QR Code detection using parallel lines

Abstract

Several basic techniques are known for QR Code detection in digital images. This paper presents an approach that utilizes a special line parameterization, along with a variant of the Hough-transform, called PClines. The premised parameterization uses a parallel coordinate system, which allows faster accumulation than the slope-intercept method used by the basic Hough-transform, due to simpler computation. The presented technique can be easily adapted to any kind of matrix code. Mathematical and algorithmic foundations of this approach are presented, along with an implementation on Microsoft's .NET framework. Test results are also presented, emphasizing the effect of different parameters on the overall detection rate. These results are then compared to other implementations using the same PClines technique.