Positioning method, positioning device and readable storage medium

Abstract

The invention discloses a positioning method. The positioning method comprises the following steps of obtaining a target image; obtaining the candidate key information of a plurality of candidate keyframes coarsely matched with the target image from a pre-constructed map, wherein the map is composed of the key information of the plurality of key frames; generating a region-of-interest (ROI) of the target image, and performing the region-of-interest matching on the target image based on the candidate key information to obtain a region-of-interest matching pair of the target image for at leastone candidate key frame; matching the feature points in the region-of-interest matching pair based on the candidate key information to obtain a feature point matching pair of the target image for theat least one candidate key frame; and calculating the relative pose of the target image and the at least one candidate key frame according to the feature point matching pair.

Description

technical field [0001] The present invention generally relates to the field of image processing, and more specifically relates to a vision-based positioning method, a positioning device and a readable storage medium. Background technique [0002] Simultaneous Localization and Mapping (SLAM) is currently a hot research topic in image processing. The so-called SLAM is the integration of mobile target positioning and environmental map creation, that is, the target builds an incremental environmental map according to its own pose estimation and sensor perception of the environment during the movement process, and uses the map to realize its own positioning. [0003] In the prior art, the relocation method is usually used to realize the self-localization of the target, that is, a given location image is used to realize positioning and pose estimation in a pre-generated map. However, relocalization is well-defined but extremely challenging in practical applications. The relocali...

AI Technical Summary

Problems solved by technology

However, retargeting is well-defined but extremely challenging in practice

The relocalization task is extremely difficult due to a series of challenges such as the appearance of the scene in the open environment may change and the observation point of view changes.

The change of viewpoint and the change of appearance will lead to sparse matching points between the target image and the map, resulting in the failure of pose estimation or the large drift between the estimated pose and the actual pose

In addition, precise image-level matching is time-consuming, and map storage is memory-intensive

All of these make relocation very challenging

Images