Robust recognition of circular coded markers with relaxed constraints on multiple features

Coded marker recognition is a significant task in computer vision and photogrammetry communities. Existing methods are only available for high-quality images, and perform poorly on images with large viewing angles, low resolutions, and unevenly distributed intensities. To this end, we design a circular coded markers based on Schneider’s algorithm and propose a robust method for its recognition on degraded images. The global images are divided into sub-images in prior to image binarization to avoid contour disappearance, which usually occurs in the binarization of the global image with unevenly distributed intensities, and then the coded marker borders are extracted in the sub-images. After that, the images inside the borders are re-binarized for extracting high-quality contours; and the concept of relaxed constraints on multiple features is applied to filter out false contours. Finally, the geometric center is located and the encrypted information is decoded based on the identified contours. The proposed method is extensively evaluated and compared with the state-of-the-art methods (the original Schneider’s and ArUco algorithms). Preliminary results demonstrate the superiority of the proposed method over the baselines regarding the recognition rates on degraded images.