WO2020006685A1

WO2020006685A1 - Method for establishing map, and terminal and computer readable storage medium

Info

Publication number: WO2020006685A1
Application number: PCT/CN2018/094329
Authority: WO
Inventors: 韩立明; 林义闽; 廉士国
Original assignee: 深圳前海达闼云端智能科技有限公司
Priority date: 2018-07-03
Filing date: 2018-07-03
Publication date: 2020-01-09
Also published as: CN109074757A; CN109074757B

Abstract

Some embodiments of the present application provide a method for establishing a map, and a terminal and a computer readable storage medium. The method for establishing the map comprises: acquiring N maps describing a same space, wherein N is an integer greater than one; extracting respective spatial invariant features of the N maps; according to the spatial invariant features of the N maps, merging information of map points in the N maps to obtain an merged map.

Description

Method, terminal and computer-readable storage medium for establishing map

Technical field

The present application relates to the field of detection, and in particular, to a method, a terminal, and a computer-readable storage medium for establishing a map.

Background technique

Current Visual Simultaneous localization and mapping (VSLAM) technology uses a visual sensor to collect a set of images, and uses feature extraction methods to extract feature points that characterize the spatial environment from the collected images, and through binocular Parallax or monocular motion parallax, etc., calculate the position information of the feature points in space, or use a camera with depth information to directly obtain the position information corresponding to the feature points. When the terminal captures a group of images, the feature points in these images can form a map for positioning in VSLAM. During the movement of the terminal, the feature points in the image captured by the sensor at the current time are matched with the feature points in the map to obtain the position and attitude of the terminal.

In the above-mentioned disclosed VSLAM technical solution, the matching of the feature points in the image captured by the terminal and the feature points in the map is the core issue of the VSLAM technology. The feature points extracted through the image have a certain invariance in a short time and a small scene, so that the current VSLAM technical solution can be applied in the above situation. However, in the case of long and large scenes, due to changes in the lighting conditions and shooting angles of the captured images, the feature points are likely to change, which makes it difficult to correctly match the feature points in the image with the map. In order to solve this problem, the existing solution is to establish multiple maps for the same space for terminal positioning. That is, in a short period of time, multiple sets of images are collected at a fixed angle, and VSLAM technology is used to establish a feature point map for each group of images, and then multiple maps are added to the labels and managed separately.

technical problem

The inventor discovered during the research of the prior art that the above-mentioned mapping scheme has at least the following disadvantages: due to the existence of multiple maps, the terminal needs to perform operations such as expanding, updating, and deleting each map separately, and the management efficiency is low. It can be seen that how to improve the management efficiency of maps is a problem that needs to be solved.

Technical solutions

A technical problem to be solved in some embodiments of the present application is how to improve map management efficiency.

An embodiment of the present application provides a method for building a map, including: obtaining N maps describing the same space, where N is an integer greater than 1; extracting spatially invariant features of each of the N maps; and according to the N maps The respective space-invariant features are combined with the information of the map points in the N maps to obtain a merged map.

An embodiment of the present application further provides a terminal including at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor. The processor executes to enable at least one processor to execute the method for establishing a map mentioned in the foregoing embodiment.

An embodiment of the present application further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the method for establishing a map mentioned in the foregoing embodiment is implemented.

Beneficial effect

Compared with the prior art, since the information of the map points in multiple maps is stored in the merged map, the embodiment of the present application only needs to perform operations such as expanding, deleting, and updating map point information. The operation of the map improves the management efficiency of the map. In addition, when the terminal uses the merged map for positioning, it is not necessary to switch the map for matching, which improves the positioning efficiency.

BRIEF DESCRIPTION OF THE DRAWINGS

One or more embodiments are exemplified by the pictures in the accompanying drawings. These exemplary descriptions do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the drawings in the drawings do not constitute a limitation on scale.

FIG. 1 is a flowchart of a method for establishing a map according to a first embodiment of the present application; FIG.

2 is a flowchart of a method of combining information of map points of N-1 maps into a reference map according to a second embodiment of the present application;

FIG. 3 is a schematic diagram showing a relationship between a map establishing method and a positioning method according to a second embodiment of the present application; FIG.

4 is a flowchart of a method for positioning using a merged map according to a second embodiment of the present application;

FIG. 5 is a schematic structural diagram of a terminal according to a third embodiment of the present application.

Embodiments of the invention

In order to make the purpose, technical solution, and advantages of the present application clearer, some embodiments of the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the application, and are not used to limit the application.

The first embodiment of the present application relates to a method for establishing a map, which is applied to a terminal. As shown in FIG. 1, the method for establishing a map includes the following steps:

Step 101: Obtain N maps describing the same space. Here, N is an integer greater than 1.

Specifically, the terminal collects multiple image sequences in the same space through a visual sensor, and extracts information and spatially invariant features of the map points extracted from each group of image sequences, and establishes based on the information and spatially invariant features of the extracted map points. Maps for each image sequence.

It should be noted that the map may be a map established by the terminal according to the image sequence collected by the visual sensor, or a map transmitted to the terminal by the cloud or other terminals, and the source of the map is not limited in this embodiment.

Step 102: Extract the spatially invariant features of each of the N maps.

Specifically, the spatially invariant feature may be any one or any combination of line features, semantic features, and tag information. Among them, the line feature refers to the characteristics of the line segments in the map, including the length, angle, and intersection of the line segments. Semantic features refer to features assigned to the same map points in different maps. The terminal recognizes the same map points in different maps by using an image recognition method, and assigns the same features to the same map points. Marker information refers to the information on the map of the location markers in space. Among them, the positioning mark is a mark arranged at a plurality of fixed positions in the space, and the mark may be a quick response (Quick Response (QR) code or data matrix (data matrix (DM) code.

Step 103: Combine the information of the map points in the N maps according to the space invariant features of the N maps to obtain a merged map.

The method of merging the information of the map points in the N maps to obtain the merged maps according to the spatially invariant features of the N maps is exemplified below.

Method 1: The terminal creates a new map as a reference map, and determines the relative pose relationship between the reference map and map A in the N maps. Among them, map A is any one of the N maps. The terminal adds the information of the spatially invariant features and map points in the map A to the reference map according to the relative pose relationship between the reference map and the map A. The terminal matches the spatially invariant features in the N-1 maps other than map A with the spatially invariant features in the reference map, and determines the corresponding N-1 maps with The relative pose relationship of the base map. The terminal combines the information of the map points of the N-1 maps into the reference map according to the relative pose relationship between the N-1 maps and the reference map, to obtain a merged map.

Method 2: The terminal selects one map from the N maps as the reference map, and matches the spatially invariant features in the N-1 maps other than the reference map in the N maps with the spatially invariant features in the reference map, respectively. , According to the respective matching results of the N-1 maps, determine the relative pose relationship between the N-1 maps and the reference map, respectively. The terminal combines the information of the map points of the N-1 maps into the reference map according to the relative pose relationship between the N-1 maps and the reference map, to obtain a merged map.

For the sake of clarity, in this embodiment, a method for determining the relative pose relationship between map B and map C is taken as an example, and in the methods 1 and 2, the terminal determines the relative pose relationship between N-1 maps and the reference map, respectively. method. The terminal matches the spatially invariant features in map B with the spatially invariant features in map C. Based on the matching results, point perspective is used. n Points (PnP) pose measurement algorithm to solve the relative pose relationship between map B and map C. Optionally, during the solution process using the PnP pose measurement algorithm, the relative pose relationship is optimized by a beam adjustment (BA) algorithm.

It should be noted that in the process of merging the information of the map points of the N-1 maps into the reference map, there may be a case where the same map points are not merged into the same map point. In view of this situation, after obtaining the merged map, the terminal may merge map points with a distance less than a preset value in the merged map. The position information of the merged map point is determined according to the position information of the map point before the merge. The value can be determined according to actual needs. For example, the spatial coordinates of map point a in the merged map are (xa, ya, za), and the spatial coordinates of map point b are (xb, yb, zb), where xa represents the abscissa of map point a, and ya represents the map. The vertical coordinate of point a, za indicates the vertical coordinate of map point a, xb indicates the horizontal coordinate of map point b, yb indicates the vertical coordinate of map point b, and zb indicates the vertical coordinate of map point b. The terminal calculates the distance between map point a and map point b, and judges whether the distance between map point a and map point b is less than a preset value. If it is determined to be smaller, merge map point a and map point b to obtain map point c. The spatial coordinates of map point c are ((xa + xb) / 2, (ya + yb) / 2, (za + zb) / 2). The description information set of map point c includes the description information set of map point a and Set of description information of map point b.

It is worth mentioning that determining the location information of the merged map points based on the location information of the map points before the merger is equivalent to collecting the location information of the same map point multiple times, which improves the accuracy of the location information of the map points .

It is worth mentioning that, based on the information of the map points of the N maps, a merged map is established, which improves the management efficiency of the map. After the merged map is established by the method for establishing a map mentioned in this embodiment, when the terminal needs to expand the map, determine the position of the information that needs to be expanded in the merged map, and add the information that needs to be expanded at that position. Each map of the map is expanded with map information. When the terminal needs to delete or update the map point information, it only needs to delete or update the map point information in the merged map, and does not need to delete or update the map point information for each of the N maps.

Compared with the prior art, in the method for establishing a map provided in this embodiment, since information of map points in multiple maps is stored in a merged map, when the terminal performs operations such as expansion, deletion, and update of map point information, , Only need to operate the merged map, which improves the management efficiency of the map. In addition, when the terminal uses the merged map for positioning, it is not necessary to switch the map for matching, which improves the positioning efficiency.

The second embodiment of the present application relates to a method for establishing a map. This embodiment is a further refinement of the first embodiment, and specifically illustrates that according to the relative pose relationship between N-1 maps and the reference map, N The process of merging the information of the map points of the -1 map into the reference map.

Specifically, a flowchart of a method for combining information of map points of N-1 maps into a reference map is shown in FIG. 2 and includes the following steps:

Step 201: Determine the corresponding relationship between the map points in the N-1 map and the map points in the reference map according to the relative pose relationship between the N-1 maps and the reference map.

Step 202: Determine the description information set of the map points in the merged map according to the corresponding relationship between the map points in the N-1 maps and the map points in the reference map, and the description information of the map points in the N-1 map.

It should be noted that after obtaining the description information set of the map points, the terminal may cluster the description information in the description information set of the map points by using a clustering algorithm to obtain the clustered description information set. The clustered description information set determines the merged map.

Specifically, the terminal clusters the description information in the description information set of the map points, classifies the description information of the similar map points into one category, and records the center point of the category as the clustered description information. For example, after determining that there is description information of unclassified map points, the terminal randomly selects description information of one map point as the center point from description information of unclassified map points. For this center point, the terminal performs the following operations: finds all description information that is within a first preset value from the center point and records it as set M; determines the vector from the center point to each element in set M, and compares all vectors Add to get the offset vector; control the center point to move in the direction of the offset vector, the distance of travel is half the modulus of the offset vector; determine whether the modulus of the offset vector is less than the second preset value, and if it is less, record the center point , Otherwise, determine the vector from the current center point to each element in the set M, add all the vectors to get the offset vector, control the center point to move in the direction of the offset vector, and move the distance by half the modulus of the offset vector ... ... until the magnitude of the offset vector is less than the second preset value. After the terminal determines that all description information of the map points is classified, the clustered description information is determined according to the recorded center point.

It is worth mentioning that, through clustering algorithms, the description information of similar map points is merged to reduce the amount of map data.

In another specific implementation, the map point information further includes shooting information corresponding to the description information of the map point, where the shooting information includes shooting brightness and shooting angle. After determining the description information set of the map points, the terminal clusters the description information in the description information set of the map points through a clustering algorithm to obtain the clustered description information set. If the clustering algorithm aggregates the L description information into one category, according to the shooting brightness corresponding to the L description information, the shooting brightness corresponding to the description information clustered by the L description information, and the shooting corresponding to the L description information. Angle to determine the shooting angle corresponding to the L description information clustered description information; based on the shooting brightness corresponding to the clustered description information and the shooting angle corresponding to the clustered description information to determine the clustered description information set Shooting information corresponding to each description information in. In specific implementation, the terminal calculates the average value of the shooting brightness corresponding to the L description information as the first average value, and uses the first average value as the shooting brightness corresponding to the description information clustered by the L description information. The terminal calculates an average value of the shooting angles corresponding to the L description information as the second average value, and uses the second average value as the shooting angle corresponding to the description information clustered by the L description information.

The following describes the process of determining the shooting information corresponding to the clustered description information in combination with the actual scene.

For example, the description information corresponding to the map point P includes first description information (ka), second description information (kb), third description information (kc), fourth description information (kd), fifth description information (ke) ... …. Among them, the shooting brightness of ka is the first shooting brightness (hka), the shooting brightness of kb is the second shooting brightness (hkb), the shooting brightness of kc is the third shooting brightness (hkc) ..., the shooting angle of ka is the first The shooting angle (tka), the shooting angle of kb is the second shooting angle (tkb), the shooting angle of kc is the third shooting angle (tkc) ... Through the clustering algorithm, ka, kb, and kc are divided into one group, The shooting brightness of the description information after the class = (hka + hkb + hkc) / 3, and the shooting angle of the description information after the cluster = (tka + tkb + tkc) / 3.

It is worth mentioning that merging N maps reduces the storage space of the maps.

The following example illustrates why combining N maps can reduce the storage space of the map. In the unmerged N maps, the storage format of the description information of the map points is: (shooting information 1, map point id, location information, map point description information 1); (shooting information 2, map point id, location information, Map point description information 2); (shooting information 3, map point id, location information, map point description information 3) ... (shooting information p, map point id, location information, map point description information p) Among them, p represents the number of description information of map points, and map point id represents the number of map points by the terminal in order to determine the same map point in different maps. After the merged map is obtained by using the method for building a map mentioned in this embodiment, the description information of the map points in the merged map is stored in the form of a description information set, the storage format is: {map point id, location information, (shooting information 1, Description information of map points 1), (shooting information 2, description information of map points 2), (shooting information 3, description information of map points 3) ... (shooting information q, description information of map points q)}. Compared with the unmerged map, the merged map has q less than or equal to p after clustering the description information in the description information set of the merged map. Assume that the data type of the shooting information is unsigned character data (unsigned char), the data size is 1 byte, the data type of the map point id is unsigned int data (unsigned int), the data size is 4 bytes, and the position information is the spatial coordinates of the map point. The data type is floating point data (float), the data size is 4 bytes, so the data size of the location information is 4 * 3 = 12 bytes, the data type of the map point description information is unsigned char, and the data size is 8 bytes. The size of the data occupied by the N maps before the merger is p + p * (4 + 12 + 8) * the number of map points = p + 24 * p * the number of map points, and the data size of the merged map is [4 + 12 + q * ( 1 + 8)] * Number of map points = (16 + q * 9) * Number of map points, because the number of map points is usually a large number, for large scenes, it will generally reach more than 100,000 orders of magnitude. The ratio k between the data and the data of the N maps is approximately (16 + q * 9) / (24 * p). When q = p> 1, k is about 37.5% -70.8%, so the storage space of the combined map is more small.

It should be noted that, after the merged map is established by using the method for establishing a map mentioned in the first embodiment or the second embodiment, the terminal may perform positioning based on the merged map. The relationship between the method of establishing a map and the positioning method is shown in FIG. 3. In the figure, map merging refers to the merging of maps corresponding to different shooting information. Single state positioning refers to positioning according to the shooting information when the terminal initiates a positioning request. The method for positioning a terminal using a merged map is shown in FIG. 4 and includes the following steps:

Step 301: Acquire an image for positioning.

Step 302: Extract description information of the map points in the image.

Step 303: Determine description information in the merged map that matches description information of map points in the image.

In specific implementation, the terminal matches the description information of the map points in the image with all the description information in the merged map, and determines the matching description information.

In another specific implementation, when the terminal extracts the description information of the map in the image, the terminal determines the shooting information of the image by analyzing the image or by using a sensor (such as a light sensor) on the terminal. The terminal determines the shooting information corresponding to the description information of the map point in the image according to the shooting information of the image. The terminal filters and obtains from the description information set of the map points in the merged map according to the shooting information corresponding to the description information of the map points in the image and each description information in the description information set of the map points in the merged map. Descriptive information for matching. The terminal determines a temporary map according to the description information used for matching, and performs positioning according to the temporary map. The terminal determines the description information that matches the description information of the map point in the image according to the description information for matching in the temporary map. When the shooting information includes shooting brightness and shooting angle, the specific process for the terminal to determine the description information for matching is as follows: the terminal calculates a first difference between the shooting brightness of the image and the shooting brightness corresponding to each description information, and the shooting angle of the image A second difference value of the shooting angle corresponding to each description information; according to the first difference value and the second difference value, selecting M description information from the description information set of the map point as the description information for matching; wherein, M is a positive integer.

The following describes the process by which the terminal obtains the descriptive information for matching from the feature information set of the map points in the map in combination with the actual scenario.

Assume that the terminal acquires the shooting brightness of the map points in the second image as H, the shooting angle is L, and the feature information set of the map points in the map is {(description information a1, shooting brightness h1, shooting angle l1), (description information a2 , Shooting brightness h2, shooting angle l2), (description information a3, shooting brightness h3, shooting angle l3) ...}, that is, the shooting brightness corresponding to the description information a1 is h1, the corresponding shooting angle is l1, and the description information a2 corresponds to The shooting brightness is h2, and the corresponding shooting angle is l2 ... The terminal calculates the difference between h1 and H and the difference between l1 and L, and sets different weights for the two differences according to actual needs, so as to determine the second image. The distance dt between the shooting information of the map points and the shooting information of the map points in the map, that is, dt = a * (H-h1) + b * (L-l1), where a is the weight of the difference in shooting brightness, b is the difference weight of the shooting angle. By analogy, the terminal calculates the distance between the shooting information of the map points of the second image and the shooting information corresponding to each description information in the feature information set of the map points in the map. The terminal sorts each description information in the feature information set of the map points in the map in the order of the distance from small to large, and selects the first M description information ranked as the description information for matching.

It should be noted that those skilled in the art can understand that in practical applications, a method of setting a distance preset value can also be used to filter the feature information set, that is, the distance of the shooting information corresponding to the description information of the map point in the second image. The description information smaller than the distance preset value is used as the description information for matching.

Step 304: Obtain position information of the map points corresponding to the matching description information from the merged map, and determine a positioning result according to the obtained position information.

Specifically, the terminal determines a positioning result by using a pose estimation algorithm, for example, a PnP pose measurement algorithm, according to a matching result between a map point in the map and a map point in the second image. The positioning result includes the pose information of the terminal.

In specific implementation, before obtaining the location information of the map points corresponding to the matching description information from the merged map and determining the positioning result based on the obtained location information, the terminal may also first determine that at least T map points in the image match the map successfully . T is a positive integer, for example, T is equal to 10.

It is worth mentioning that after the number of successfully matched map points reaches T, the pose information of the terminal is calculated to avoid the insufficient number of successful matches and the inability to locate the resource waste caused by the positioning result.

Compared with the prior art, in the method for establishing a map provided in this embodiment, since information of map points in multiple maps is stored in a merged map, when the terminal performs operations such as expansion, deletion, and update of map point information, , Only need to operate the merged map, which improves the management efficiency of the map. In addition, when the terminal uses the merged map for positioning, it is not necessary to switch the map for matching, which improves the positioning efficiency. In addition, the terminal merges the description information of similar map points through a clustering algorithm, reducing the amount of map data.

A third embodiment of the present application relates to a terminal. As shown in FIG. 5, the terminal includes at least one processor 401; and a memory 402 that is communicatively connected to the at least one processor 401. The memory 402 stores instructions that can be executed by the at least one processor 401, and the instructions are executed by the at least one processor 401, so that the at least one processor 401 can execute the method for establishing a map described above.

In this embodiment, the processor 401 is a central processing unit (Central Processing Unit (CPU) as an example, the memory 402 is a readable and writable memory (Random Access Memory, RAM) as an example. The processor 401 and the memory 402 may be connected through a bus or other methods. In FIG. 5, the connection through the bus is taken as an example. The memory 402 is a non-volatile computer-readable storage medium, and can be used to store non-volatile software programs, non-volatile computer executable programs, and modules. For example, description information of map points in the embodiments of the present application is stored in In memory 402. The processor 401 executes various functional applications and data processing of the device by running the non-volatile software programs, instructions, and modules stored in the memory 402, that is, the above method for establishing a map is implemented.

The memory 402 may include a storage program area and a storage data area, where the storage program area may store an operating system and an application program required for at least one function; the storage data area may store a list of options and the like. In addition, the memory 402 may include a high-speed random access memory, and may further include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 402 may optionally include a memory remotely set relative to the processor, and these remote memories may be connected to an external device through a network. Examples of the above network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

One or more modules are stored in the memory, and when executed by one or more processors, execute the method for establishing a map in any of the foregoing method embodiments.

The above product can execute the method provided in the embodiment of the present application, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not described in this embodiment, refer to the method provided in the embodiment of the present application.

A fourth embodiment of the present application relates to a computer-readable storage medium storing a computer program. When the computer program is executed by the processor, the method for building a map described in any of the above method embodiments is implemented.

That is, those skilled in the art can understand that all or part of the steps in the method of the above embodiments can be implemented by a program instructing related hardware. The program is stored in a storage medium and includes several instructions to make a device ( It may be a single-chip microcomputer, a chip, or the like) or a processor that executes all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disks or optical disks and other media that can store program code .

Those of ordinary skill in the art can understand that the foregoing embodiments are specific embodiments for implementing the present application, and in practical applications, various changes can be made in form and details without departing from the spirit and range.

Claims

A method of building a map, including:

Get N maps describing the same space, where N is an integer greater than 1;

Extracting spatially invariant features of each of the N maps;

The information of the map points in the N maps are combined according to the spatially invariant characteristics of the N maps to obtain a merged map.
The method of claim 1, wherein the spatially invariant features include any one or any combination of line features, semantic features, and marker information;

Wherein, the line feature refers to a feature of a line segment in a map; the semantic feature refers to a feature assigned to the same map point in different maps; and the marker information refers to information of a positioning mark in space on a map .
The method for establishing a map according to claim 1 or 2, wherein the information of the map points in the N maps are combined according to the spatially invariant features of the N maps to obtain a merged map, This includes:

Selecting a map from the N maps as a reference map;

Matching spatially invariant features in the N-1 maps other than the reference map with the spatially invariant features in the reference map, respectively;

Determine the relative pose relationship between the N-1 map and the reference map according to the corresponding matching results of the N-1 maps;

According to the relative pose relationship between the N-1 maps and the reference map, information of map points of the N-1 maps is merged into the reference map to obtain the merged map.
The method for establishing a map according to any one of claims 1 to 3, wherein the information of the map points in the N maps are combined according to the spatially invariant features of the N maps to obtain a merge. After the map, the method for establishing a map further includes:

Merging map points with a distance less than a preset value in the merged map, where the position information of the merged map point is determined according to the position information of the map point before the merge.
The method for establishing a map according to claim 3, wherein the information of the map points includes description information of the map points;

Combining the information of the map points of the N-1 maps into the reference map according to the relative pose relationship between the N-1 maps and the reference map, to obtain the merged map, specifically include:

Determine the corresponding relationship between the map points in the N-1 map and the map points in the reference map according to the relative pose relationship between the N-1 maps and the reference map;

Determine the map points in the merged map according to the corresponding relationship between the map points in the N-1 maps and the map points in the reference map and the description information of the map points in the N-1 map A collection of descriptive information.
The method for building a map according to claim 5, wherein, according to the corresponding relationship between the map points in the N-1 maps and the map points in the reference map, and the N-1 maps After determining the description information of the map points in the map, after determining the set of description information of the map points in the merged map, the method for establishing a map further includes:

The clustering algorithm is used to cluster the description information in the description information set of the map points to obtain a clustered description information set.
The method for establishing a map according to claim 6, wherein the information of the map points further comprises: shooting information corresponding to the description information of the map points.
The method of claim 7, wherein the shooting information includes shooting brightness and shooting angle;

After the clustering algorithm is used to cluster the description information in the description information set of the map points to obtain the clustered description information set, the method for establishing a map further includes:

If the clustering algorithm aggregates L description information into one category, then according to the shooting brightness corresponding to the L description information, determine the shooting brightness corresponding to the description information after the L description information is clustered, Describe the shooting angles corresponding to the L description information, and determine the shooting angles corresponding to the description information clustered by the L description information;

Determine the shooting corresponding to each clustered description information in the clustered description information set according to the shooting brightness corresponding to the clustered description information and the shooting angle corresponding to the clustered description information information.
The method of claim 8, wherein the clustering is determined based on the shooting brightness corresponding to the clustered description information and the shooting angle corresponding to the clustered description information. After the following description information set, the method for establishing a map further includes:

Obtaining images for positioning;

Extracting description information of a map point in the image;

Determining description information in the merged map that matches description information of a map point in the image;

Obtaining position information of a map point corresponding to the matched description information from the merged map, and determining a positioning result according to the obtained position information.
The method for establishing a map according to claim 9, wherein the determining the description information in the merged map that matches the description information of a map point in the image specifically comprises:

Determining shooting information of the image;

Filtering and obtaining from the description information set of map points in the merged map for matching according to the shooting information of the image and each piece of description information in the description information set of map points in the merged map Descriptive information

From the description information for matching, description information that matches the description information of a map point in the image is determined.
The method for establishing a map according to claim 10, wherein the shooting information includes shooting brightness and shooting angle;

Filtering according to the shooting information of the image and each piece of description information in the description information set of the map points in the map, and filtering and obtaining from the description information set of the map points in the merged map for Match description information, including:

Calculating a first difference between the shooting brightness of the image and the shooting brightness corresponding to each description information, and a second difference between the shooting angle of the image and the shooting angle corresponding to each description information;

According to the first difference value and the second difference value, from the description information set of the map point, M description information is selected as the description information for matching; wherein M is a positive integer.
The method for establishing a map according to any one of claims 9 to 11, wherein after the determining the description information in the map that matches the description information of the map points in the image, the obtained Before the location information determines the positioning result, the positioning method further includes:

It is determined that at least T map points in the image match the map successfully, where T is a positive integer.
A terminal comprising at least one processor; and

A memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processing The device can execute the method for establishing a map according to any one of claims 1 to 12.
A computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the method for establishing a map according to any one of claims 1 to 12 is implemented.