ARTag

ARTag is a fiducial marker system used in augmented reality, computer vision, and robotics to enable robust camera pose estimation. It employs small square markers with a distinctive black border and a binary interior pattern. Each marker encodes a unique identifier, and the pattern is designed to be detectable even under perspective distortion and partial occlusion. The system is designed to provide reliable identification and orientation information for a given marker.

Detection and pose estimation are core components. In a captured image, the algorithm searches for square shapes with a solid border, extracts the corner points, and thresholds the interior to reveal a binary code. The code is decoded to determine the marker’s ID and its orientation. Using the known physical size of the marker and the camera's intrinsic parameters, the four corner correspondences enable estimation of the marker’s 3D pose relative to the camera, typically yielding a six-degree-of-freedom transformation (rotation and translation).

ARTag supports a dictionary of markers and can detect multiple markers within a single scene. The interior codes incorporate error-checking to reduce false detections and misidentifications. Related fiducial systems include ARToolkit, ARuco, and AprilTag; ARTag is among the earlier widely used schemes for AR calibration, pose-based object alignment, and robotics navigation.

Limitations include dependence on lighting and image resolution; detection can degrade with heavy occlusion, large perspective distortion, or markers that are far away. Despite these constraints, ARTag remains a foundational reference in fiducial marker design and pose estimation research.