作者: Christoforos Kanellakis , Petros Karvelis , George Nikolakopoulos
DOI: 10.1007/978-3-030-34995-0_10
关键词:
摘要: This work establishes a novel framework for characterizing the open space of featureless dark tunnel environments Micro Aerial Vehicles (MAVs) navigation tasks. The proposed method leverages processing single camera to identify deepest area in scene order provide collision free heading command MAV. In sequel and inspired by haze removal approaches, idea is structured around image depth map estimation scheme, without metric measurements. core contribution developed stems from extraction 2D centroid plane that characterizes center tunnel’s darkest area, which assumed represent space, while robustness scheme being examined under varying light/dusty conditions. Simulation experimental results demonstrate effectiveness challenging underground [1].