This paper present and discusses algorithms suitable for visual navigation for mobile and flying robots. Three different algorithms were used to explore the direct method for the vision system. Those methods are Homography, Iterative Closest Point (ICP), and Horn’s Absolute Orientation. Those algorithms were tested on the camera with moving baseline. The relations between optimal baseline and depth distance were discussed. The camera’s calibration process has been presented and discussed. Several experiments with the different image noise level were performed. The noise levels influence on distance and pose estimation accuracy were discussed. Measurements and estimation errors for both mobile and flying robots were shown and compared with different methods.

Additional Metadata
Persistent URL dx.doi.org/10.1007/978-3-319-94517-0_6
Series GeoPlanet: Earth and Planetary Sciences
Citation
Al-Isawi, M.M.A. (Malik M. A.), & Sasiadek, J. (2019). Pose estimation for mobile and flying robots via vision system. In GeoPlanet: Earth and Planetary Sciences. doi:10.1007/978-3-319-94517-0_6