| Abstract | ARToolkit is a very successful and robust marker system used for Augmented Reality (AR), its robust performance has spawned many applications in AR and computer vision. ARToolkit consists of several 2D planar fiducial marker patterns and software that recognizes and identifies these markers in images. It functions well in finding markers; however its performance with respect to false positive detections and inter-marker confusion could use improvement. Quite often markers are confused for one another of falsely detected in the background. ARToolkit markers consist of a square black border enclosing a pattern that is compared to several stored patterns by correlation. This paper proposes a method to robustify this marker system by replacing this correlation step with a digital symbol method. The interior greyscale pattern is replaced by a digital pattern of 36 bits with contains a unique ID number protected from false detection with the digital code techniques of checksums and forward error correction (FEC). This proposed new system, ARTag has a very low and numerically quantifiable error rate, has a reduced processing item, and can encode up to 2046 different unique Id's with no need to store patterns. Experimental results are shown validating this method. |
|---|
