PCA-Based Detection of Drone Landing Sites Using Normal Vector Estimation

This study proposes a principal component analysis (PCA)-based algorithm for the detection of drone landing sites, designed to be both simple and efficient while considering the hardware limitations inherent in drones. The algorithm utilizes point cloud data (PCD) to identify potential landing sites through PCA-based normal vector estimation, demonstrating its applicability even in resource-constrained systems. Experimental validation, utilizing data acquired from drones, confirmed that the proposed method identified safe landing sites more rapidly and reliably than algorithms presented in previous studies. The proposed approach ensures both efficiency and accuracy, rendering it a promising solution for resource-limited systems, such as drones.