OCR Training and Simulation for Person Follower Robot in an Indoor Environment

Abstract

This paper explores the usage of Optical character recognition (OCR) for a person follower behavior. The character on a person's uniform is captured using a network camera, which is interfaced on a Robotic system. In particular, this paper addresses the training of the characters to be used for the person following behaviour of a Robot, to ensure that the tracking is done properly. The graphical design is developed with LabVIEW software of National Instruments. Simulation results are furnished for the proposed training scheme. Aim: To train and simulate an OCR for person follower behavior. Study Design: The study of OCR method, exploring its use for person follower behavior is proposed and a LabVIEW based graphic design is developed. Place and Duration of Study: Department of Electronics and Communications engineering, B V Raju Institute of Technology, Narasapur, Medak (Dt), Telangana, India, between July 2014 and July 2016. Methodology: We included different image samples of each character 'L', 'F' and 'R' at various distances from camera which covered maximum scenarios. OCR module present in LabVIEW was used for training samples. Results: The screenshots of simulated results of the proposed training scheme are presented. Conclusion: The training of sample images of the characters is done in various positions and to a maximum field of view of the camera. The test results have been found to be satisfactory.