Visual place recognition

An graphic with arrows, boxes, and captions depicting the visual place recognition architecture. An image of an outdoor environment labelled "query image" points to feature extractor, which then points to a box labelled "feature aggregator," which then points to a box labelled "similarity search." A set of images labeled "database images" points to a box labelled "VPR model" which points to similarity search as well. Similarity search then points to another image of the environment with the label "database image."
A visualization of the modern approach to the visual place recognition task.

Visual Place Recognition (VPR) is a content-based image retrieval task in which, given a database of images and a query image, the goal is to return the image in the database that is closest in geographic location to the query image. This task is primarily focused on real-world images of outdoor urban locations, but can be applied to indoor environments. The modern approach to the VPR task is to train machine learning algorithms that can extract features which encode the geographic information of the image. VPR is primarily used in robotics and self-driving applications for localization, mapping, and planning.