Investigation of wall inspection algorithm considering image recognition by a mobile robot

Abstract

After a large earthquake happens, it is necessary to investigate infrastructure as soon as possible. To prevent suffering from secondary disaster, mobile robots should be utilized for the investigation. By using an image recognition technology, mobile robots can find where is a trouble such as a crack. In the case of utilizing the image recognition technology, a camera should be moved by considering recognizable condition. In this paper, we describe wall inspections by a mobile robot using image recognition technology. We discussed how to inspect the wall thoroughly and constraints of an algorithm considering recognizable condition. Then, we proposed an algorithm to give the range of the robot position and the parameter of PTZ camera. Finally, we demonstrated proposed algorithm to verify it by an experiment that a mobile robot recognizes QR codes that imitated trouble points on a wall.