Indoor Place Recognition System for Localization of Mobile Robots
Abstract: In this paper we present a method for robots to do visual place recognition and categorization. The robot learns from experience and then recognizes previously observed places in known environments and categorizes previously unseen places in new environments. This system has been practically tested with a novel dataset developed by us to validate the theoretical results of the proposed system. A Histogram of Oriented Uniform Patters (HOUP) descriptor has been used to represent an image and then appropriate classifiers have been used to perform the classification tasks. It is shown that our method not only performs well on our dataset but also on existing datasets. A major contribution of this work is that this is the first real time implementation of a HOUP descriptor on two mobile robot platforms. Finally we built a novel dataset of seventeen indoor places for doing place recognition and validated our method in real time on this dataset.
Robots we used:
- Raghavender Sahdev and John K. Tsotsos, “Indoor Place Recognition for Localization of Mobile Robots,” In 13th Conference on Computer and Robot Vision, 2016, Victoria BC, June 1-3, 2016.
- Raghavender Sahdev and John K. Tsotsos, “Place Recognition for Localization of Mobile Robots”, poster abstract in International Conference on Perceptual Organization, June 2015.
- Raghavender Sahdev, Asheer Bachoo and John K. Tsotsos, “Place Recognition for Localization of Mobile Robots”, poster at the 2015 NCFRN Field Trials in Kelowna.
If you use the dataset, please cite the paper as
R. Sahdev and J. K. Tsotsos, “Indoor Place Recognition for Localization of Mobile Robots,” In 13th International Conference on Computer and Robot Vision, 2016, Victoria, BC, June 1-3, 2016.