Positioning initialization method, storage medium and electronic device

Abstract

The present disclosure relates to a positioning initialization method, a storage medium and an electronic device. The method comprises: determining a reference frame from a key frame in a point cloud database according to GNSS positioning information; matching the reference frame and a current point cloud frame; in response to determining that the matching result is a matching success, determining a target frame from the reference frame; and determining an initialization pose according to the target frame, the current point cloud frame and a point cloud registration algorithm. In this way, the GNSS positioning information can be used to preliminarily screen the key frames in the point cloud database, while ensuring the reliability of the screened reference frames, the number of image frames matched with the current point cloud frame can be reduced, the data processing amount can be reduced, and the efficiency of determining the initialization pose can be improved. The use of the current point cloud frame and the key frame previously stored in the point cloud database can improve the accuracy of the determined initialization pose, reduce the positioning error, and improve the robustness of the positioning system.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the field of positioning technology, in particular to a positioning initialization method, a storage medium and an electronic device. BACKGROUND

[0002] With the rapid development of artificial intelligence technology, automatic driving has changed people's way of travel, and plays an important role in logistics, catering, medical treatment and other fields, and has important significance for promoting intelligent travel. Positioning initialization is an important link in navigation positioning SLAM (Simultaneous Localization and Mapping), that is, when the sensor is initially located in a certain scene, it is necessary to determine its initial position in the current environment through positioning initialization to provide initial information for subsequent navigation positioning and ensure that the platform positioning works normally.

[0003] The current mainstream initialization positioning method, for example, the brute force search matching method based on ICP (Iterative Closest Point) / NDT (Normal Distribution Transform), relies on a key frame map. When the map size is large, the search time of this method is long, and it is difficult to quickly realize positioning initialization. SUMMARY

[0004] The purpose of the present disclosure is to provide a positioning initialization method, a storage medium and an electronic device to speed up the initialization and quickly and accurately realize positioning initialization.

[0005] To achieve the above purpose, the first aspect of the present disclosure provides a positioning initialization method, which comprises:

[0006] According to the GNSS positioning information, a reference frame is determined from the key frames in the point cloud database;

[0007] The reference frame and the current point cloud frame are matched;

[0008] In response to determining that the matching result is a matching success, a target frame is determined from the reference frame;

[0009] According to the target frame, the current point cloud frame and the point cloud registration algorithm, an initialization pose is determined.

[0010] Optionally, the reference frame is determined from the key frames in the point cloud database according to the GNSS positioning information, comprising:

[0011] The reference distance between the GNSS positioning information and the key pose corresponding to each key frame is determined;

[0012] The key frame corresponding to the minimum reference distance of the target number is determined as the reference frame.

[0013] Optionally, the matching the reference frame and the current point cloud frame comprises:

[0014] A first feature vector of each of the reference frames is determined.

[0015] A second feature vector of the current point cloud frame is determined.

[0016] A matching result is determined according to the first feature vector and the second feature vector.

[0017] Optionally, the determining the matching result according to the first feature vector and the second feature vector comprises:

[0018] A vector distance between the second feature vector and each of the first feature vectors is determined.

[0019] The matching result is determined according to the vector distance and a vector threshold.

[0020] Optionally, the determining the matching result according to the vector distance and a vector threshold comprises:

[0021] If the vector distance less than the vector threshold exists, the matching result is determined as matching success.

[0022] Optionally, the determining the target frame from the reference frames in response to the determination that the matching result is matching success comprises:

[0023] If the matching result is matching success, the reference frame corresponding to the minimum vector distance is determined as the target frame.

[0024] Optionally, the determining the matching result according to the vector distance and a vector threshold comprises:

[0025] If the vector distance less than the vector threshold does not exist, the matching result is determined as matching failure.

[0026] Optionally, the method further comprises:

[0027] performing particle filtering on the GNSS positioning information in response to the determination that the matching result is matching failure;

[0028] determining the initial pose according to a filtering result, the current point cloud frame and a point cloud registration algorithm.

[0029] A second aspect of the present disclosure provides a non-transitory computer readable storage medium having a computer program stored thereon, the program being executed by a processor to implement the steps of the method provided by the first aspect of the present disclosure.

[0030] The third aspect of the present disclosure provides an electronic device, comprising:

[0031] a memory having a computer program stored thereon;

[0032] a controller, wherein the computer program is executed by the controller to implement the steps of the method provided by the first aspect of the present disclosure.

[0033] In the above technical solution, according to the GNSS positioning information, the reference frame is determined from the key frames in the point cloud database; the reference frame and the current point cloud frame are matched; in response to determining that the matching result is a matching success, the target frame is determined from the reference frame; and the initialization pose is determined according to the target frame, the current point cloud frame and the point cloud registration algorithm. In this way, the GNSS positioning information can be used to preliminarily screen the key frames in the point cloud database, which can reduce the number of image frames matched with the current point cloud frame, reduce the data processing amount, shorten the time required to determine the target frame, and further improve the efficiency of determining the initialization pose while ensuring the reliability of the screened reference frame. The current point cloud frame and the key frames previously stored in the point cloud database can be used to improve the accuracy of the determined initialization pose, reduce the positioning error, and improve the robustness of the positioning system.

[0034] Other features and advantages of the present disclosure will be described in detail in the following detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0035] The accompanying drawings are included to provide a further understanding of the present disclosure, and constitute a part of the specification, and are used to explain the present disclosure together with the following detailed description, but do not constitute a limitation of the present disclosure. In the drawings:

[0036] Figure 1 is a flowchart of a positioning initialization method provided by an exemplary embodiment of the present disclosure.

[0037] Figure 2 is a flowchart of a positioning initialization method provided by an exemplary embodiment of the present disclosure.

[0038] Figure 3 is a flowchart of a positioning initialization method provided by an exemplary embodiment of the present disclosure.

[0039] Figure 4 is a block diagram of a positioning initialization device provided by an exemplary embodiment of the present disclosure.

[0040] Figure 5 is a flowchart of an electronic device provided by an exemplary embodiment of the present disclosure. DETAILED DESCRIPTION

[0041] The specific embodiments of the present disclosure are described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to illustrate and explain the present disclosure, and are not used to limit the present disclosure.

[0042] Figure 1 is a flowchart of a positioning initialization method provided by an exemplary embodiment of the present disclosure. The method can be applied to a controller of an electronic device such as a smart robot. As shown in Figure 1 , the method can include S101 to S104.

[0043] S101, determining a reference frame from a key frame in a point cloud database according to GNSS positioning information.

[0044] For example, a signal indicating latitude and longitude information provided by a GNSS (Global Navigation Satellite System) can be obtained. If the signal strength is less than a preset strength threshold, the signal can be determined as an invalid signal, and if the signal strength is not less than the preset strength threshold, the signal can be determined as a valid signal. In the case where the signal is determined to be valid, the latitude and longitude information indicated by the signal can be converted into a coordinate system corresponding to a point cloud map, i.e., GNSS positioning information can be obtained, wherein the coordinate system used can be MGRS (Military Grid Reference System).

[0045] For example, the point cloud database can be pre-constructed by algorithms such as LOAM, LIO-SAM, FAST-LIO, and the storage format of the data in the point cloud database can be PCD or ROSBAG format, and the corresponding loading method when calling the content in the point cloud database can be PCL or ROS loading. For example, the environment information can be collected by an environment collection device (such as a laser radar, a depth camera or other sensors) on the robot, so that the robot can determine and record the point cloud map and the pose information at the current time based on the collected environment information at every interval in the target scene, save the point cloud map at the current time as a key frame, each key frame corresponds to a key pose, until the robot finishes the target scene and completes the construction of the point cloud database.

[0046] For example, the association between each key frame in the point cloud database and the GNSS positioning information can be determined, and the n key frames with the greatest association can be determined as reference frames. The GNSS positioning information can be used to represent the position, and the key pose corresponding to the key frame can also be used to represent the position, so the distance between the position represented by the GNSS positioning information and the position represented by each key pose can be determined, the smaller the distance, the greater the association between the GNSS positioning information and the key frame corresponding to the key pose.

[0047] In this way, the key frames in the point cloud database can be preliminarily screened through GNSS positioning information. While ensuring the reliability of the screened reference frames, the number of image frames matched with the current point cloud frame can be reduced, the data processing volume can be reduced, and the efficiency of positioning initialization can be improved.

[0048] S102: Match the reference frame and the current point cloud frame.

[0049] For example, the current point cloud frame can be generated based on the lidar signal acquired at the current moment. Matching between the reference frame and the current point cloud frame can be achieved by comparing the similarity between the reference frame and the current point cloud frame. For example, the similarity between the reference frame and the current point cloud frame can be determined based on their feature vectors. If the similarity is greater than a similarity threshold, it can be determined that the current point cloud frame and the reference frame are matched successfully. If the similarity is less than or equal to the similarity threshold, it can be determined that the current point cloud frame and the reference frame are matched unsuccessfully.

[0050] S103 : In response to determining that the matching result is successful, determining a target frame from the reference frame.

[0051] For example, if the match result is successful, it can be determined that there is an image frame with high similarity to the current point cloud frame among the filtered reference frames, that is, there is a keyframe in the point cloud database that can be used to determine the initial pose. If there is a reference frame with a similarity greater than the similarity threshold, then this reference frame can be determined as the target frame. If there are multiple reference frames with similarities greater than the similarity threshold, the reference frame with the greatest similarity can be determined as the target frame.

[0052] S104: Determine an initialization pose based on the target frame, the current point cloud frame, and the point cloud registration algorithm.

[0053] For example, point cloud registration algorithms such as ICP and NDT can be used to process the target frame and the current point cloud frame to achieve point cloud registration and obtain an initial pose. This can improve the accuracy of the initial pose determined, reduce positioning errors, and enhance the robustness of the positioning system by utilizing the current point cloud frame and keyframes previously stored in the point cloud database.

[0054] In the above technical solution, a reference frame is determined from the key frames in the point cloud database based on GNSS positioning information; the reference frame and the current point cloud frame are matched; in response to determining that the matching result is a successful match, a target frame is determined from the reference frame; and the initialization pose is determined based on the target frame, the current point cloud frame, and the point cloud registration algorithm. In this way, the key frames in the point cloud database can be preliminarily screened using GNSS positioning information. While ensuring the reliability of the screened reference frames, the number of image frames that are matched with the current point cloud frame can be reduced, the amount of data processing can be reduced, the time required to determine the target frame can be shortened, and the efficiency of determining the initialization pose can be improved; using the current point cloud frame and the key frames previously stored in the point cloud database can improve the accuracy of the determined initialization pose, reduce positioning errors, and improve the robustness of the positioning system.

[0055] In an optional embodiment, in S101, determining a reference frame from key frames in a point cloud database based on GNSS positioning information may include:

[0056] Determine the reference distance between the GNSS positioning information and the key pose corresponding to each key frame;

[0057] The key frame corresponding to the minimum reference distance of the target number is determined as the reference frame.

[0058] For example, the GNSS positioning information can be represented by (gx, gy, gz), where gx is the data in the x-axis direction of the GNSS positioning information, gy is the data in the y-axis direction of the GNSS positioning information, and gz is the data in the z-axis direction of the GNSS positioning information. The key pose corresponding to the i-th key frame can be represented by (mk_x i ,mk_y i ,mk_z i ,mk_yaw i ,mk_pitch i ,mk_roll i ) indicates that, where mk_x i is the data of the i-th key frame in the x-axis direction, mk_y i is the data of the i-th key frame in the y-axis direction, mk_z i is the data of the i-th key frame in the z-axis direction, mk_yaw i is the yaw angle of the i-th key frame, mk_pitch i is the pitch angle of the i-th key frame, mk_roll i The reference distance D between the GNSS positioning information and the key pose corresponding to the i-th key frame can be determined by the following formula: i : The target number can be preset based on actual needs, for example, can be set to 2. If the point cloud database includes 5 key frames in total, the reference distances D1, D2, D3, D4 and D5 between the GNSS positioning information and each key frame can be determined respectively, and the 5 reference distances are sorted. If the 5 reference distances are sorted in descending order as D2, D3, D5, D4 and D1, the 2nd key frame and the 3rd key frame corresponding to D2 and D3 are determined as the reference frames. In this way, the key frames in the point cloud database can be preliminarily screened through the GNSS positioning information, ensuring that the reference frames screened have a certain relevance with the GNSS positioning information, that is, ensuring the reliability of the screened reference frames, while ensuring the accuracy of the initialized pose, and reducing the number of image frames matched with the current point cloud frame in the subsequent process.

[0059] In an optional embodiment, in S102, matching the reference frames and the current point cloud frame can include:

[0060] S1021, determining a first feature vector of each reference frame.

[0061] S1022, determining a second feature vector of the current point cloud frame.

[0062] S1023, determining a matching result according to the first feature vector and the second feature vector.

[0063] For example, the feature vector of the image frame can be extracted through a preset feature vector extraction algorithm. The feature vector extraction method can be any one of BOW algorithm, Scan Context algorithm, IRIS algorithm, and OVERLAP-Net algorithm. The feature vector extraction algorithm to be used can be determined according to different scene requirements. For example, if the current scene is a complex scene, the IRIS algorithm can be used, and if the current scene is a single scene or has a higher requirement for efficiency, the Scan Context algorithm can be used. The similarity between the current point cloud frame and each reference frame can be determined according to the second feature vector of the current point cloud frame and the first feature vector of each reference frame, and then the matching result is determined.

[0064] In S1023, the matching result is determined according to the first feature vector and the second feature vector, which can include:

[0065] determining a vector distance between the second feature vector and each first feature vector;

[0066] determining the matching result according to the vector distance and a vector threshold.

[0067] For example, the matching result can be determined in the following manner:

[0068] If there is a vector distance less than the vector threshold, it is determined that the matching result is a matching success.

[0069] If there is no vector distance less than the vector threshold, it is determined that the matching result is a matching failure.

[0070] In continuation of the above, if the determined reference frame is the second key frame and the third key frame in the point cloud database, the first feature vectors of the second key frame and the third key frame can be determined respectively, and the vector distance FD2 between the first feature vector of the second key frame and the second feature vector of the current point cloud frame and the vector distance FD3 between the first feature vector of the third key frame and the second feature vector of the current point cloud frame can be determined. The vector threshold T can be pre-set, and whether the similarity between the reference frame and the current point cloud frame meets the requirement can be determined by the vector threshold T.

[0071] If any one of FD2 and FD3 is less than the vector threshold T, it can be determined that there is an image frame in the reference frame that is highly similar to the current point cloud frame (i.e., meets the similarity requirement), i.e., there is an image frame in the reference frame that can be used to determine the initialization pose, at this time, it can be determined that the matching result is a matching success. On the contrary, if neither DF2 nor FD3 is less than the vector threshold T, it can be determined that there is no image frame in the reference frame that is highly similar to the current point cloud frame, i.e., there is no image frame in the reference frame that can be used to determine the initialization pose, at this time, it can be determined that the matching result is a matching failure.

[0072] In an optional embodiment, in S103, in response to determining that the matching result is a matching success, determining the target frame from the reference frames can include:

[0073] If the matching result is a matching success, the corresponding reference frame of the minimum vector distance is determined as the target frame.

[0074] For example, if the matching result is a matching success, it can be determined that there is at least one image frame in the reference frame whose corresponding vector distance is less than the vector threshold. At this time, the corresponding reference frame of the minimum vector distance is determined as the target frame, the reference frame that matches the current point cloud frame successfully and has the highest similarity with the current point cloud frame is determined as the target frame, and thus the accuracy of determining the initialization pose can be improved.

[0075] In an optional embodiment, the positioning initialization method provided by the present disclosure can further include:

[0076] In response to determining that the matching result is a matching failure, performing particle filtering on the GNSS positioning information;

[0077] Determining the initialization pose according to the filtering result, the current point cloud frame and the point cloud registration algorithm.

[0078] For example, the particle filter can include randomly generating m positions (gx+offse_x j ,gy+offse_y j ,gz+offse_z j ) near the position (gx,gy,gz) represented by the GNSS positioning information, where j can be in the range of 1 to m, offse_x j is the jth random quantity in the x-axis direction, offse_y j is the jth random quantity in the y-axis direction, and offse_z j is the jth random quantity in the z-axis direction. Then, based on the GNSS positioning information and the m randomly generated positions, a new image frame can be obtained, and the current point cloud frame and the newly obtained image frame can be processed by an algorithm such as ICP or NDT point cloud registration to achieve point cloud registration and obtain the initial pose.

[0079] The GNSS positioning information in the present disclosure can be used for reference frame selection on the one hand, and as an initial position for particle filtering when the reference frame and the current point cloud frame fail to match on the other hand. When the reference frame and the current point cloud frame fail to match, i.e., loop detection fails, the particle filter can be used to determine the initial pose, decouple loop detection and particle filtering, and increase the robustness of the positioning system.

[0080] Figure 3 is a flowchart of a positioning initialization method provided by an example embodiment of the present disclosure. Through the Figure 3 , the implementation process of the positioning initialization method provided by the present disclosure can be more clearly understood. As shown in Figure 3 , the method can include S301 to S313.

[0081] S301, acquiring GNSS positioning information.

[0082] S302, generating a current point cloud frame based on the laser radar signal acquired at the current time.

[0083] S303, determining the reference distance between the GNSS positioning information and the key pose corresponding to each key frame in the point cloud database.

[0084] S304, determining the key frame corresponding to the target number of minimum reference distances as the reference frame.

[0085] S305, determining the first feature vector of each reference frame.

[0086] S306, determining the second feature vector of the current point cloud frame.

[0087] S307, determining the vector distance between the second feature vector and each first feature vector.

[0088] S308, if there is a vector distance less than the vector threshold, determining that the matching result is a matching success.

[0089] S309, determining the corresponding reference frame of the minimum vector distance as a target frame.

[0090] S310, determining an initialization pose according to the target frame, the current point cloud frame and a point cloud registration algorithm.

[0091] S311, if there is no vector distance less than the vector threshold, determining that the matching result is a matching failure.

[0092] S312, performing particle filtering on the GNSS positioning information.

[0093] S313, determining an initialization pose according to the filtering result, the current point cloud frame and the point cloud registration algorithm.

[0094] In this way, the GNSS positioning information can be used to preliminarily screen the key frames in the point cloud database, while ensuring the reliability of the screened reference frames, the number of image frames matched with the current point cloud frame can be reduced, the data processing amount can be reduced, the time required to determine the target frame can be shortened, and the efficiency of determining the initialization pose can be improved; the current point cloud frame and the key frames previously stored in the point cloud database are used to improve the accuracy of the determined initialization pose, reduce the positioning error, and improve the robustness of the positioning system. When the reference frame and the current point cloud frame fail to match, i.e., the loop detection fails, the particle filtering on the GNSS positioning information can be performed to determine the initialization pose, so that the loop detection and the particle filtering can be decoupled, and the robustness of the positioning system can be improved.

[0095] Based on the same inventive concept, the disclosure also provides a positioning initialization device. Figure 4 is a block diagram of a positioning initialization device 400 provided by an example embodiment of the disclosure. Referring to Figure 4 , the positioning initialization device 400 can include:

[0096] A first determination module 401 is configured to determine a reference frame from key frames in a point cloud database according to GNSS positioning information.

[0097] A matching module 402 is configured to match the reference frame and a current point cloud frame.

[0098] A second determination module 403 is configured to determine a target frame from the reference frame in response to determining that the matching result is a matching success.

[0099] A first positioning module 404 is configured to determine an initialization pose according to the target frame, the current point cloud frame and a point cloud registration algorithm.

[0100] In the technical solution, the reference frame is determined from the key frames in the point cloud database according to the GNSS positioning information; the reference frame and the current point cloud frame are matched; in response to determining that the matching result is a matching success, the target frame is determined from the reference frame; and the initialization pose is determined according to the target frame, the current point cloud frame and the point cloud registration algorithm. In this way, the key frames in the point cloud database can be preliminarily screened through the GNSS positioning information, the reliability of the screened reference frame is ensured, the number of image frames matched with the current point cloud frame is reduced, the data processing amount is reduced, the time required to determine the target frame is shortened, and the efficiency of determining the initialization pose is improved; the current point cloud frame and the key frames previously stored in the point cloud database are used, the accuracy of the determined initialization pose is improved, the positioning error is reduced, and the robustness of the positioning system is improved.

[0101] Optionally, the first determining module 401 comprises:

[0102] a first determining sub-module, configured to determine a reference distance between the GNSS positioning information and a key pose corresponding to each key frame;

[0103] a second determining sub-module, configured to determine, as the reference frame, the key frame corresponding to the minimum reference distance of the target number.

[0104] Optionally, the matching module 402 can comprise:

[0105] a third determining sub-module, configured to determine a first feature vector of each reference frame;

[0106] a fourth determining sub-module, configured to determine a second feature vector of the current point cloud frame;

[0107] a fifth determining sub-module, configured to determine a matching result according to the first feature vector and the second feature vector.

[0108] Optionally, the fifth determining sub-module is configured to determine the matching result in the following manner:

[0109] If the vector distance is less than the vector threshold, it is determined that the matching result is a matching success.

[0110] Optionally, the second determining module 403 is configured to determine the target frame from the reference frame in the following manner:

[0111] If the matching result is a matching success, the reference frame corresponding to the minimum vector distance is determined as the target frame.

[0112] Optionally, the fifth determining sub-module is configured to determine the matching result in the following manner:

[0113] If the vector distance is not less than the vector threshold, determining that the matching result is a matching success.

[0114] Optionally, the positioning initialization apparatus 400 can further include:

[0115] a filtering module, configured to perform particle filtering on the GNSS positioning information in response to determining that the matching result is a matching failure.

[0116] a second positioning module, configured to determine the initialization pose according to the filtering result, the current point cloud frame, and a point cloud registration algorithm.

[0117] As to the apparatus in the above-described embodiments, the specific manners in which various modules perform operations have been described in detail in the embodiments of the method, and thus will not be described in detail here.

[0118] Figure 5 is a flowchart of an electronic device 700 provided by an example embodiment of the present disclosure. As shown in Figure 5 the electronic device 700 can include a processor 701 and a memory 702. The electronic device 700 can also include one or more of a multimedia component 703, an input / output (I / O) interface 704, and a communication component 705.

[0119] The processor 701 is configured to control overall operations of the electronic device 700 to complete all or part of the steps of the positioning initialization method described above. The memory 702 is configured to store various types of data to support operations of the electronic device 700, which can include, for example, instructions for any application or method operating on the electronic device 700, and application-related data, such as contact data, transmitted and received messages, pictures, audio, video, and the like. The memory 702 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic storage, a flash memory, a magnetic disk, or an optical disk. The multimedia component 703 can include a screen and an audio component. The screen can be, for example, a touch screen, and the audio component is configured to output and / or input audio signals. For example, the audio component can include a microphone configured to receive external audio signals. The received audio signals can be further stored in the memory 702 or transmitted through the communication component 705. The audio component also includes at least one speaker configured to output audio signals. The I / O interface 704 provides an interface between the processor 701 and other interface modules, which can be a keyboard, a mouse, a button, and the like. The buttons can be virtual buttons or physical buttons. The communication component 705 is configured to perform wired or wireless communication between the electronic device 700 and other devices. The wireless communication, such as Wi-Fi, Bluetooth, near field communication (NFC), 2G, 3G, 4G, NB-IOT, eMTC, or other 5G, and the like, or a combination of one or more of them, is not limited herein. Therefore, the corresponding communication component 705 can include a Wi-Fi module, a Bluetooth module, an NFC module, and the like.

[0120] In an exemplary embodiment, the electronic device 700 can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-mentioned positioning initialization method.

[0121] In another exemplary embodiment, a computer-readable storage medium including program instructions is further provided. When the program instructions are executed by a processor, the steps of the aforementioned positioning initialization method are implemented. For example, the computer-readable storage medium may be the aforementioned memory 702 including the program instructions. The program instructions may be executed by the processor 701 of the electronic device 700 to perform the aforementioned positioning initialization method.

[0122] In another exemplary embodiment, a computer program product is further provided. The computer program product includes a computer program executable by a programmable device, and has a code portion for executing the above positioning initialization method when executed by the programmable device.

[0123] The preferred embodiments of the present disclosure are described in detail above in conjunction with the accompanying drawings. However, the present disclosure is not limited to the specific details of the above embodiments. Within the technical concept of the present disclosure, various simple modifications can be made to the technical solutions of the present disclosure, and these simple modifications all fall within the scope of protection of the present disclosure.

[0124] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the present disclosure will not further describe various possible combinations.

[0125] In addition, the various embodiments of the present disclosure may be arbitrarily combined, and as long as they do not violate the concept of the present disclosure, they should also be regarded as the contents disclosed by the present disclosure.

Claims

1. A positioning initialization method, characterized in that: include: Based on the GNSS positioning information, the reference frame is determined from the key frames in the point cloud database; Matching the reference frame and the current point cloud frame; In response to determining that the matching result is a successful match, determining a target frame from the reference frame; Determining an initialization pose according to the target frame, the current point cloud frame, and a point cloud registration algorithm; In response to determining that the matching result is a matching failure, performing particle filtering on the GNSS positioning information; The initialization pose is determined according to the filtering result, the current point cloud frame and the point cloud registration algorithm.

2. The method according to claim 1, characterized in that Determining a reference frame from a key frame in a point cloud database according to the GNSS positioning information includes: Determining a reference distance between the GNSS positioning information and a key pose corresponding to each key frame; The key frame corresponding to the minimum reference distance of the target number is determined as the reference frame.

3. The method according to claim 1, characterized in that The matching of the reference frame and the current point cloud frame includes: determining a first eigenvector of each of the reference frames; Determining a second eigenvector of the current point cloud frame; A matching result is determined according to the first feature vector and the second feature vector.

4. The method according to claim 3, characterized in that Determining a matching result according to the first feature vector and the second feature vector includes: determining a vector distance between the second eigenvector and each of the first eigenvectors; A matching result is determined according to the vector distance and the vector threshold.

5. The method according to claim 4, characterized in that Determining a matching result according to the vector distance and the vector threshold includes: If there is a vector distance that is smaller than the vector threshold, the matching result is determined to be a successful match.

6. The method according to claim 4, characterized in that In response to determining that the matching result is a successful match, determining a target frame from the reference frame includes: If the matching result is a successful match, the reference frame corresponding to the minimum vector distance is determined as the target frame.

7. The method according to claim 4, characterized in that Determining a matching result according to the vector distance and the vector threshold includes: If there is no vector distance less than the vector threshold, the matching result is determined to be a matching failure.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

9. An electronic device, characterized in that: include: a memory having a computer program stored thereon; A processor, configured to execute the computer program in the memory to implement the steps of the method according to any one of claims 1 to 7.

Images