Visual positioning method and apparatus, and computer-readable storage medium

A visual positioning apparatus, comprising:

• at least one processor;
• and
• a memory communicatively coupled to the at least one processor,
• wherein the at least one processor is configured to:
• collect a first image;
• identify the first image to determine entity types for a plurality of first entities presented in the first image and a distance between every two first entities;
• generate a first graph based on the distance between every two first entities, wherein the first graph comprises nodes corresponding to the plurality of first entities and edges for connecting different nodes in the first graph, the node is configured to indicate the entity type for the corresponding first entity, and the edge is configured to indicate that a distance between two first entities corresponding to two connected nodes is less than a threshold distance;
• determine description information of the first entity corresponding to each node in the first graph based on the entity types indicated by nodes that are in random walk paths starting from the corresponding node;
• determine a candidate entity for each first entity from a plurality of second entities based on pre-stored description information for the plurality of second entities, wherein the description information of the candidate entity for the first entity matches the description information of the first entity, the description information of each second entity is configured to indicate entity types for second entities adjacent to the corresponding second entity;
• and
• position a collection area of the first image based on an area where the candidate entities are located.

A non-transitory computer-readable storage medium having computer programs stored thereon, wherein when the programs are executed by a processor, a visual positioning method is implemented, the method comprising:

• collecting a first image;
• identifying the first image to determine entity types for a plurality of first entities presented in the first image and a distance between every two first entities;
• generating a first graph based on the distance between every two first entities, wherein the first graph comprises nodes corresponding to the plurality of first entities and edges for connecting different nodes in the first graph, the node is configured to indicate the entity type for the corresponding first entity, and the edge is configured to indicate that a distance between two first entities corresponding to two connected nodes is less than a threshold distance;
• determining description information of the first entity corresponding to each node in the first graph based on the entity types indicated by nodes that are in random walk paths starting from the corresponding node;
• determining a candidate entity for each first entity from a plurality of second entities based on pre-stored description information for the plurality of second entities, wherein the description information of the candidate entity for the first entity matches the description information of the first entity, the description information of each second entity is configured to indicate entity types for second entities adjacent to the corresponding second entity;
• and
• positioning a collection area of the first image based on an area where the candidate entities are located.