Map construction method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a map construction method and device, electronic equipment and a storage medium. The map construction method comprises the steps of taking the size of the relative displacement length detected by a physical sensor as a threshold value for opening and closing map construction, and constructing sub-maps in a segmented manner; and aligning the shape and the scale of the sub-maps to a map track generated by the physical sensor according to the timestamp, and updating all map points. According to the invention, map distortion can be effectively reduced, the scale is completely aligned to a physical space, track calculation of the physical sensor is completely consistent with representation of the map in the map reuse process, meanwhile, segmentation expression is conducted on the map, the occupied space of the map is greatly reduced, and the performance requirement for a hardware system in the SLAM process is lowered.

Description

technical field [0001] The embodiments of the present application relate to the technical field of computer vision, and in particular to a map construction method, device, electronic device and storage medium. Background technique [0002] Real-time positioning and map construction technologies are currently widely used in the field of intelligent robots. Most of the maps constructed by SLAM are required to be able to meet the later reuse, thus contributing to accurate positioning. Common types of SLAM maps include point cloud maps, probabilistic grid maps, visual feature maps, visual / laser semantic maps, and high-precision vector graphics. Among them, because the visual map has some advantages such as low sensor price and low establishment cost, the visual map and visual positioning method are more universal in many application scenarios. The process of visual SLAM from the successful initialization to the end will form a map with many descriptive visual features. The map ...

AI Technical Summary

Problems solved by technology

The process of visual SLAM from the successful initialization to the end will form a map with many descriptive visual features. The map can be reused in the visual feature positioning system. However, in the prior art, the map occupies too much space, and the visual SLAM process cannot be established. For a complete map with a certain scene scale, the map reuse process cannot load such a large amount of data map; the average cost of computer memory and computing power is too large, and the system pressure for the map building process and subsequent map reuse is high; map distortion and scale distortion, Map reuse is not robust

Images