Positional markers, also known as fiducial markers, have found widespread use for camera pose estimation. The camera's pose can be estimated at low cost and high speed by detecting fiducial markers while dealing with occlusion and distortions. The estimated pose can then be combined with other sensor data or directly be used for AGV navigation. The markers only have to be printed out, placed in the desired environment, and then detected. However, a large variety of different marker types and different libraries for their detection exist. Choosing the correct marker type for the proper application can therefore be a time-consuming process. This thesis presents an overview of the different marker types and testing results using a generated test set of ArUco, Apriltag, and ARTag markers. Through the construction of prototype implementations, the project validated the effectiveness of the ArUco and the AprilTags libraries. Both libraries were tested and analyzed regarding detection time, accuracy and robustness. Finally, the thesis attempted to apply the advantages of the 2D bar code Aztec Code to fiducial markers. For this purpose, multiple styles of combined markers were constructed and tested. This style of combined marker can be used to encode information and execute pose estimation using a single marker, however, the high robustness towards occlusion of Aztec Code could not be preserved through modification. All results were evaluated with a focus on applicability to AGV Navigation.     «

Positional markers, also known as fiducial markers, have found widespread use for camera pose estimation. The camera's pose can be estimated at low cost and high speed by detecting fiducial markers while dealing with occlusion and distortions. The estimated pose can then be combined with other sensor data or directly be used for AGV navigation. The markers only have to be printed out, placed in the desired environment, and then detected. However, a large variety of different marker types and...     »