Picture construction/vehicle positioning method and system, terminal and computer storage medium

Abstract

The invention provides a picture construction / vehicle positioning method and system, a terminal and a computer storage medium, and the picture construction method comprises the steps: loading a feature map of a vehicle in an area where a historical travel path belongs, so as to obtain all key frames in the feature map; the method comprises the following steps: acquiring an initial pose of a vehicle, selecting N key frames in an electronic map by taking the initial pose as a center, and forming a current local slice sub-graph by the N key frames; the distance between the N key frames and the center is within a preset distance range; wherein N is greater than or equal to 1; and searching the key frames with the common-view relationship in the current local slice sub-graph, and constructing the local slice sub-graph required by vehicle positioning through the key frames with the common-view relationship. According to the method, through the operation of deserializing the map first but not allocating the memory at the same time, the memory occupied by loading the map is greatly reduced, meanwhile, the loading rate is also greatly improved, and the burden caused by an oversized map to vehicle end hardware is reduced.

Description

technical field [0001] The invention belongs to the technical field of vehicle intelligent parking, and relates to a method and a system, in particular to a picture construction / vehicle positioning method, system, terminal and computer storage medium. Background technique [0002] At present, the relocation methods commonly used by vehicle positioning systems on the market are based on known feature point maps. By extracting feature points of the current visual image and matching with known map feature points, the initial position of the vehicle in the global coordinate system is calculated. pose, realize the global positioning of the vehicle, and complete the initialization of the system. During the entire driving process, the real-time pose of the vehicle is determined based on global map matching. There are many drawbacks in the above scheme in use. First of all, when the indoor parking environment is large enough, it will directly lead to a sudden increase in the memor...

AI Technical Summary

Problems solved by technology

First of all, when the indoor parking environment is large enough, it will directly lead to a sudden increase in the memory usage of the established feature point map. Therefore, each time the vehicle is started, it will go through a long map loading process, which seriously affects the parking efficiency and user experience.

Secondly, when the map is large enough, the hardware environment of the car is not enough to allocate such a large memory for storing the visual feature point map, resulting in failure to load

At the same time, it is more difficult and accurate to match the global map to locate the current pose than to locate in the local small map, and the mis-match and mis-positioning are serious.

[0003] Therefore, how to provide a picture construction / vehicle positioning method, system, terminal and computer storage medium to solve the problem that when the surrounding environment is large enough in the prior art, the memory usage of the established map directly increases, the map loading process is long, and the When the hardware environment of the car is not enough to allocate memory, it will lead to loading failure and the difficulty of matching and locating the current pose in the global map, the accuracy rate is low, and the frequency of mislocation is high. These defects have become urgent problems for those skilled in the art. question

Images