Visual Landmarks for Place Recognition in Changing Environments

Place recognition based on local image features is an approach to localization.

Changing environments pose a serious problem to current robotic systems aiming at long term operation. While place recognition systems perform reasonably well in static or low-dynamic environments, severe appearance changes that occur between day and night, between different seasons or different local weather conditions remain a challenge. We aim at adapting the idea of landmark-based localization for changing environments.

Example segmentation based landmarks.

Related Publications

Neubert, P., Protzel, P. (2016). Beyond Holistic Descriptors, Keypoints and Fixed Patches: Multiscale Superpixel Grids for Place Recognition in Changing Environments. IEEE Robotics and Automation Letters and presentation at IEEE International Conference on Robotics and Automation (ICRA), Stockholm, Sweden. Code

Krajník, T., Cristóforis, P., Kusumam, K., Neubert, P. and Duckett, T. (2016). Image features for visual teach-and-repeat navigation in changing environments. Robotics and Autonomous Systems (2016),DOI: http://dx.doi.org/10.1016/j.robot.2016.11.011

Neubert, P. (2015). Superpixels and their Application for Visual Place Recognition in Changing Environments. Dissertation, TU Chemnitz.

Neubert,P., Protzel, P. (2015). Local Region Detector + CNN based Landmarks for Practical Place Recognition in Changing Environments. Proc of. European Conference on Mobile Robotics (ECMR), Lincoln, England. Code