Multi-UAV forest area inspection path planning based on concave polygon region decomposition

Forested areas, as important ecological resources, need to be patrolled regularly and efficiently to prevent fires, monitor pests and protect the ecology. Traditional methods relying on manual patrols or manned aircrafts face challenges such as low efficiency, high cost, and high safety risks, especially in complex terrains. In this paper, a coverage path planning algorithm based on spatial position ordering and convex decomposition is proposed for the coverage path planning problem of multiple unmanned aerial vehicles (UAVs) in concave polygonal regions. The algorithm decomposes the complex forested area into consecutive convex polygonal subregions by selectively removing concave vertices and improving the pre-ordered ear-shearing method. To further achieve balanced and spatially contiguous coverage, a region allocation method based on spatial position ordering and an explicit adjacency constraint (via a shared-edge-checking candidate set) is proposed to ensure region balance and neighbourhood constraints. Experimental results in a simulated forest area with 14 concave vertices show that the proposed method can reduce the coverage time by 54.7% and the total path length by 26.3% compared with single UAV operation. Further validation in conjunction with real terrain data from Xishuangbanna, Yunnan Province, reveals that through dynamic flight altitude optimisation (H = 150 m), the coverage rate is increased to 97.8% and the shading coefficient is reduced to 0.09, which verifies that the multi-UAV cooperative strategy still maintains its high efficiency in real scenarios. Additional comparisons performed under multi-UAV configurations also validated the efficiency of the algorithm, outperforming existing methods in minimising coverage time and total path length while maintaining a balanced task allocation. The algorithm greatly improves the coverage efficiency of forest area detection and provides technical support for dynamic monitoring of forest resources.