Classification of fiducial markers in challenging conditions with SVM

Abstract

Square fiducial markers are a popular tool for camera pose estimation because of their high robustness and performance. However, state-of-the-art methods perform poorly under difficult image conditions, such as camera defocus, motion blur, small scale or non-uniform lighting. This paper tackles the marker identification problem as a classification one, proposing a methodology to train such classifiers by creating a synthetic dataset of markers affected by several transformations. Our approach employes a SVM for marker identification. Statistical analyses have been performed in order to determine the best SVM configuration for our problem, and the best one is compared to the ArUco fiducial marker systems in challenging video sequences. The results obtained show that the proposed method performs significantly better.