Exploiting Repetitive Patterns for Fast Succinct Map Matching

Map matching is a critical problem of robotic mapping and localization applications which has attracted broad interests in robot vision community. Despite its accuracy and efficiency, the popular RANSAC-based algorithm suffers from large memory requirements, which is proportional to number and size of the maps. In this paper, our goal is to realize fast succinct map matching by introducing a compact and discriminative approach for part-based scene modeling. The main difficulty is that part modeling of environment maps is a novel task and there is no accepted definition available in literature. To overcome this limitation, we exploit the fact that typical environments (e.g. indoor, street, forests, suburban, etc) contain highly repetitive patterns, and take the approach of repetitiveness-based scene compression. Our method has two key steps. First, our method mines a small number of large common patterns that well explain an input scene from a known reference map, by introducing an efficient common pattern discovery (CPD) using integral images. Second, we represent parts by using traditional bounding box (BB)-based object annotation technique. Since BB is a far lower dimensional representation compared with the original map data and matching between BBs can be fast using a simple box intersection operation, both compactness and efficiency are achieved.