UAV identifies passable paths in unknown indoor environments based on Deep-learning

In order to identify passable paths in unknown indoor environments for UAV, we implement a method to help UAV perceive the indoor environments based on deep-learning. At first, in the indoor environments, UAV’s passage to the next area is through an open door. Therefore, we need to identify and distinguish the image area of the door from the UAV cameras. And then, we identify the opening or closing state of a door by means of a feature determination. In this paper, the simulation experiment can identify the real indoor environment and is able to identify if it is a passable path.