UAV-based aerial imaging and path optimization to combat mosquito-borne diseases.

Area of Science:

• Environmental Science
• Public Health
• Computer Vision

Background:

• The Aedes aegypti mosquito transmits diseases like Dengue, Chikungunya, Yellow Fever, and Zika.
• No specific treatment exists for Dengue; prevention focuses on eliminating mosquito breeding sites.
• Identifying Possible Breeding Habitats (PBH) is crucial for mosquito control.

Purpose of the Study:

• To develop an automated system for detecting mosquito breeding habitats using Unmanned Aerial Vehicles (UAVs).
• To optimize UAV flight paths for efficient larvicide application.
• To improve disease prevention by targeting mosquito reproduction.

Main Methods:

• An aerial dataset of PBH was created using UAVs and manually annotated.
• Object detection was performed using YOLOv8 and YOLOv11 algorithms.
• The Travelling Salesman Problem was employed for optimizing UAV path planning for larvicide spraying.

Main Results:

• YOLOv11 significantly outperformed YOLOv8 in detecting PBH, achieving high metrics (mAP50: 0.97, mAP50-90: 0.61, Precision: 0.96, Recall: 0.88).
• Optimized path planning reduced energy and battery consumption during UAV operations.
• The integrated system demonstrated effective detection and treatment of mosquito habitats.

Conclusions:

• UAVs combined with YOLOv11 offer a cost-effective and efficient solution for identifying and treating mosquito breeding sites.
• Optimized path planning enhances the operational efficiency of UAVs in mosquito control.
• This approach represents a significant advancement in preventing mosquito-borne diseases.