Development of a doctor following mobile robot with mono-vision based marker detection

Abstract

This work presents a simple marker-based person detection method for a mobile robot that can follow a person in indoor environments. Expensive laser range finders or RFID which can provide very accurate range measurements have been used by several researchers for person detection. Recently, vision based approaches have been popular for person detection within a group of multiple people using stereo cameras. In this paper of proposed implementation, an inexpensive single camera has been used to acquire video frames to detect a specific target person and determine its position. A new detection method using color and shape based marker technique has been advanced in this work. The experimental results show that the proposed algorithm can detect a target person under various conditions such as marker features and lighting conditions. Index Terms-Following robot; Mono-vision; Object tracking; Marker-based detection.