A map online updating method and device, a storage medium and a terminal device
By acquiring the difference data between image data and historical image data, the map data of the SLAM system is updated, which solves the problem of map data not being updated in a timely manner and improves positioning accuracy and application scenarios.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-09-13
- Publication Date
- 2026-05-15
AI Technical Summary
The map data of existing SLAM systems cannot be updated in a timely manner, which affects the positioning accuracy or makes positioning impossible, especially when the position or orientation of objects in the scene changes.
By acquiring historical image data that matches the image data with pre-stored map data, the differences between the two are determined, and the map data is updated based on these differences.
It enables timely updates of map data, improves positioning accuracy, expands application scenarios, and provides convenience without requiring additional user intervention.
Smart Images
Figure CN115638783B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of map updating technology, and in particular to a method, apparatus, storage medium, and terminal device for online map updating. Background Technology
[0002] SLAM (Simultaneous Localization and Mapping), a fundamental technology in early mobile robotics, allows systems to capture environmental information to build maps and simultaneously use this information to calculate the device's current position and orientation. Therefore, it has been widely adopted in many new applications. For example, maps built using SLAM systems can be used for 3D reconstruction. Furthermore, using pre-built maps, the system can calculate the device's current position and orientation, providing positioning and navigation capabilities for systems and users in AR / VR, mobile robotics, and autonomous driving applications. Compared to other traditional positioning technologies such as GPS, SLAM systems can calculate not only the device's position but also its 3D orientation. Moreover, SLAM systems are less restricted by the environment and can be used both indoors and outdoors.
[0003] While SLAM technology boasts a series of advantages, several current drawbacks significantly limit its application scenarios. One such issue is the online autonomous updating of maps. Traditional SLAM systems typically assume a static scene during map construction, meaning that objects in the scene will not change position or orientation during or after map creation. However, this assumption is often not met in real-world applications. Recent research has shown that some systems use object detection techniques to remove common dynamic objects like people and animals during the map building phase, but fail to update map elements during the localization phase. When infrequently moving objects, such as tables or potted plants, change position or orientation, the lack of timely map updates affects the localization accuracy of the SLAM system, potentially leading to a failure to locate the target.
[0004] Therefore, existing technologies still need to be improved and enhanced. Summary of the Invention
[0005] The technical problem to be solved by the present invention is to provide a method, apparatus, storage medium and terminal device for online map updating, in order to address the above-mentioned defects of the prior art. The aim is to solve the problem that the map data in the prior art is not updated in a timely manner, which affects the positioning accuracy or even makes positioning impossible.
[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:
[0007] In a first aspect, the present invention provides a method for online map updating, wherein the method includes: acquiring image data and acquiring historical image data matching the image data from pre-stored map data; determining difference data between the image data and the historical image data based on the image data and the historical image data, wherein the difference data is used to reflect the differences between objects in the image data and the historical image data; and updating the map data based on the difference data.
[0008] Secondly, embodiments of the present invention provide an online map update device, wherein the device includes: an image matching module, used to acquire image data and acquire historical image data that matches the image data from map data; a difference analysis module, used to determine the difference data between the image data and the historical image data based on the image data and the historical image data, the difference data being used to reflect the differences between objects in the image data and the historical image data; and a map update module, used to update the map data based on the difference data.
[0009] Thirdly, embodiments of the present invention also provide a terminal device, which includes a data acquisition device, a memory, and a processor; wherein the data acquisition device is used to acquire image data of a preset target scene; the memory is used to pre-store map data; and the processor is used to acquire image data and process the image data according to the online map update method provided in this application to update the pre-stored map data.
[0010] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing a map online update program, which, when executed by a processor, implements the steps of the map online update method as described in any of the above solutions.
[0011] Beneficial Effects: Compared with existing technologies, this invention provides an online map updating method. First, it acquires image data and then retrieves historical image data matching the image data from the map data. Next, based on the image data and historical image data, it determines the differences between the two, reflecting the differences between objects in the image data and the historical image data. Finally, it updates the map data based on these differences. This invention analyzes the currently acquired image data and corresponding historical image data, and updates the map data promptly based on the differences between them. The entire process requires no additional user intervention, providing great convenience and expanding the application scenarios of map data. The updated map data also facilitates accurate positioning. Attached Figure Description
[0012] Figure 1A flowchart illustrating a specific implementation of the online map update method provided in this embodiment of the invention.
[0013] Figure 2 This is a schematic diagram of the online map update device provided in an embodiment of the present invention.
[0014] Figure 3 A schematic diagram of a terminal device provided in an embodiment of the present invention. Detailed Implementation
[0015] To make the objectives, technical solutions, and effects of this invention clearer and more explicit, the invention 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 merely illustrative of the invention and are not intended to limit the invention.
[0016] This embodiment provides a method for online map updates. Based on this method, map data can be updated in a timely manner to achieve accurate positioning. Specifically, this embodiment first acquires image data and then retrieves historical image data matching the image data from the map data. Next, based on the image data and historical image data, the differences between the image data and historical image data are determined. These differences reflect the differences between objects in the image data and historical image data. Finally, the map data is updated based on these differences. This embodiment analyzes the currently acquired image data and corresponding historical image data, and updates the map data in a timely manner based on the differences between them. The entire process requires no additional user intervention, providing great convenience and expanding the application scenarios of map data. The updated map data helps achieve accurate positioning.
[0017] For example, a user uses a mobile phone to acquire image data of an indoor scene. Assume the image data contains objects such as a table, two chairs, and a sofa, and there is a certain relative positional relationship between them. Then, the map data can be matched based on this image data to retrieve corresponding historical image data—images of the same indoor scene that were previously captured. After obtaining this historical image data, this embodiment compares and analyzes the newly acquired image data with the historical image data to determine the differences between them. Specifically, it determines whether the historical image data also contains the same table, two chairs, and a sofa, and whether the relative positional relationship between them is the same as in the image data. Once the difference data is obtained, this embodiment can update the map data based on this difference data. Since the difference data reflects changes in objects between the image data and the historical image data, when changes occur, the map data needs to be updated to obtain new map data for more accurate positioning.
[0018] Exemplary methods
[0019] The online map update method in this embodiment can be applied to terminal devices, such as mobile phones, computers, robots, handheld scanners, and other intelligent products. It can even be applied to monitoring equipment on various roads. Figure 1 As shown in the figure, the online map update method in this embodiment specifically includes the following steps:
[0020] Step S100: Obtain image data of the target scene and retrieve historical image data that matches the image data from the pre-stored map data.
[0021] This embodiment first acquires image data, which is image data of a specific target scene. After obtaining the image data, this embodiment matches the image data with map data, and retrieves historical image data that matches the image data from the map data. This embodiment needs to update the pre-stored map data, which consists of corresponding image data collected for each target scene during creation, and can also be understood as historical image data. Therefore, when the image data is re-acquired, this embodiment matches historical image data of the same target scene from the pre-stored map data, which is beneficial for analyzing the image data and historical image data in subsequent steps to determine whether the map data needs to be updated.
[0022] In one implementation, this embodiment includes the following steps when performing image data matching:
[0023] Step S101: Acquire image data of the preset target scene;
[0024] Step S102: Perform feature recognition on the image data to obtain image features in the image data, wherein the image features include: image feature points corresponding to objects in the image data or geometric shape features corresponding to objects in the image data;
[0025] Step S103: Based on image features, perform feature matching in map data to obtain historical image data that matches the image features.
[0026] In practical implementation, this embodiment can acquire images for a preset target scene. For example, when the map data in this embodiment is used for indoor positioning, images can be acquired for the indoor scene to obtain image data of that indoor scene. The image data might contain a table, two chairs, and a sofa, with a certain relative positional relationship between them. After obtaining the image data, this embodiment analyzes the image data, performs feature recognition, and identifies the image features within the image data. In practical application, the image data acquired in this embodiment is for a preset target scene; therefore, the image data will definitely contain objects, such as people, buildings, roads, and other objects.
[0027] To determine whether and how to update the map data in subsequent steps, this embodiment requires identifying all objects in the image data. In this embodiment, objects in the image data can be identified through image features. Therefore, the image features in this embodiment include: image feature points corresponding to objects in the image data or geometric shape features corresponding to objects in the image data. Image feature points are pixels in the image data. Based on the image feature points and geometric shape features, the object, its shape, and location can be determined. Then, based on the identified image features, feature matching is performed in the map data to obtain historical image data that matches the image features. After obtaining image features from the currently acquired image data, this embodiment can perform feature matching with previously acquired scene images in the map data to determine the most matching scene image, i.e., to determine the historical image data.
[0028] In one implementation, the timing of image data acquisition for the preset target scene in this embodiment can be set during the map reconstruction phase of the SLAM system. This embodiment sets the timing of map reconstruction for the SLAM system; for example, when the map is activated, it indicates that the user needs to use map data for positioning, requiring map data updates. Therefore, this embodiment requires image data acquisition for the preset target scene. After obtaining image features through feature recognition of the image data, this embodiment can determine the object corresponding to the image feature based on the image feature and set the image features corresponding to the same object as a feature group. For example, the objects in the image data may include a table, two chairs, and a sofa. Each object is composed of multiple image features. Therefore, this embodiment can set the image features corresponding to the same object as a feature group. For example, all image features corresponding to the table can be set as feature group 1, all image features corresponding to the first chair as feature group 2, all image features corresponding to the second chair as feature group 3, and all image features corresponding to the sofa as feature group 4. This facilitates the matching of historical image data from the map data. When matching with map data, this embodiment can match all scene images in the map data based on these identified feature groups. Since some scene images may also contain the same tables or chairs, this embodiment selects the scene image with the highest similarity to all feature groups in the image data based on the principle of highest similarity, and then uses the scene image as historical image data.
[0029] In another implementation, after classifying and grouping image features to obtain feature groups, this embodiment further determines the adjacent association groups based on the k nearest neighbors algorithm. That is, after recognizing the image data, all image features of the table are obtained, forming feature group 1. At this point, the K neighboring feature groups of feature group 1 can be found. These feature groups are beneficial for accurately updating the map data in subsequent updates.
[0030] Step S200: Based on the image data and historical image data, determine the difference data between the image data and historical image data. The difference data is used to reflect the differences between objects in the image data and historical image data.
[0031] After obtaining the image data and historical image data, this embodiment compares and analyzes the image data and historical image data to determine the difference data between the image data and historical image data. This difference data is used to reflect the differences between the object features in the image data and historical image data, that is, to reflect the changes in the objects in the two image data. Only in this way can it be determined which objects in the current preset target scene have changed and how they have changed, so as to facilitate the updating of map data in subsequent steps.
[0032] In one implementation, this embodiment includes the following steps when analyzing image data and historical image data:
[0033] Step S201: Based on the feature group, determine the group difference data of the feature group in the image data and historical image data, wherein the group difference data is used to reflect the change of the object corresponding to the feature group between the image data and historical image data;
[0034] Step S202: Treat the group difference data as variance data.
[0035] Specifically, since this embodiment groups the image features in the image data into feature groups, it can determine the changes of objects in the image data and historical image data based on these feature groups. Specifically, for objects in a preset target scene, only the object's position may have changed (e.g., the placement of a table has changed). In this case, the map data also needs to be updated. Both the image data and historical image data contain feature groups corresponding to the table, but these feature groups differ. Therefore, this embodiment can determine the group difference data of the feature group corresponding to the table in the image data and historical image data. The group difference data reflects the changes of the table between the image data and historical image data. After calculating the group difference data, this embodiment can use this group difference data as the difference data. Specifically, the group difference in this embodiment can be the reprojection error of image feature points or the geometric shape matching difference. The calculation method for the group difference is as follows:
[0036]
[0037] Where n is the total number of features in the feature group, x i Let y be the i-th feature of this feature group in the image data. i For this feature group in historical image data, it is related to x i Matching features.
[0038] In another implementation, after grouping image features in the image data into feature groups, two scenarios may occur. First, a feature group for a certain object may exist only in the current image data but not in the historical image data. This indicates that the object corresponding to that feature group has been added to the current target scene. For example, a vase has been added to the table in the current target scene, but it doesn't exist in the historical image data. This situation also falls under the category of discrepancies between the current image data and historical image data, and the discrepancy is the feature group itself, which exists only in the current image data. Second, a feature group for a certain object may exist only in the historical image data but not in the current image data. This indicates that the object corresponding to that feature group has been removed from the current target scene. For example, a table has been removed from the current target scene, so it doesn't exist in the current image data but does exist in the historical image data. This situation also falls under the category of discrepancies between the current image data and historical image data, and the discrepancy is the feature group itself, which exists only in the current historical image data.
[0039] Step S300: Update the map data based on the discrepancy data.
[0040] Since the difference data in this embodiment reflects the differences between objects in the image data and historical image data, once the difference data between the image data and historical image data is obtained, it is possible to know which objects have changed in the image data relative to the historical image data, and what specific changes have occurred. This allows for timely updates to the map data to achieve accurate positioning.
[0041] In one implementation, this embodiment includes the following steps when updating the map:
[0042] Step S301: Compare the group difference data with the preset threshold;
[0043] Step S302: If the difference data of the groups is greater than the preset threshold, the image features at the positions corresponding to the feature groups in the historical image data are deleted, and the image features corresponding to the feature groups in the image data are written into the corresponding positions in the historical image data.
[0044] Since group difference data is used to reflect the changes of the objects corresponding to the feature group between the image data and historical image data, this embodiment compares the group difference data with a preset threshold after obtaining the group difference data. The preset threshold is used to measure the degree of change of the objects corresponding to the feature group between the image data and historical image data. If the group difference data is greater than the preset threshold, it indicates that the objects corresponding to the feature group have changed significantly between the image data and historical image data, and the map data needs to be updated.
[0045] Specifically, this embodiment can delete the image features at the locations corresponding to feature groups in historical image data and write the image features corresponding to the feature groups in the image data into the corresponding locations in the historical image data, thereby completing the map data update. In one implementation, the preset threshold in this embodiment can be set to 30% of the average value of the group difference data corresponding to all feature groups, which facilitates a more accurate determination of whether the map data needs to be updated. It should be noted that this embodiment can also delete all image features in the historical image data and write all image features in the image data, but this method involves a large amount of computation and is inefficient.
[0046] In another implementation, if the discrepancy data indicates that the object corresponding to the feature group exists only in the image data, it means that the object corresponding to the feature group was newly added to the preset target scene. In this embodiment, the object corresponding to the associated group is matched in the historical image data. If the match is successful, it can be determined that the object has a relative positional relationship with the previous object, further confirming that the object was newly added to the preset target scene, and also determining the position of the newly added object. Therefore, in this embodiment, the historical image data corresponding to the feature group position in the historical image data can be deleted, and the image features corresponding to the feature group in the image data can be added to the historical image data to complete the map data update.
[0047] In other implementations, if the discrepancy data indicates that the object corresponding to the feature group only exists in historical image data, then the group lifecycle of the object corresponding to the feature group is obtained. In this embodiment, the group lifecycle is used to reflect the number of times the feature group has not been detected in the image data. If the group lifecycle has been exhausted, it means that the object corresponding to the feature group has been removed from the preset target scene. Therefore, according to this embodiment, the image features corresponding to the feature group can be deleted from the historical images.
[0048] Specifically, the only object present in the historical image data is the table. The initial lifespan of the table is 3. When the feature group corresponding to the table is not present in the image data (i.e., the table is not detected), the table's lifespan is reduced by 1. When the table's lifespan reaches 0 (meaning its lifespan has expired, i.e., the number of times the table has not been detected in the image data has been used up), it is confirmed that the table is no longer in the preset target scene, and therefore the table is deleted from the map data. Furthermore, if the table was not detected twice in previous image data, its lifespan was 1, but it is detected again in the current image, then the table's lifespan will return to the initial value of 3. In practical applications, an object may be in the preset target scene but not captured in the current image data; however, the object may still exist. Therefore, this embodiment sets a lifespan to avoid accidental deletion. If the lifespan becomes 0, it means the object is truly no longer in the preset target scene, and at this point, the object is deleted from the map data, thus ensuring the accuracy of map data updates.
[0049] In summary, this embodiment first acquires image data and then retrieves historical image data matching the image data from the map data. Next, based on the image data and the historical image data, it determines the differences between the two, reflecting the differences between objects in the image data and the historical image data. Finally, it updates the map data based on these differences. This embodiment analyzes the currently acquired image data and its corresponding historical image data, and updates the map data promptly based on the differences between them. The entire process requires no additional user intervention, providing great convenience and expanding the application scenarios of map data. The updated map data also facilitates accurate positioning.
[0050] Exemplary device
[0051] Based on the above embodiments, the present invention also provides an online map updating device, such as... Figure 2 As shown, the device includes an image matching module 10, a difference analysis module 20, and a map update module 30. Specifically, the image matching module 10 is used to acquire image data containing a preset target scene and to acquire historical image data that matches the image data from the map data; the difference analysis module 20 is used to determine the difference data between the image data and the historical image data based on the image data and the historical image data, and the difference data is used to reflect the differences between objects in the image data and the historical image data; the map update module 30 is used to update the map data based on the difference data.
[0052] In one implementation, the image matching module 10 includes: an image acquisition unit for acquiring image data of a preset target scene; a feature recognition unit for performing feature recognition on the image data to obtain image features in the image data, wherein the image features include: image feature points corresponding to objects in the image data or geometric shape features corresponding to objects in the image data; and a feature matching unit for performing feature matching in map data based on the image features to obtain historical image data that matches the image features.
[0053] In one implementation, the feature matching unit includes: a feature grouping subunit, used to determine the object corresponding to the image feature based on the image feature, and set the image features corresponding to the same object as a feature group; and an image filtering subunit, used to filter out the scene image with the highest similarity to all feature groups in the map data based on all feature groups, and use the scene image as historical image data.
[0054] In one implementation, the feature matching unit includes an association group determination subunit, which is used to determine the association group adjacent to the feature group based on the nearest neighbor algorithm.
[0055] In one implementation, the difference analysis module 20 includes: a group difference data determination unit, used to determine the group difference data of the feature group in the image data and historical image data based on the feature group, wherein the group difference data is used to reflect the change of the object corresponding to the feature group between the image data and historical image data; and a difference data determination unit, used to use the group difference data as difference data.
[0056] In one implementation, the map update module 30 includes: a data comparison unit for comparing group difference data with a preset threshold; and a first map update unit for deleting image features at the locations corresponding to feature groups in historical image data and writing image features corresponding to feature groups in image data if the group difference data is greater than the preset threshold.
[0057] In one implementation, the map update module 30 includes: an association matching unit, used to match the object corresponding to the association group in the historical image data if the object corresponding to the feature group exists only in the image data; and a second map update unit, used to add the image features corresponding to the feature group to the historical image data if the matching is successful.
[0058] In one implementation, the map update module 30 includes: a lifecycle determination unit, used to obtain the group lifecycle of the object corresponding to the feature group if the differential data is that the object corresponding to the feature group only exists in the historical image data, the group lifecycle being used to reflect the number of times the feature group has not been detected in the image data; and a third map update unit, used to delete the image features corresponding to the feature group in the historical image if the group lifecycle has been exhausted.
[0059] The working principle of each module in the online map update device of this embodiment is the same as that of each step in the above method embodiment, and will not be repeated here.
[0060] Based on the above embodiments, the present invention also provides a terminal device, the schematic diagram of which can be as follows: Figure 3 As shown, the terminal device includes a data acquisition unit, a processor, and a memory. The data acquisition unit is used to acquire image data of a preset target scene and transmit it to the processor. The processor is used to acquire the image data and update the pre-stored map data according to the online map update method provided in the above embodiment. The memory is used to pre-store the map data of the preset target scene. It should be noted that the data acquisition unit and the processor can communicate via wired or wireless means, which is not limited here.
[0061] In one embodiment, the acquisition device includes any one of a depth camera, an RGB camera, or a laser scanner. The depth camera includes any one of a structured light camera, a TOF camera, a binocular camera, or a multi-view camera; no limitation is made here. It should be noted that when the acquisition device is a depth camera, the acquired image data includes depth images or point cloud data; when the acquisition device is an RGB camera, the acquired image data includes RGB images; and when the acquisition device is a laser scanner, the acquired image data includes point cloud data. Furthermore, the corresponding pre-stored map data is the same as the image data acquired by the acquisition device.
[0062] In one embodiment, the processor is also used to control the operation of the data acquisition device. It can be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, neural network chips, digital processing chips, graphics processors, and various control chips. The processor is the control unit of the terminal device, connecting various components of the terminal device through various interfaces and lines. It executes programs or modules (e.g., a program for an online map update method) and calls data to perform various functions of the terminal device and process data.
[0063] In one embodiment, the memory is further used to store a map online update method program that can run on the processor. When the processor executes the map online update method program, it implements the following operation instructions: acquiring image data and acquiring historical image data that matches the image data from the map data; determining the difference data between the image data and the historical image data based on the image data and the historical image data, the difference data being used to reflect the differences between objects in the image data and the historical image data; and updating the map data based on the difference data. It should be noted that when the processor is a neural network processing chip, the memory may not include the map online update method program, and the map online update method program may be integrated on the processor so that it can be directly run by the processor; this is not limited here.
[0064] In one embodiment, the memory includes at least one type of readable storage medium, including flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. The memory can be an internal storage unit of the terminal device, such as the portable hard drive of the terminal device. The memory can also be an external storage device of the terminal device, such as a plug-in portable hard drive, SmartMedia Card (SMC), Secure Digital (SD) card, Flash Card, etc., equipped on the terminal device. Furthermore, the memory can include both internal storage units and external storage devices of the terminal device. The memory can be used not only to store application software and various types of data installed on the terminal device, such as the code of a map online update method program, but also to temporarily store data that has been output or will be output. It should be noted that this embodiment does not limit the memory; the memory is not a necessary component of this application and can be designed according to the processor type.
[0065] Those skilled in the art will understand that Figure 3 The block diagram shown is merely a partial structural diagram related to the present invention and does not constitute a limitation on the terminal device to which the present invention is applied. The specific terminal device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.
[0066] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the methods described above. Any references to memory, storage, operational databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual operating data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for online map updating, characterized in that, The method includes: Acquire image data containing the target scene, and retrieve historical image data that matches the image data from pre-stored map data; Based on the image data and the historical image data, determine the difference data between the image data and the historical image data; wherein, the difference data is used to reflect the differences between objects in the image data and the historical image data; The map data is updated based on the discrepancies. The step of acquiring image data containing the target scene and retrieving historical image data matching the image data from pre-stored map data includes: Acquire image data of a preset target scene; The image data is subjected to feature recognition to obtain image features in the image data; wherein, the image features include image feature points corresponding to objects in the target scene in the image data or geometric shape features corresponding to objects in the target scene in the image data; Based on image feature points and geometric shape features in image data, determine the object category, object shape, and object location in the image data; Based on the image features, feature matching is performed in the map data to obtain the historical image data that matches the image features; The step of performing feature matching in the map data based on the image features to obtain the historical image data that matches the image features includes: Based on the image features, determine the object corresponding to the image features, and set the image features corresponding to the same object as a feature group; The nearest neighbor algorithm is used to determine the associated groups adjacent to the feature group; Based on all the feature groups, the scene image with the highest similarity to all the feature groups is selected from the map data, and the scene image is used as the historical image data; The step of updating the map data based on the discrepancy data includes: If the difference data indicates that the object corresponding to the feature group exists only in the image data, then match the object corresponding to the association group in the historical image data; If a match is successful, the image features corresponding to the feature group are added to the historical image data; If a match is successful, the image features corresponding to the feature group are added to the historical image data, including: If the match is successful, it is determined that the object has a relative positional relationship with the previous object, indicating that the object is newly added to the preset target scene, and the position of the newly added object is determined. Delete the historical image data corresponding to the feature group in the historical image data, and add the image features corresponding to the feature group in the image data to the historical image data.
2. The online map update method according to claim 1, characterized in that, The step of determining the difference data between the image data and the historical image data based on the image data includes: Based on the feature group, group difference data of the feature group in the image data and the historical image data are determined, wherein the group difference data is used to reflect the change of the object corresponding to the feature group between the image data and the historical image data; The difference data of the group is used as the difference data.
3. The online map updating method according to claim 2, characterized in that, The step of updating the map data based on the discrepancy data includes: The group difference data is compared with a preset threshold. If the difference data of the group is greater than the preset threshold, the image features at the position corresponding to the feature group in the historical image data are deleted, and the image features corresponding to the feature group in the image data are written into the corresponding position in the historical image data.
4. The online map updating method according to claim 1, characterized in that, The step of updating the map data based on the discrepancy data includes: If the differential data indicates that the object corresponding to the feature group exists only in the historical image data, then the group lifecycle of the object corresponding to the feature group is obtained. The group lifecycle is used to reflect the number of times the feature group has not been detected in the image data. If the lifecycle of the group has expired, the image features corresponding to the feature group in the historical images are deleted.
5. A map online updating device, characterized in that, The apparatus, used to implement the online map update method as described in any one of claims 1-4, comprises: The image matching module is used to acquire image data containing a preset target scene and to acquire historical image data that matches the image data from map data; The difference analysis module is used to determine the difference data between the image data and the historical image data based on the image data and the historical image data. The difference data is used to reflect the differences between objects in the image data and the historical image data. The map update module is used to update the map data based on the discrepancy data.
6. A terminal device, characterized in that, The terminal device includes a data acquisition device, a memory, and a processor; wherein, the data acquisition device is used to acquire image data of a preset target scene; the memory is used to pre-store map data; and the processor is used to acquire the image data and process the image data according to the online map update method as described in any one of claims 1-4 to update the pre-stored map data.
7. A computer-readable storage medium, characterized in that, A computer-readable storage medium stores a map online update program, which, when executed by a processor, implements the steps of the map online update method as described in any one of claims 1-4.