Autonomous Deforestation Monitoring in Pakistan: A Multi-Drone, AI-Based System for Real-Time Deforestation Protection

Illegal deforestation is a growing concern for ecologically fragile areas like the Kaghan Valley in Pakistan [1]. Despite being an economic engine and a significant geographical link, the area has no real-time monitoring system to log illegal deforestation. Conventional methods like satellite imaging and physical ranger patrolling are faced with limitations, only able to conclusively detect deforestation after the damage is done. These predicaments motivate the development of a multi-drone robotic surveillance system that combines aerial imaging with AI-based analysis to detect and suppress illegal logging activity in the region.

The system is prototyped using a simulation-based approach nested in Ubuntu, ROS, and Gazebo [2]. The system includes a modulation of a forest environment with a fixed-wing drone system (equipped with a thermal imaging camera) to effectively monitor canopy loss, unauthorised trails, and unregulated timber movement in real time. Designed to function in low-visibility conditions within simulation, the system is capable of running simulated night-time monitoring operations which is usually when most illicit logging is believed to occur.

Alongside developing an effective monitoring system, the research strives to include the sentiments and social understanding, using a survey-based input, of the people who are directly affected by deforestation and forest access limitations, an indigenous community called the Bakrawal community. As per previous research, UAVs and AI can significantly enhance environmental monitoring [3, 4, 5], but most studies focus on tropical or low-altitude regions during daytime. This study extends their application to a mountainous, high-altitude region with real-time detection goals.

This research aims to highlight the significance of simulating algorithm testing and prototyping before real-world deployment [2] and emphasises community collaboration as an important aspect of future system implementation. Findings from this research stand to inform how drone technology can support remote forest surveillance, improve transparency in enforcement, and potentially empower local stakeholders through shared monitoring tools.