant on Edge Computing in Agriculture, represents a
groundbreaking advancement in agricultural surveillance. By integrating unmanned aerial
vehicle (UAV) technology with cutting-edge deep-learning models like You Only Look Once
version 9 (YOLOv9), this initiative aims to revolutionize precision agriculture practices. The
project's innovative approach involves utilizing UAV imagery to detect diseases in melon leaves,
focusing on enhancing agricultural productivity and sustainability. Moreover, an automated
and online plant disease detection system based on the YOLOv9 model eliminates the need
for farmers to physically inspect the greenhouse, allowing them to allocate their time more
efficiently.
175
