Vision based square matrix design and detection for robotic routing

Robot vision is essential to aid in performing robotic tasks. Existing robotic path tracking methods used in the industry such as line following are simple but not capable to cater for continuous floor plan changes. This paper proposes a new vision based pattern design and detection algorithm that can be implemented on an existing mobile robot structure to provide path tracking and routing. A unique pattern that is capable to embed information is successfully designed. The detection algorithm is tested at different lighting conditions, pattern size and orientation. A larger pattern size and good lighting conditions has produced favorable high detection of the pattern. Subsequently, after pattern detection, a unique identifier of the pattern is generated in the form of a binary matrix. The generated binary matrix is proposed to be used as a location identifier to aid in mobile robot routing.