As two-dimensional barcodes can store data efficiently with error correction ability, they are widely used as landmarks in Automatic Ground Vehicle (AGV) guidance systems. AGV with a scanner can obtain the approximate position in its operation field if the position information is decoded into landmarks pasted on the ground, and a more specific position can be calculated if the landmark center is determined in an image. In this paper, we propose an algorithm to detect the position for the two-dimensional barcode landmark composed of four Data Matrix symbols since four symbols are more robust to a single one. Firstly, a voting strategy is adopted to detect the finder pattern (L shape) of a Data Matrix symbol after candidate contours are selected. Then each Data Matrix code is isolated and decoded by libdmtx—an open-source software library. Finally, each block’s quality is evaluated by bipolarity as a weight to calculate the position of the landmark. The experiment shows that our algorithm can achieve real-time detection with high accuracy on embedded systems.
A Real-Time Algorithm for Multiple Data Matrix Codes Localization
Lect. Notes Electrical Eng.
2021-10-30
11 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
SCOOP: A Real-Time Sparsity Driven People Localization Algorithm
British Library Online Contents | 2014
|SWARMON—Real-Time Localization System
Springer Verlag | 2015
|A relaxation algorithm for real-time multiple view 3D-tracking
British Library Online Contents | 2002
|