The ability to recognize known places is an essential competence of any intelligent system that operates autonomously over longer periods of time. Approaches that rely on the visual appearance of distinct scenes have recently been developed and applied to large scale SLAM scenarios. Recognizing known places despite severe changes in their appearance due to weather, illumination changes, time of day, or seasons is a particularly challenging task.
In the above illustration, the image in the top-left corner is compared to the images from the map. We aim at identifying the two corresponding images of the same place (on the right side) although they are subject to severe appearance changes due to the different seasons.
Our research comprises three approaches to the problem of visual place recognition:
