A method and device for initializing point cloud map

By calculating the bag-of-word vector similarity in the point cloud map to determine the initialization area, the problems of low initialization positioning efficiency and difficulty in maintenance in the point cloud map are solved, and efficient and accurate point cloud map initialization is achieved.

CN113129369BActive Publication Date: 2024-11-19BEIJING JINGDONG QIANSHITECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202010048424.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-01-16
Publication Date
2024-11-19
Estimated Expiration
2040-01-16

AI Technical Summary

Technical Problem

In the prior art, point cloud map initialization positioning is inefficient and difficult to maintain, violent point cloud matching is time-consuming, and memory location methods cannot solve the kidnapping problem in shutdown.

Method used

By obtaining the point cloud data of the current frame, generating the current image, calculating the word bag vector, and matching the similarity between the word bag vector of each sub-picture of the point cloud map and the current word bag vector, determining the initialization area, thereby obtaining the current frame Position.

Benefits of technology

It improves the efficiency and accuracy of initialization positioning in point cloud maps, avoids the waste of time of violent matching, solves the problem of post-maintenance of memory location methods, and realizes accurate initialization positioning in point cloud maps.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113129369B_ABST
    Figure CN113129369B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and device for initializing a point cloud map, and relates to the field of computer technology. A specific implementation of the method includes obtaining point cloud data of a current frame, generating a current image, and calculating a current bag-of-words vector; based on the bag-of-words vectors corresponding to each sub-image of the point cloud map, respectively calculating the similarity with the current bag-of-words vector to determine the sub-image as the initialization area, and then obtaining the pose of the current frame. Therefore, the implementation of the present invention can solve the problems of low efficiency and difficult maintenance of initialization positioning in existing point cloud maps.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a point cloud map initialization method and device. Background Art

[0002] In the field of autonomous driving, initial positioning in point cloud maps plays an extremely important role, such as the use of brute force point cloud matching methods, memory position methods, etc.

[0003] In the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art:

[0004] The current initialization positioning has some defects. The brute force point cloud matching method is time-consuming when the point cloud map range is large. The memory position method is simple and easy, but it cannot solve the kidnapping problem (i.e. moving the position when the power is turned off and initializing when the power is turned on), which is not convenient for later maintenance. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides a point cloud map initialization method and device, which can solve the problems of low efficiency and difficult maintenance of initialization positioning in existing point cloud maps.

[0006] To achieve the above-mentioned purpose, according to one aspect of an embodiment of the present invention, a point cloud map initialization method is provided, including obtaining point cloud data of a current frame, generating a current image, and calculating a current bag-of-words vector; based on the bag-of-words vectors corresponding to each sub-image of the point cloud map, respectively calculating the similarity with the current bag-of-words vector to determine the sub-image as the initialization area, and then obtaining the pose of the current frame.

[0007] Optionally, before respectively calculating the similarity with the current bag-of-words vector based on the bag-of-words vectors corresponding to each sub-graph of the point cloud map, the following steps are included:

[0008] A point cloud map is obtained to project all point cloud data onto a plane to generate a corresponding image, and then the image is divided into several sub-images, and the bag-of-words vector corresponding to each sub-image is calculated.

[0009] Optionally, after obtaining the point cloud map, include:

[0010] All point cloud data in the point cloud map are obtained to downsample the point cloud data.

[0011] Optionally, downsampling the point cloud data includes:

[0012] The geometric center of the point cloud within a preset range is calculated, the geometric center is used as the representative point of the range, and other point cloud data within the range is cleared.

[0013] Optionally, the image is divided into a plurality of sub-images, including:

[0014] The image is divided into a plurality of sub-images based on the resolution of the point cloud data of the current frame when the current image is generated.

[0015] Optionally, the bag-of-words vector corresponding to each subgraph is calculated, including:

[0016] Extract feature points from each sub-image and calculate the descriptor of each sub-image;

[0017] Based on the feature points of each subgraph, the corresponding bag-of-words vector is calculated through the descriptor.

[0018] Optionally, based on the feature points of each subgraph, the corresponding bag-of-words vector is calculated through the descriptor, including:

[0019] The ORB feature in the visual library is used to extract the feature points in each sub-image, and the descriptor consisting of 256 bits is calculated. Then, the DBoW2 library is used to calculate the bag-of-words vector uniquely corresponding to the sub-image from the descriptor.

[0020] Optionally, based on the bag-of-words vectors corresponding to each sub-graph of the point cloud map, similarities with the current bag-of-words vector are calculated respectively, including:

[0021] The sparse rule operator is used to calculate the similarity with the current bag-of-words vector based on the bag-of-words vector corresponding to each sub-graph of the point cloud map.

[0022] Optionally, determining a sub-image as an initialization area, and then obtaining the pose of the current frame, includes:

[0023] The point cloud data of the current frame is matched with the point cloud data corresponding to the sub-image as the initialization area through the nearest point iteration algorithm to iteratively solve the pose of the current frame.

[0024] In addition, the present invention also provides a point cloud map initialization device, including an acquisition module for acquiring point cloud data of the current frame, generating a current image, and calculating the current word bag vector; a processing module for calculating the similarity with the current word bag vector based on the word bag vectors corresponding to each sub-image of the point cloud map, so as to determine the sub-image as the initialization area, and then obtain the posture of the current frame.

[0025] An embodiment of the above invention has the following advantages or beneficial effects: because the point cloud data of the current frame is obtained to generate the current image to calculate the current word bag vector; based on the word bag vectors corresponding to each sub-image of the point cloud map, the similarity with the current word bag vector is calculated respectively to determine the sub-image as the initialization area, and then obtain the technical means of the posture of the current frame, so the existing technical problems of low efficiency and difficult maintenance of initialization positioning in point cloud maps are overcome, thereby achieving the technical effect of precise initialization positioning in point cloud maps.

[0026] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0028] Figure 1 is a schematic diagram of the main process of the point cloud map initialization method according to the first embodiment of the present invention;

[0029] Figure 2 is a schematic diagram of the main process of the point cloud map initialization method according to the second embodiment of the present invention;

[0030] Figure 3 is a schematic diagram of main modules of a point cloud map initialization device according to an embodiment of the present invention;

[0031] Figure 4 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0032] Figure 5 It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0034] Figure 1 is a schematic diagram of the main process of the point cloud map initialization method according to the first embodiment of the present invention. Figure 1 As shown, the point cloud map initialization method includes:

[0035] Step S101, obtaining the point cloud data of the current frame, generating the current image, and calculating the current bag-of-words vector.

[0036] In some embodiments, when calculating the current word bag vector, feature points may be extracted from the current image and descriptors may be calculated. The word bag vector is then calculated based on the feature points and the descriptors.

[0037] In a further embodiment, the ORB feature in OpenCV is used to extract feature points in the current image, and a descriptor consisting of 256 bits is calculated, and then the DBoW2 library is used to calculate the bag-of-words vector uniquely corresponding to the current image from the descriptor.

[0038] Step S102, based on the bag-of-words vectors corresponding to each sub-image of the point cloud map, the similarity with the current bag-of-words vector is calculated respectively to determine the sub-image as the initialization area, and then obtain the pose of the current frame.

[0039] In some embodiments, a sparse rule operator (ie, L1 norm) is used to calculate the similarity with the current bag-of-words vector based on the bag-of-words vector corresponding to each subgraph.

[0040] As some other embodiments, the point cloud data of the current frame is matched with the point cloud data corresponding to the sub-image as the initialization area through the closest point iteration algorithm (ie, ICP) to iteratively solve the pose of the current frame.

[0041] In some other embodiments, before executing step S102, a point cloud map may be obtained to project all point cloud data onto a plane to generate a corresponding image, and then the image may be segmented into a plurality of sub-images, and a bag-of-words vector corresponding to each sub-image may be calculated.

[0042] In some embodiments, after obtaining the point cloud map, it is necessary to obtain all point cloud data in the point cloud map to downsample the point cloud data. In a further embodiment, the geometric center of the point cloud within a preset range is calculated, the geometric center is used as a representative point of the range, and other point cloud data within the range is cleared, thereby downsampling the point cloud data, reducing a large number of invalid points (such as ground points), reducing the density of the point cloud map, and reducing the scale of calculation.

[0043] Preferably, a voxel filter may be used for downsampling to calculate the geometric center of the point cloud within a certain range (eg, 15 cm).

[0044] As some other embodiments, the image is divided into a plurality of sub-images based on the resolution of the current image generated by the point cloud data of the current frame. For example, if the resolution of the image generated by the point cloud map is 10000×10000, and the resolution of the current image generated by the point cloud data of the current frame is 100×100, then the image generated by the point cloud map is divided into 10000 sub-images according to 100×100.

[0045] As some other embodiments, when calculating the bag-of-words vector corresponding to each sub-graph, feature points may be extracted from each sub-graph, and descriptors of each sub-graph may be calculated. Then, based on the feature points of each sub-graph, the corresponding bag-of-words vector may be calculated through the descriptors.

[0046] In a further embodiment, the ORB feature (ORB feature is a feature point detection and description algorithm based on visual information) in OpenCV (OpenCV is a cross-platform computer vision library released under a BSD license) is used to extract feature points in each sub-graph, and a descriptor consisting of 256 bits is calculated, and then the DBoW2 library (DBoW2 is an open source software library) is used to calculate the bag-of-words vector uniquely corresponding to the sub-graph from the descriptor.

[0047] The specific implementation process includes: Assuming that 500 feature points are extracted from the subgraph, 500 descriptors can be generated. The DBoW2 library can provide a K-ary tree with N layers and M nodes in each layer, so there are a total of M N Each intermediate node has a corresponding key value, and each descriptor of each subgraph is compared with each intermediate node, and a deep search is performed until the corresponding leaf node is found. Then each subgraph can be uniquely described by the 01 vector composed of the presence or absence of the leaf node, which is the bag-of-words vector corresponding to the subgraph.

[0048] In summary, the point cloud map initialization method proposed in the present invention uses image similarity evaluation to screen out the area of ​​the current frame point cloud in the point cloud map, that is, converting the point cloud into an image, and then using the bag-of-words matching method of the image to screen out the most similar area, thereby avoiding the blind matching solution of the brute force matching method and wasting time and resources.

[0049] At the same time, the present invention can quickly and accurately determine the initialization area, thereby greatly reducing the solution space during brute force point cloud matching and improving the efficiency and accuracy of pose solving. Moreover, the present invention does not need to construct and store visual word bag information separately, so it does not take up too many computing resources and storage resources. In addition, the present invention is more versatile and does not have the later maintenance problem of the memory position method.

[0050] Figure 2 is a schematic diagram of the main process of the point cloud map initialization method according to the second embodiment of the present invention. The point cloud map initialization method may include:

[0051] Step S201, obtaining all point cloud data in the point cloud map to downsample the point cloud data.

[0052] Preferably, the geometric center of the point cloud within a preset range is calculated, the geometric center is used as a representative point of the range, and other point cloud data within the range is cleared.

[0053] Step S202: Project the point cloud data onto a plane to generate a corresponding bird's-eye view.

[0054] Preferably, all point cloud data are projected onto an xy plane at z=z1, thereby generating a bird's-eye view corresponding to the point cloud map, wherein z1 is a preset fixed value.

[0055] Step S203 , based on the resolution of the point cloud data of the current frame when generating the current bird's-eye view, divide the bird's-eye view into a plurality of sub-images.

[0056] Step S204, calculating the bag-of-words vector corresponding to each subgraph.

[0057] Preferably, feature points are extracted from each sub-graph, and a descriptor of each sub-graph is calculated. Then, based on the feature points of each sub-graph, a corresponding bag-of-words vector is calculated through the descriptor.

[0058] Step S205, obtaining the point cloud data of the current frame, generating the current bird's-eye view, and calculating the current bag-of-words vector.

[0059] Preferably, the point cloud data of the current frame is projected onto the xy plane of z=z1, thereby generating the current bird's-eye view, wherein z1 is a preset fixed value.

[0060] Step S206, based on the bag-of-words vector corresponding to each subgraph, respectively calculate the similarity with the current bag-of-words vector.

[0061] Preferably, a sparse rule operator (ie, L1 norm) is used to calculate the similarity with the current bag-of-words vector based on the bag-of-words vector corresponding to each subgraph.

[0062] Preferably, assuming that the bag-of-words vector of the bird's-eye view of the current frame is a, and the bag-of-words vector of the bird's-eye view of a sub-image is b, their L1 norms are defined as:

[0063]

[0064] Step S207: taking the sub-graph with the highest similarity as the initialization region.

[0065] Step S208, matching the point cloud data of the current frame with the point cloud data corresponding to the sub-image serving as the initialization area, so as to iteratively solve the pose of the current frame.

[0066] Preferably, the point cloud data of the current frame is matched with the point cloud data corresponding to the sub-image as the initialization area by using an iterative closest point algorithm ICP to iteratively solve the pose of the current frame.

[0067] It is worth noting that steps S201 to S204 can be executed simultaneously with step S205, or steps S201 to S204 can be executed first and then step S205, or step S205 can be executed first and then step S201 to S204.

[0068] Figure 3 is a schematic diagram of the main modules of the point cloud map initialization device according to an embodiment of the present invention, such as Figure 3 As shown, the point cloud map initialization device 300 includes an acquisition module 301 and a processing module 302. The acquisition module 301 acquires the point cloud data of the current frame and generates the current image to calculate the current word bag vector; the processing module 302 calculates the similarity with the current word bag vector based on the word bag vectors corresponding to each sub-image of the point cloud map, so as to determine the sub-image as the initialization area, and then obtain the posture of the current frame.

[0069] In a preferred embodiment, the acquisition module 301 is further used to: acquire a point cloud map to project all point cloud data onto a plane, generate a corresponding image, and then segment the image into a plurality of sub-images, and calculate a bag-of-words vector corresponding to each sub-image.

[0070] In some embodiments, after the acquisition module 301 acquires the point cloud map, it can acquire all point cloud data in the point cloud map to downsample the point cloud data.

[0071] In a further embodiment, the acquisition module 301 downsamples the point cloud data, including: calculating the geometric center of the point cloud within a preset range, taking the geometric center as a representative point of the range, and clearing other point cloud data within the range.

[0072] As some other embodiments, the acquisition module 301 divides the image into a plurality of sub-images, including: dividing the image into a plurality of sub-images based on the resolution when the current image is generated based on the point cloud data of the current frame.

[0073] As some further embodiments, the acquisition module 301 calculates the bag-of-words vector corresponding to each sub-graph, including: extracting feature points from each sub-graph and calculating the descriptor of each sub-graph; and calculating the corresponding bag-of-words vector through the descriptor based on the feature points of each sub-graph.

[0074] In a preferred embodiment, the acquisition module 301 calculates the corresponding bag-of-words vector through the descriptor based on the feature points of each subgraph, including:

[0075] The ORB feature in the visual library is used to extract the feature points in each sub-image, and the descriptor consisting of 256 bits is calculated. Then, the DBoW2 library is used to calculate the bag-of-words vector uniquely corresponding to the sub-image from the descriptor.

[0076] In some other embodiments, the processing module 302 uses a sparse rule operator to calculate the similarity with the current bag-of-words vector based on the bag-of-words vectors corresponding to each sub-graph of the point cloud map.

[0077] As some further embodiments, the processing module 302 determines the sub-image as the initialization area, and then obtains the pose of the current frame, including: matching the point cloud data of the current frame with the point cloud data corresponding to the sub-image as the initialization area through a nearest point iteration algorithm to iteratively solve the pose of the current frame.

[0078] It should be noted that the point cloud map initialization method and the point cloud map initialization device of the present invention have a corresponding relationship in terms of specific implementation contents, so the repeated contents will not be described again.

[0079] Figure 4 An exemplary system architecture 400 is shown to which a point cloud map initialization method or a point cloud map initialization device according to an embodiment of the present invention may be applied.

[0080] like Figure 4 As shown, system architecture 400 may include terminal devices 401, 402, 403, network 404 and server 405. Network 404 is used to provide a medium for communication links between terminal devices 401, 402, 403 and server 405. Network 404 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0081] Users can use terminal devices 401, 402, 403 to interact with server 405 through network 404 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 401, 402, 403, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only as examples).

[0082] The terminal devices 401 , 402 , and 403 may be various electronic devices having a point cloud map initialization screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0083] The server 405 may be a server that provides various services, such as a backend management server (only an example) that provides support for shopping websites browsed by users using the terminal devices 401, 402, and 403. The backend management server may analyze and process the received data such as product information query requests, and feed back the processing results (such as target push information, product information - only an example) to the terminal device.

[0084] It should be noted that the point cloud map initialization method provided in the embodiment of the present invention is generally executed by the server 405 , and accordingly, the computing device is generally arranged in the server 405 .

[0085] It should be understood that Figure 4The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0086] Reference below Figure 5 , which shows a schematic diagram of the structure of a computer system 500 of a terminal device suitable for implementing an embodiment of the present invention. Figure 5 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0087] like Figure 5 As shown, the computer system 500 includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage part 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the computer system 500 are also stored. The CPU 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0088] The following components are connected to the I / O interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a cathode ray tube (CRT), a liquid crystal point cloud map initializer (LCD), etc., and a speaker; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed, so that a computer program read therefrom is installed into the storage section 508 as needed.

[0089] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or installed from the removable medium 511. When the computer program is executed by the central processing unit (CPU) 501, the above-mentioned functions defined in the system of the present invention are executed.

[0090] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0091] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0092] The modules involved in the embodiments of the present invention may be implemented by software or hardware. The modules described may also be set in a processor. For example, it may be described as follows: a processor includes an acquisition module and a processing module. The names of these modules do not, in some cases, limit the modules themselves.

[0093] As another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiment; or it may exist independently without being assembled into the device. The above computer-readable medium carries one or more programs. When the above one or more programs are executed by a device, the device includes obtaining point cloud data of the current frame, generating a current image, and calculating the current word bag vector; based on the word bag vectors corresponding to each sub-image of the point cloud map, respectively calculating the similarity with the current word bag vector to determine the sub-image as the initialization area, and then obtaining the pose of the current frame.

[0094] According to the technical solution of the embodiment of the present invention, it is possible to solve the problems of low efficiency and difficult maintenance of initialization positioning in the existing point cloud map.

[0095] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A point cloud map initialization method, characterized in that: include: Get the point cloud data of the current frame and project it onto the plane to generate the current image. Extract feature points from the current image and calculate the descriptor to calculate the current bag-of-words vector. Based on the bag-of-words vectors corresponding to each sub-image of the point cloud map, the method includes: dividing the current image into a plurality of sub-images, extracting feature points from each sub-image, and calculating a descriptor of each sub-image; Based on the feature points of each subgraph, the corresponding bag-of-words vector is calculated through the descriptor; The similarity with the current bag-of-words vector is calculated respectively to determine the sub-graph as the initialization area, and then the pose of the current frame is obtained.

2. The method according to claim 1, characterized in that Based on the bag-of-words vectors corresponding to each sub-graph of the point cloud map, before calculating the similarity with the current bag-of-words vector, it includes: A point cloud map is obtained to project all point cloud data onto a plane to generate a corresponding image, and then the image is divided into several sub-images, and the bag-of-words vector corresponding to each sub-image is calculated.

3. The method according to claim 2, characterized in that After obtaining the point cloud map, including: All point cloud data in the point cloud map are obtained to downsample the point cloud data.

4. The method according to claim 3, characterized in that The point cloud data is downsampled, including: The geometric center of the point cloud within a preset range is calculated, the geometric center is used as the representative point of the range, and other point cloud data within the range is cleared.

5. The method according to claim 2, characterized in that: The image is divided into several sub-images, including: The image is divided into a plurality of sub-images based on the resolution of the point cloud data of the current frame when the current image is generated.

6. The method according to claim 1, characterized in that Based on the feature points of each subgraph, the corresponding bag-of-words vector is calculated through the descriptor, including: The ORB feature in the visual library is used to extract the feature points in each sub-image, and the descriptor consisting of 256 bits is calculated. Then, the DBoW2 library is used to calculate the bag-of-words vector uniquely corresponding to the sub-image from the descriptor.

7. The method according to claim 1, characterized in that Based on the bag-of-words vectors corresponding to each sub-graph of the point cloud map, the similarity with the current bag-of-words vector is calculated respectively, including: The sparse rule operator is used to calculate the similarity with the current bag-of-words vector based on the bag-of-words vector corresponding to each sub-graph of the point cloud map.

8. The method according to any one of claims 1 to 7, characterized in that: Determine the sub-image as the initialization area, and then get the pose of the current frame, including: The point cloud data of the current frame is matched with the point cloud data corresponding to the sub-image as the initialization area through the nearest point iteration algorithm to iteratively solve the pose of the current frame.

9. A point cloud map initialization device, characterized in that: include: The acquisition module is used to acquire the point cloud data of the current frame and project it onto a plane to generate the current image, extract feature points from the current image and calculate the descriptor to calculate the current bag-of-words vector; A processing module is used for dividing the current image into a plurality of sub-images, extracting feature points from each sub-image, and calculating a descriptor of each sub-image based on the bag-of-words vectors corresponding to each sub-image of the point cloud map; Based on the feature points of each sub-image, the corresponding bag-of-words vector is calculated through the descriptor; the similarity with the current bag-of-words vector is calculated respectively to determine the sub-image as the initialization area, and then the pose of the current frame is obtained.

10. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 8.

11. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.