Map data processing method, device and product

Abstract

The invention provides a map data processing method, device and product, and the method comprises the steps: obtaining to-be-tested map data which comprises a plurality of to-be-tested map elements; a target real-scene image corresponding to the to-be-detected map data is determined, and the target real-scene image comprises at least one to-be-detected map element; vector difference processing is carried out on the to-be-measured map data based on map parent library data, a vector difference result is obtained, and the vector difference result comprises a geometric matching relation and / or attribute information between the map parent library data and the to-be-measured map data; and detecting the current situation of the to-be-detected map data based on the vector difference result and the target real scene image, and generating a detection result for the to-be-detected map data. According to the method and the device, the accuracy and the reliability of map data quality detection are improved, so that more accurate and credible basic data support is provided for application based on map data.

Description

Technical Field

[0001] This application relates to the field of information processing technologies, and particularly to a method, device, and product for map data processing. Background Art

[0002] The vector data to be measured is a high-quality vector map data generated through crowd-sourcing collection, extraction, and fusion processing, including but not limited to vector data of various map elements such as lane lines, ground markings, poles, and signs. For example, the vector data to be measured can be vector data obtained after processing the data transmitted back from vehicle terminals such as online car-hailing vehicles, taxis, and private cars; it can also be vector data obtained after being processed by the above vehicle terminals. To meet the requirements of high-freshness maps, the vector data to be measured will be regularly delivered to vector map data, so as to push the vector data to be measured to a high-quality map data production line for data fusion and update the master database map data. Before data fusion, it is necessary to detect and process the vector data to be measured, so as to obtain the quality information of the vector data to be delivered and to determine whether the vector data to be delivered meets the push production standard.

[0003] The traditional manual detection scheme for vector data to be measured has high detection costs and low efficiency, and cannot meet the requirements of rapid response batch detection. For the automatic detection method based on manually marked ground truth, since manual labor cannot directly mark three-dimensional vector ground truth data, it is necessary to build a ground truth production line at high cost for ground truth production.

[0004] However, due to the complexity of the changes in road signs in the actual scene and the characteristics of crowd-sourcing collection and batch production and delivery of vector data to be measured in multiple cities, if there are current changes such as road reconstruction and expansion, the ground truth marked manually will become invalid. Therefore, the traditional detection scheme based on manually marked ground truth cannot handle the changing scenes of real-world road signs, and the detection accuracy is low. Summary of the Invention

[0005] The main purpose of the embodiments of this application is to provide a method, device, and product for map data processing, which improves the accuracy and reliability of map data quality detection, thereby providing more accurate and reliable basic data support for applications based on map data.

[0006] In a first aspect, an embodiment of the present application provides a method for processing map data, including: obtaining to-be-measured map data, where the to-be-measured map data includes a plurality of to-be-measured map elements; determining a target real-scene image corresponding to the to-be-measured map data, where the target real-scene image includes at least one of the to-be-measured map elements; performing vector difference processing on the to-be-measured map data based on map master library data to obtain a vector difference result, where the vector difference result includes a geometric matching relationship and / or attribute information between the map master library data and the to-be-measured map data; detecting the currency of the to-be-measured map data based on the vector difference result and the target real-scene image, and generating a detection result for the to-be-measured map data.

[0007] In a second aspect, an embodiment of the present application provides a map data processing device, including:

[0008] An obtaining module, configured to obtain to-be-measured map data, where the to-be-measured map data includes a plurality of to-be-measured map elements;

[0009] A determining module, configured to determine a target real-scene image corresponding to the to-be-measured map data, where the target real-scene image includes at least one of the to-be-measured map elements;

[0010] A difference module, configured to perform vector difference processing on the to-be-measured map data based on map master library data to obtain a vector difference result, where the vector difference result includes a geometric matching relationship and / or attribute information between the map master library data and the to-be-measured map data;

[0011] A detection module, configured to detect the currency of the to-be-measured map data based on the vector difference result and the target real-scene image, and generate a detection result for the to-be-measured map data.

[0012] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0013] At least one processor; and

[0014] A memory communicatively connected to the at least one processor;

[0015] Wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the electronic device to execute the method described in any of the above aspects.

[0016] In a fourth aspect, an embodiment of the present application provides a cloud device, including:

[0017] At least one processor; and

[0018] A memory communicatively connected to the at least one processor;

[0019] Wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to cause the cloud device to execute the method described in any of the above aspects.

[0020] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium storing computer-executable instructions, and when a processor executes the computer-executable instructions, the method described in any of the above aspects is implemented.

[0021] In a sixth aspect, an embodiment of the present application provides a computer program product including a computer program, and when the computer program is executed by a processor, the method described in any of the above aspects is implemented.

[0022] The map data processing method, device and product provided by the embodiments of the present application obtain the map data to be measured and determine the corresponding target real-scene image, and perform vector difference processing on the map data to be measured by using the map master library data to generate a vector difference result that can represent the geometric matching relationship and / or attribute information between the map data to be measured and the map master library data. Then, in combination with the target real-scene image and the vector difference result, the currency of the map data to be measured is detected to generate a detection result for the map data to be measured. Among them, the target real-scene image provides a real-time environmental background and actual road conditions, so that the final detection result can cope with the changes of actual scene road signs, improve the accuracy and reliability of the map data quality detection, and thus provide more accurate and credible basic data support for applications based on map data. Description of the Drawings

[0023] The drawings here are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application. Obviously, the drawings in the following description are some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0024] Figure 1 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application;

[0025] Figure 2 It is a schematic application scenario diagram of a map data processing system provided by an embodiment of the present application;

[0026] Figure 3 It is a schematic flowchart of a map data processing method provided by an embodiment of the present application;

[0027] Figure 4A schematic diagram of the correspondence between each trajectory point in the sampling point trajectory in the real-scene database provided by this application and the corresponding real-scene image;

[0028] Figure 5 A schematic diagram of the up-to-date semantic difference process provided by this application;

[0029] Figure 6 A schematic diagram of the map data detection scenario provided by this application;

[0030] Figure 7 A schematic diagram of the real-scene image matching process provided by this application;

[0031] Figure 8 A schematic diagram of the high-order vector difference sub-process provided by the embodiment of this application;

[0032] Figure 9 A schematic diagram of the OBJ vector difference method process for an embodiment of this application;

[0033] Figure 10 A schematic diagram of the structure of a map data processing device provided by the embodiment of this application;

[0034] Figure 11 A schematic diagram of the structure of a cloud device provided by the embodiment of this application.

[0035] Through the above-mentioned drawings, the specific embodiments of this application have been shown, and there will be more detailed descriptions hereinafter. These drawings and textual descriptions are not intended to limit the scope of the concept of this application in any way, but to illustrate the concept of this application to those skilled in the art by referring to specific embodiments. Detailed implementation manners

[0036] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all the implementation manners consistent with this application.

[0037] The term "and / or" in this article is used to describe the association relationship of associated objects, and specifically represents that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone.

[0038] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data that have been authorized by the user or fully authorized by all parties. Moreover, the collection, use, and processing of relevant data need to comply with the relevant laws, regulations, and standards of relevant countries and regions, and corresponding operation entrances are provided for users to choose to authorize or reject.

[0039] To clearly describe the technical solutions of the embodiments of this application, the nouns involved in this application are first defined as follows:

[0040] Map element: It refers to various facilities and elements existing on the road, including but not limited to traffic signs, traffic lights, street lights, road markings, signposts, signs, guardrails, pedestrian crossing facilities, etc. These topographic elements play an important role in road traffic management, helping to guide and regulate traffic flow and ensuring driving safety and traffic order.

[0041] Vector data to be measured: It is a high-quality vector map data generated through crowdsourcing collection, extraction, and fusion processing, including but not limited to vector data of various map elements such as lane lines, ground markings, poles, and signs. For example, it can be vector map data to be delivered.

[0042] Currency: It means that the geographical space information provided by the map should reflect the current latest situation as much as possible. To ensure the currency of topographic maps, it is necessary to promptly reflect the actual changes of human and natural elements and conduct surveying or resurveying of the changed geographical elements on the ground. The shorter the map update cycle, the stronger the currency. In the field of map data processing, "currency change" usually refers to the actual dynamic changes of geographical features, infrastructure, land use, or environmental conditions reflected in a geographic information system or map data, such as the addition, deletion, or attribute change of lane lines, ground markings, poles, and signs.

[0043] OBJ: Object, that is, in the embodiments of this application, it refers to the vector data of discrete and independent map elements on the road, including but not limited to poles, signs, arrows, channelizing islands, stop lines, zebra crossings, and other ground markings.

[0044] Geometric matching: It is a positioning and matching technology based on geometric information such as the shape, edge, and contour of an object.

[0045] OCR: Optical Character Recognition, optical character recognition.

[0046] IOU: Intersection over Union, the intersection over union ratio, is an important metric in the field of object detection for measuring the degree of overlap between a predicted bounding box and a ground truth bounding box.

[0047] Elevation difference: Usually used to describe the vertical height difference between two points.

[0048] JSON: JavaScript Object Notation, JavaScript object notation, is an open standard lightweight data interchange format.

[0049] Ratio of intersection area: The ratio of the intersection area of two regions to the area of one or both regions.

[0050] Ratio of area difference: The ratio of the difference between the areas of two regions to the area of one of the regions.

[0051] The map data processing method of the embodiments of this application can be applied to any application field that requires map data.

[0052] The vector data to be measured is a high-quality vector map data generated through crowdsourcing collection, extraction, and fusion processing, including but not limited to vector data of various map elements such as lane lines, ground markings, poles, and signs. For example, the vector data to be measured can be vector data obtained after processing the data transmitted back from vehicle terminals such as online car-hailing, taxis, and private cars; it can also be vector data obtained after processing by the above vehicle terminals.

[0053] In order to meet the requirements of high freshness of the map, the vector data to be measured will be regularly delivered with vector map data, so as to push the vector data to be measured to the high-quality map data production line for data fusion and update the map data in the master database. Before data fusion, it is necessary to perform detection processing on the vector data to be measured, so as to obtain the quality information of the vector data to be delivered and to be measured, and then determine whether the vector data to be delivered meets the push production standard.

[0054] The traditional manual detection scheme for vector data to be measured has high costs and low efficiency, and cannot meet the requirements of rapid response and batch evaluation. For the automatic detection method based on manually labeled ground truth, since manual annotation of three-dimensional vector ground truth data is not directly possible, it is necessary to build a high-cost ground truth production line for ground truth production.

[0055] However, due to the complexity of the changes in road signs in the actual scenario, and the characteristics of the vector data to be measured, which are collected through crowdsourcing and mass-produced and delivered to multiple cities. If there are currency changes such as road reconstruction and expansion, the true values marked manually will become invalid. Therefore, the traditional detection scheme based on manually marked true values cannot handle the changing scenarios of real-world road signs, and the detection accuracy is low. The evaluation of the current vector data to be measured faces at least two major problems: how to perform efficient and batch automated detection, and how to make the detection method have the ability to evaluate currency.

[0056] To solve at least one of the above problems, the embodiments of the present application provide a map data processing solution. By obtaining the map data to be measured and determining the corresponding target real-world image, and using the map master database data to perform vector difference processing on the map data to be measured, a vector difference result that can represent the geometric matching relationship and / or attribute information between the map data to be measured and the map master database data is generated. Then, by combining the target real-world image and the vector difference result, the currency of the map data to be measured is detected, and a detection result for the map data to be measured is generated. Among them, the target real-world image provides the real-time environmental background and actual road conditions, so that the final detection result can handle the changes in real-world road signs, improve the accuracy and reliability of the map data quality detection, and thus provide more accurate and credible basic data support for applications based on map data.

[0057] The following will describe in detail some embodiments of the present application with reference to the accompanying drawings. Without conflict between the embodiments, the embodiments and the features in the embodiments can be combined with each other. In addition, the step sequence in the following method embodiments is only an example and is not strictly limited.

[0058] As Figure 1 shown, this embodiment provides an electronic device 1, including: at least one processor 11 and a memory 12. Figure 1 Taking one processor as an example. The processor 11 and the memory 12 are connected through a bus 10. The memory 12 stores instructions executable by the processor 11. The instructions are executed by the processor 11, so that the electronic device 1 can execute all or part of the processes of the methods in the following embodiments, so as to improve the accuracy and reliability of the map data quality detection, and thus provide more accurate and credible basic data support for applications based on map data.

[0059] In one embodiment, the electronic device 1 can be a mobile phone, a tablet computer, a laptop computer, a desktop computer, or a large computing system composed of multiple computers.

[0060] Figure 2 It is a schematic diagram of an application scenario 200 of a map data processing system provided by the embodiments of the present application. As Figure 2As shown in the figure, the system includes: a server 210 and a terminal 220, where:

[0061] The server 210 may be a data platform that provides map data processing services. For example, it may be a data platform that provides data processing services for a navigation map service engine. In an actual scenario, a data platform may have multiple servers 210. Figure 2 In this case, one server 210 is taken as an example.

[0062] The terminal 220 may be a mobile device used to log in to the data platform. For example, it may be an on-vehicle crowdsourcing collection device for map data collection, a road monitoring camera, etc. It may also be a device such as a computer, a mobile phone, or a tablet used when evaluating the map data to be measured. There may also be multiple terminals 220. Figure 2 In this case, two terminals 220 are taken as an example for illustration.

[0063] Information can be transmitted between the terminal 220 and the server 210 through the Internet so that the terminal 220 can access the data on the server 210. The above-mentioned terminal 220 and / or server 210 can both be implemented by the electronic device 1.

[0064] The map data processing solution of the embodiments of the present application can be deployed on the server 210, or can be deployed on the terminal 220, or partially deployed on the server 210 and partially deployed on the terminal 220. In an actual scenario, it can be selected based on actual needs, and this embodiment does not make a limitation.

[0065] When the map data processing solution is fully or partially deployed on the server 210, a call interface can be opened to the terminal 220 to provide algorithm support for the terminal 220.

[0066] The method provided by the embodiments of the present application can be implemented by the electronic device 1 executing corresponding software code and by performing data interaction with the server. Among them, the electronic device 1 can be a local terminal device. When the method runs on the server, the method can be implemented and executed based on a cloud interaction system, where the cloud interaction system includes a server and a client device.

[0067] In a possible implementation manner, the method provided by the embodiments of the present application provides a graphical user interface through a terminal device, where the terminal device can be the aforementioned local terminal device or the client device in the aforementioned cloud interaction system.

[0068] Please refer to Figure 3 , which is a map data processing method according to an embodiment of the present application. This method can be executed by Figure 1 the electronic device 1 shown in the figure and can be applied to Figure 2In the map data processing application scenario shown in the figure, it is to improve the accuracy and reliability of map data quality detection, so as to provide more accurate and credible basic data support for applications based on map data. In this embodiment, taking the terminal 220 as the execution end as an example, the method includes the following steps:

[0069] Step 301: Obtain the map data to be tested, where the map data to be tested includes multiple map elements to be tested.

[0070] In this step, map elements refer to various facilities and elements existing on the road, including but not limited to traffic signs, traffic lights, street lights, road markings, sign poles, signboards, guardrails, pedestrian crossing facilities, etc. The map data to be tested refers to the map data that needs to be subjected to data quality detection. For example, it can be the vector map data to be delivered. The map data to be tested can include map data of multiple regions, such as vector map data of multiple cities to be delivered in batches. In the map data to be tested, the map elements to be tested can be represented by vector data. For example, the vector data of various facilities and elements existing on a certain road.

[0071] In one embodiment, step 301 may specifically include: obtaining the original map data to be tested and the original master database data, where the original map data includes vector data of multiple map elements to be tested, and the original master database data includes multiple preset reference map elements. Filter out the vector data of specified types of map elements from the original map data. Determine the road surface area that includes both the map elements to be tested and the reference map elements as the evenly covered area, and determine the data in the original map data located in the evenly covered area as the map data to be tested. And / or determine the data in the original master database data located in the evenly covered area as the map master database data.

[0072] In this embodiment, the master database is a database that contains comprehensive map data and can be used as data output. For example, it can be a high-quality map database of a specified area. In the actual scenario, if it is found that the map data in the real scene has changed, the changes can be updated to the master database. The reference map element refers to the data object contained in the master database data. First, obtain the original map data to be tested and the original master database data. The original map data is the initial vector data to be tested, and the original master database data is the preset standard map data. Taking the vector map data to be delivered as the original map data as an example, based on the batch granularity, the core map data of multiple cities within a regular delivery batch of a specified high-speed rail can be stored as the initial vector data to be tested according to the preset map granularity, and the downstream process nodes all flow according to the map granularity. Here, the preset map granularity refers to dividing the map data into regions according to the preset division rules. For example, a 10-kilometer * 10-kilometer range area on the map is divided into one map sheet. According to this rule, the vector map data to be delivered can be divided into multiple map sheet data, and subsequent data processing can be performed based on the map granularity to improve the data processing efficiency.

[0073] The specified type of map element refers to the map element that does not need to be detected, which can filter out the vector data to be measured (i.e., non-evaluation type data) concerned by the map element types that do not need to be detected. By filtering out the vector data of the preset type of map element from the original map data, the streamlining and accuracy of the map data are optimized, and the data calculation amount is reduced. The specified type of map element can be specified by the user in real time or determined based on an external file. In the external file, the true value ID of the type of map element that does not need to be detected can be preset according to actual needs. The map element corresponding to this true value ID does not participate in the subsequent data detection. For example, the specified type of map element in the original map data can be marked as a non-evaluation type, and no calculation is performed downstream, reducing unnecessary data calculation and improving the calculation efficiency.

[0074] Optionally, in order to further improve the data detection efficiency, coverage calculation can be performed. Taking the road surface as an example, the road surface area that includes both the vector data of the map element to be measured and the reference map element is determined as the evenly covered area. Taking the detection scenario of the vector map data to be delivered as an example, assuming that the road surface at a high-precision level (where the map data at this level can include the vector data of the map elements within the road surface) is used as the granularity, calculate whether the preset type of vector data to be measured and the original master database data exist simultaneously within the area of road surface A. If so, define the area of road surface A as the evenly covered area, and the vector ratio of the evenly covered area can be calculated based on this. Here, the preset type refers to the type of map element, which can be set according to actual needs. For example, in an actual scenario, there must be lane lines and speed limit signs on highways, so the preset type can refer to lane lines and speed limit signs.

[0075] Taking the lane line with the preset type being the specified type as an example, specifically, the calculation of the evenly covered area can be as follows:

[0076] i. Assume that the road surface has 2 geometric expressions: the road surface HD Road and the intersection surface HD Object. Extract these two types of data from the vector data to be measured and the original master database data respectively. To avoid errors caused by an overly long geometric road surface in highway and urban expressway scenarios, the road surface can be divided into small surfaces of equal length according to the driving direction.

[0077] ii. Identify whether a certain specified type of OBJ and lane lines exist simultaneously within the segmented road surface in the vector data to be measured and the original master database data, and the lane lines meet a certain preset length. If both exist, mark this road surface as the evenly covered area.

[0078] iii. Mark the vector data located in the road surface of the evenly covered area in the vector data to be tested as evenly covered, and determine it as the map data to be tested. Mark the vector data located in the road surface of the evenly covered area in the original mother library data as evenly covered, and determine it as the map mother library data. Mark the OBJ in the road surface of the non-evenly covered area as uncovered, and the data downstream of the non-evenly covered area does not participate in the calculation, thereby improving data processing efficiency.

[0079] iv. At the same time, the coverage rate of the total proportion of the vector data to be tested in the coverage area can be calculated, and the coverage rate of the map elements with vector data to be tested but no corresponding mother library true value can be calculated.

[0080] In one embodiment, filtering out the vector data of the map elements of the preset type from the original map data may specifically include: detecting a nested traffic sign existing in the original map data, the nested traffic sign including a first traffic sign and a second traffic sign, the area of the first traffic sign being larger than the area of the second traffic sign, assigning the text information of the second traffic sign in the nested traffic sign to the first traffic sign, and removing the vector data of the second traffic sign.

[0081] In the present embodiment, the traffic nesting sign refers to the embedding of the second traffic sign with a smaller area in the first traffic sign with a larger area in the actual scene, so as to identify the traffic information more finely, but the original map data may identify one traffic nesting sign as two map elements, which may lead to data redundancy. In the process of filtering out the vector data of the map elements of the preset type from the original map data, individual filtering can be performed for the traffic nesting sign. First, the traffic nesting sign existing in the original map data is detected, and then the text information of the second traffic sign (small sign) is assigned to the first traffic sign (big sign), thereby retaining all necessary information content. Subsequently, the vector data of the second traffic sign (small sign) is eliminated to reduce redundant information and data volume. This processing method not only effectively simplifies the map data structure, reduces the complexity of data storage and processing, but also ensures the integrity and consistency of information, improves the practicality and efficiency of map data, and particularly in navigation systems and geographic information systems, can significantly improve the processing speed and response capability of data.

[0082] Taking the detection scenario of the vector map data to be delivered as an example, identifying the non-ground OBJ traffic nested signs in the vector data to be tested will generate two (big sign inside small sign) vector OBJs, which will increase the redundancy index of the vector data to be tested during geometric matching. Therefore, it is possible to perform big sign inside small sign detection, mark the detected nested small sign with a non-detection type mark, and assign the OCR text of the small sign to the big sign.

[0083] Optionally, nested cards can be detected as follows:

[0084] Calculate the distance dist between the center points of two adjacent cards, project the two adjacent cards to the XOZ plane and YOZ plane of the Cartesian coordinate system respectively, calculate the intersection area ratio xzArea of the two card geometries in the projection plane XOZ, and calculate the intersection area ratio yzArea of the two card geometries in the projection plane YOZ. If the intersection area ratio in any one of the projection planes is greater than the set ratio threshold and the distance dist is less than the set distance threshold, the two cards are determined to be nested cards.

[0085] Optionally, the data preprocessing process may include data range filtering and coverage calculation, and range filtering includes but is not limited to: filtering out non-evaluation type data, filtering out small cards (correcting traffic nested cards), and deleting non-evaluation type data recorded in the true value plug-in file. Then jump to coverage calculation, including but not limited to: road surface cutting, calculating the covered road surface, calculating the covered intersection surface, and calculating the OBJ covered by the covered road surface.

[0086] Step 302: Determine a target real scene image corresponding to the map data to be measured, wherein the target real scene image includes at least one map element to be measured.

[0087] In this step, the real-scene image refers to an image that can reflect the basic characteristics of the map elements to be tested in the map data to be tested in the real scene. For example, the target real-scene image of the map elements to be tested includes the appearance characteristics of the map elements to be tested, thereby providing real-time environmental background and actual road conditions, so that the final detection results can cope with changes in road signs in actual scenes, thereby improving the accuracy and reliability of map data quality detection.

[0088] In one embodiment, step 302 may specifically include: determining a target real scene image corresponding to each map element to be measured in the map data to be measured according to a preset real scene database, wherein the real scene database includes sampling point trajectories of multiple roads and a real scene image corresponding to each sampling point in the sampling point trajectories.

[0089] In this embodiment, a sampling point refers to the position where an image acquisition device on a road captures a real-scene image. It is assumed that a crowdsourcing collection vehicle takes a photo at the sampling point, and the obtained photo serves as the real-scene image corresponding to the sampling point. When an actual collection vehicle captures road images, photos may be taken at multiple position points on a road, and all the photo-taking position points on a road are connected to form the sampling point trajectory of the road. The real-scene images of multiple map elements can be stored in a real-scene database. For example, real-scene images can be collected in advance at different sampling points on each road, and the OBJ semantic information on the real-scene images can be extracted and bound to the collection points. Based on this, a real-scene database is established. Then, the real-scene database can be used to determine the current real-scene image corresponding to each map element to be measured as the target real-scene image. The current real-scene image provides a real-time environmental background and actual road conditions, enabling the final detection result to cope with changes in actual scene road signs and improving the accuracy and reliability of map data quality detection.

[0090] In one embodiment, determining the target real-scene image corresponding to each map element to be measured in the map data to be measured according to a preset real-scene database includes: determining the sampling point trajectory bound in the real-scene database for the target road to which each map element of the road object to be measured belongs; projecting the center point of the map element to be measured onto the corresponding sampling point trajectory to obtain a target projection point; selecting a target sampling point from the sampling point trajectory that is at a preset distance from the target projection point, and determining the current real-scene image corresponding to the target sampling point in the real-scene database as the target real-scene image.

[0091] In this embodiment, the real-scene database can be photo data of roads collected in real time through crowdsourcing. For example, for a road, multiple sampling points can be determined according to a preset sampling frequency to form the sampling point trajectory of the road, and real-scene images are taken at each sampling point. Here, the relationship between the real-scene image and the sampling point is that the image of the actual surrounding scene captured by the image acquisition device located at the sampling point is bound to the sampling point. The real-scene database has characteristics such as high freshness and rich semantics, and the semantics are the OBJ semantic information extracted from the real-scene images.

[0092] As Figure 4 shown, it is a schematic diagram of the correspondence between each sampling point on the sampling point trajectory in a real-scene database provided by this application and its corresponding real-scene image. Figure 4 In (1) is a schematic diagram of the sampling point trajectory of a certain road. Figure 4 In (2) is the real-scene image corresponding to the sampling point P in (1). Subsequently, the semantic information shown in this real-scene image can be used as a reference. Based on Figure 4The real-scene database shown performs a currency evaluation on the to-be-tested map data. For example, through the semantics of real images with a very high freshness, it further verifies whether the matching results (new additions, deletions, and attribute changes) based on geometric matching in step 303 are caused by currency changes, thus solving the currency problem.

[0093] When determining the target real-scene image corresponding to the to-be-tested map feature, first determine the target road to which each to-be-tested map feature belongs, and find the sampling point trajectory bound to the target road in the real-scene database. Then, project the center point of the to-be-tested map feature onto the sampling point trajectory to obtain the target projection point, ensuring the spatial relevance between the to-be-tested object and the real-scene data. Subsequently, select the target sampling point from the sampling point trajectory. The selection method of the target sampling point can be to move a preset distance in a specified direction from the target projection point along the direction of travel of the target road, and use the sampling point at the moved position as the target sampling point corresponding to the target projection point. Taking the direction of travel of the target road as the front, the specified direction here includes but is not limited to the front, rear, left, and right of the target projection point. Since in the actual scene, the real-scene image collected when the camera is at the target projection point is generally an image of the scene around the target projection point and cannot truly reflect the real-scene image at the position of the target projection point, the real-scene image collected at the target sampling point outside the preset distance by moving a preset distance from the target projection point can include the real-scene image at the position where the target projection point is located. The preset distance can be set according to actual needs. For example, it can be 30 meters. That is, in the direction of travel of the target road, moving 30 meters backward from the target projection point can obtain the target sampling point, and use the current real-scene image corresponding to the target sampling point in the real-scene database and the semantic information of the image as the target real-scene image of the to-be-tested map feature. Through precise spatial projection and distance screening, it is ensured that the target real-scene image can accurately and truly reflect the real-scene information of the corresponding to-be-tested map feature, thus providing more real and reliable support for the verification and application of map data.

[0094] During the real - scene image matching process, the stored original map data to be measured can be used to start the real - scene image matching sub - processes respectively in a concurrency - driven manner according to the preset map sheet granularity. The real - scene image matching sub - process mainly completes the road matching between the map data to be measured and the high - quality map data (map master database data), realizes the road attribution at the map element level, and realizes the mapping (drawing) of the map data to be measured and the road. Here, the road level can be the lane level or the road level without distinguishing lanes. For example, if road A has 3 lanes, during the real - scene image matching process, it can be regarded as a high - quality road or can be distinguished by lane lines and divided into 3 sub - roads. Based on the target road to which the map element to be measured belongs, the system automatically recommends real - scene images for each target road to which the map element to be measured belongs. Since the real - scene images at night have poor light, the images are dim and unclear. For example, the real - scene images on the target road can be screened by time, the real - scene images collected at night can be excluded, and multiple latest clear real - scene images can be selected for downstream semantic analysis to improve data accuracy.

[0095] Step 303: Perform vector difference processing on the map data to be measured based on the map master database data to obtain a vector difference result, where the vector difference result includes the geometric matching relationship and / or attribute information between the map master database data and the map data to be measured.

[0096] In this step, the attribute information here refers to the basic information of the map element, such as information about the color, shape, line type, text content, etc. of the map element. For example, the attribute information of a non - crossable lane line can include a white solid line. The attribute information of a crossable lane line can include: a white dashed line. The attribute information of a straight - ahead arrow can include: a straight arrow. The attribute information of a bus - only lane can include information such as the ground text "bus - only lane". Vector difference processing refers to performing mathematical operations and analysis on vector data to calculate the differences between two sets of vector data. By performing vector difference processing on the map data to be measured based on the map master database data, the geometric matching relationship and / or attribute information between the map data to be measured and the map master database data can be revealed. Vector difference processing can identify and quantify the geometric differences between the map data to be measured and the map master database data, such as position offset, shape variation, and scale change, etc., and can also analyze the differences in the attribute information between these two sets of map data, such as differences in road type, color of ground markings, text content of road signs, line type of lane lines, etc. It provides reliable technical support for the update, correction, and verification of map data, and can then assist in quickly identifying anomalies or changes in map data.

[0097] It should be noted that the execution order of the above - mentioned step 302 and step 303 is only an example. In the actual scenario, step 303 can also be executed first and then step 302, or step 302 and step 303 can be executed simultaneously. The embodiments of the present application do not make any limitations in this regard.

[0098] The reference map element refers to the vector data of the standard map elements stored in the map master database.

[0099] Optionally, before step 303, vector difference processing grouping data preparation can be performed first: perform data specification conversion, such as converting the to-be-tested map data into the JSON specification required for road matching, and perform coordinate transformation, such as uniformly transforming the to-be-tested map data and the map master database data into the 2019 coordinate system, so as to facilitate subsequent data calculation in the unified coordinate system.

[0100] Optionally, the to-be-tested map data, the map master database data, and the target real-scene image corresponding to each to-be-tested map element can be uniformly transformed into the 2019 coordinate system and then into the Cartesian rectangular coordinate system, which is convenient for downstream vector geometric difference matching to calculate vector relationships.

[0101] In one embodiment, step 303 may specifically include: performing geometric matching between the to-be-tested map element and the reference map element in the map master database data, forming a matching pair between the target map element with the largest overlap degree between the reference map element and the to-be-tested map element and the to-be-tested map element, obtaining the geometric matching relationship between the to-be-tested map element and the reference map element, and recording the attribute information of the matching pair. The vector difference result includes the geometric matching relationship and the attribute information.

[0102] In this embodiment, the attribute information of the matching pair refers to the attribute information jointly belonging to the two map elements forming the geometric matching relationship. For example, for the matching pair formed by two left-turn arrows, its attribute information may include: white left-turn arrow. By traversing each to-be-tested map element in the to-be-tested map data and performing geometric matching with the reference map elements in the map master database data one by one, the matching pairs between the to-be-tested map elements and the reference map elements can be effectively identified and established. By calculating the overlap degree, the target map element with the largest overlap degree with the to-be-tested map element is selected from the map master database data as the matching object, ensuring the accuracy of geometric matching. The establishment of the matching pair not only reveals the geometric matching relationship between the two, but also comprehensively reflects the differences between the two sets of map data by recording the attribute information of the matching pair. The finally obtained vector difference result contains detailed geometric matching relationships and attribute information, providing a solid foundation for the analysis, update, and correction of map data, and improving the efficiency and accuracy of map data processing.

[0103] Optionally, during the geometric matching process, the to-be-tested map elements in the to-be-tested map data can be traversed, and for the current to-be-tested map element, a set of candidate reference map elements intersecting with the current to-be-tested map element is queried from the map master database data. The target map element with the largest overlap degree between the set of candidate reference map elements and the current to-be-tested map element forms a matching pair with the current to-be-tested map element.

[0104] In this embodiment, first, each map element to be measured in the map data to be measured is traversed. For the current map element to be measured, a set of candidate reference map elements that geometrically intersect with it is queried from the map master database data to ensure comprehensive consideration of all potential matching elements. Then, by calculating the overlap degree, the target map element with the largest overlap degree with the current map element to be measured in the set of candidate reference map elements is determined, so as to form a geometric matching pair between the current map element to be measured and the target map element, improving the accuracy and reliability of geometric matching.

[0105] In one embodiment, before traversing the map elements to be measured in the map data to be measured in step 303, it further includes: creating a spatial index for the map data to be measured and the map master database data. And querying the set of candidate reference map elements that intersect with the current map element to be measured from the map master database data, which may specifically include: after spatially expanding the current map element to be measured, querying the set of candidate reference map elements that geometrically intersect with the current map element to be measured from the map master database data according to the spatial index.

[0106] In this embodiment, the spatial index refers to a data structure arranged in a certain order according to the position and shape of spatial elements or a certain spatial relationship between spatial elements, which contains summary information of spatial elements, such as the identifier of the element, the circumscribed rotated rectangle, and the pointer pointing to the spatial element entity. By creating a spatial index before traversing the map elements to be measured in the map data to be measured, the efficiency and accuracy of the map data matching process can be significantly improved. First, creating a spatial index for the map data to be measured and the map master database data lays a foundation for subsequent fast querying and matching. For example, the search tree STRtree of the spatial index (which is a tree structure and a variant of the hash tree) is convenient for quickly finding spatially adjacent geometries during downstream geometric calculations. When querying the set of candidate reference map elements that intersect with the current map element to be measured, after caching and spatially expanding the current map element to be measured, and using the spatial index to quickly retrieve the set of candidate reference map elements that geometrically intersect from the map master database data, it not only speeds up the query speed, reduces the consumption of computing resources, but also improves the accuracy and reliability of matching. By effectively using the spatial index and spatial expansion technology, this solution can more quickly identify the reference map elements related to the map elements to be measured, thus optimizing the matching and verification process of map data.

[0107] In one embodiment, the process of determining the target map feature with the largest overlap degree between the candidate reference map feature set and the current map feature to be measured in step 303 may specifically include: calculating the intersection-over-union ratio and the first elevation difference between each candidate reference map feature in the candidate reference map feature set and the current map feature to be measured respectively. The reference map feature in the candidate reference map feature set with an intersection-over-union ratio greater than zero, a first elevation difference less than the first threshold, and the smallest center distance from the current map feature to be measured is determined as the target map feature.

[0108] In this embodiment, first, the intersection-over-union ratio and the first elevation difference between each candidate reference map feature in the candidate reference map feature set and the current map feature to be measured are calculated respectively. These calculations provide detailed information on spatial overlap and elevation consistency. Then, by screening out the reference map features with an intersection-over-union ratio greater than zero and a first elevation difference less than the preset first threshold, the rationality of the map features forming the matching pairs in terms of space and elevation is ensured. If there are multiple reference map features that meet the above screening conditions, the reference map feature with the smallest center distance from the current map feature to be measured can be selected as the final target map feature, further improving the accuracy of the matching, optimizing the quality of geometric matching, and providing more reliable and accurate data support for applications such as navigation systems and geographic information systems.

[0109] Taking the OBJ geometric matching process in the detection scenario of the vector map data to be delivered as an example:

[0110] i. Loop through the map features to be measured gtObjs within the area covered by all map sheets and marked with the evaluation type (i.e., evaluation range).

[0111] ii. After expanding the buf (buffer space) based on each map feature to be measured gtObj, query the master database truth value hdObjs (i.e., the candidate reference map feature set) on the expanded buf space.

[0112] iii. Traverse the candidate reference map feature set hdObjs, and compare each candidate reference map feature hdObj with the current map feature to be measured gtObj one by one.

[0113] iv. Calculate the minimum bounding rotated rectangles of the current map feature to be measured gtOb and each candidate reference map feature hdObj. Here, the long side of the minimum bounding rotated rectangle can be used for subsequent calculation of the long side direction angle difference between the two objects. Calculating the minimum bounding rotated rectangles in advance can facilitate directly obtaining the long side in downstream calculations to calculate the corresponding direction and angle difference for determining the geometric accuracy of the OBJ. For example, given the minimum bounding rotated rectangle of OBJ-A, its long side can be taken out to calculate the long side direction, which represents the direction of OBJ-A; if the minimum bounding rotated rectangles of two objects, OBj-A and OBJ-B, are known, the long sides of both can be obtained, and the long side direction angle difference can be calculated from the long sides of the two, thereby determining whether the directions of the two OBJs are consistent.

[0114] v. Calculate the IOU and the first elevation difference dZ between the current map feature to be measured gtOb and each candidate reference map feature hdObj. Among them:

[0115] 1. Calculate the IOU of the ground OBJ projected onto the XOY plane.

[0116] 2. Calculate the IOU of the non-ground OBJ projected onto the XOZ / YOZ plane.

[0117] vi. Select N reference map features hdObjs from the candidate reference map feature set hdObjs whose above-mentioned IOU is greater than zero and the first elevation difference dZ is less than the first threshold. Here, N is an integer, and N is less than or equal to the total number of objects in the candidate reference map feature set hdObjs.

[0118] vii. Calculate the distance from the current map feature to be measured gtObj to the center point of each of the N reference map features hdObjs.

[0119] viii. Determine the reference map feature hdObj with the shortest distance from the center point of the current map feature to be measured gtObj among the N reference map features hdObjs as the target map feature, and record the current map feature to be measured gtObj and this target map feature as a geometric matching pair.

[0120] In an embodiment, after forming the geometric matching pair, the geometric matching pair can also be verified. Specifically, it can further include: if the target map feature has already formed a matching pair with a historical map feature, calculate the first distance between the target map feature and the current map feature to be measured, calculate the second distance between the target map feature and the historical map feature, and determine the matching pair corresponding to the minimum value of the first distance and the second distance as the final matching pair.

[0121] In this embodiment, a historical map feature refers to a map feature in the map data to be measured that has already formed a matching pair with a certain reference map feature. By introducing a distance calculation and a comparison mechanism for historical matching pairs, the matching process between the map feature to be measured and the reference map feature can be further optimized to ensure that the generated vector difference result is more accurate and reasonable. Specifically, when the selected target map feature has formed a matching pair with a historical map feature, calculate the first distance between the target map feature and the current map feature to be measured, and the second distance between the target map feature and the historical map feature. By comparing these two distances, select the matching pair corresponding to the minimum distance as the final matching pair. This effectively solves the possible matching conflict problem and ensures that the final matching pair is the closest in space, thereby improving the accuracy and stability of geometric matching.

[0122] Taking the OBJ geometric matching process of the above detection scenario of the vector map data to be delivered as an example, after step viii, it may further include:

[0123] ix. Determine whether the target map feature hdObj has already been matched by the vector OBJ of the historical map feature in the map data to be measured in other loops. If so, go to step x; otherwise, go to step xi.

[0124] x. Calculate the center point distance (i.e., the first distance di) of the current matching pair (referring to the geometric matching pair formed by the current map feature to be measured gtObj and the target map feature) and the center point distance (i.e., the second distance ds) of the historical matching pair. Compare which of the two distances is smaller, and take the matching pair corresponding to the minimum center point distance min(di, ds) as the final geometric matching pair.

[0125] xi. Record and update the geometric matching relationship in the vector difference result.

[0126] In one embodiment, step 303 may specifically further include: screening out the first matching pairs corresponding to the preset map feature types from the vector difference result. The first matching pairs include the third map feature to be measured from the map data to be measured and the second reference map feature from the map master database data. Search for the candidate set of reference map features in the map master database data that are of the same type as the second reference map feature and have not been matched. Calculate the second elevation difference and the long side angle difference of the minimum circumscribed rotated rectangle between each reference map feature in the candidate set of reference map features and the second reference map feature respectively. Select the third reference map feature from the candidate set of reference map features whose second elevation difference is less than the second threshold and whose long side angle difference is less than the third threshold. Add the matching relationship between the third reference map feature and the third map feature to be measured.

[0127] In this embodiment, since the annotation methods of the map data to be measured and the map master database data may be inconsistent, there may be a phenomenon where the geometric sizes and quantities of map elements at the same location (such as guide strips, zebra crossings, and ground OCR) are inconsistent between the map master database data and the map data to be measured. For example, a zebra crossing may be a vector data in the map data to be measured, but it is cut into multiple vector data in the map master database data, which may lead to redundant problems in the vector difference results. These types are the preset map element types. By further screening and expanding the vector difference results, the comprehensiveness and accuracy of the matching between the map elements to be measured and the reference map elements can be enhanced. First, the first matching pairs of specific preset map element types are screened out from the vector difference results. The first matching pairs are composed of the third map element to be measured in the map data to be measured and the second reference map element in the map master database data. Then, a candidate set of reference map elements of the same type as the second reference map element and not yet matched is searched for in the map master database data, and the second elevation difference and the long side angle difference of the minimum circumscribed rotated rectangle between these reference map elements in the candidate set and the second reference map element are calculated. Through this refined calculation and comparison, the third reference map elements with a second elevation difference less than the second threshold and a long side angle difference less than the third threshold are screened out from the candidate set of reference map elements. Although the third reference map elements are not matched in the geometric matching, their spatial positions highly overlap with the second reference map elements. Therefore, it is possible that the same map element is cut into the second reference map element and the third reference map element. Therefore, the matching relationship between the third reference map element and the third map element to be measured is newly added. This process not only expands the coverage of the matching but also ensures the rationality and accuracy of the newly added matching relationship through strict geometric and elevation condition screening, improves the accuracy of the data detection results, and thus provides more reliable and detailed support for the application of map data.

[0128] Taking the detection scenario of the above-mentioned vector map data to be delivered as an example, assume that the preset map element types include guide strips, zebra crossings, and ground OCR. Due to the inconsistent annotation methods, the geometric sizes and quantities of the OBJ (guide strips, zebra crossings, and ground OCR) at the same location may be inconsistent between the map master database data and the map data to be measured. For example, a zebra crossing may be a vector data in the map data to be measured, but it is cut into multiple vector data in the map master database data, which may lead to recall and redundancy problems. Therefore, after the OBJ geometric matching process, OBJ supplementary recall can be performed for these three types, and the process can be as follows:

[0129] i. Screen out the following 3 types of geometric matching pairs from the vector difference results: the third map element to be measured gtObj - the second reference map element hdObj for guide strips, zebra crossings, and ground OCR.

[0130] ii. Search for neighboring reference map elements hdobjs of the same type within the geometric expansion buf range of the second reference map element hdObj in the map master database data.

[0131] iii. Take out the candidate set hdobjs of reference map elements that have not been matched from the reference map elements hdobjs of the same type.

[0132] iv. Calculate the second elevation difference dZ between the candidate set hdobjs of unmatched reference map elements and the second reference map element hdObj that has been matched, and the angle difference in the long side direction of the minimum circumscribed rotated rectangle.

[0133] v. Take out the third reference map element hdobijs from the candidate set hdobjs of reference map elements where the second elevation difference dZ < the second threshold and the direction angle difference < the third threshold.

[0134] vi. Supplement the matching relationship between the third map element to be measured gtObj - the third reference map element hdObjs in the vector difference result, where the third map element to be measured and the reference map element can have a one-to-many matching relationship.

[0135] In one embodiment, after step 303, it further includes: detecting the accuracy of the geometric matching relationship in the vector difference result.

[0136] In this embodiment, the geometric matching accuracy of the matching pairs in the vector difference result can be detected to ensure a high degree of consistency in the spatial position and shape between the map elements to be measured and the reference map elements, thereby eliminating inaccurate matching problems caused by geometric errors.

[0137] In one embodiment, taking the detection scenario of the vector map data to be delivered as an example, the OBJ matching pairs in the vector difference result can be divided into ground OBJ matching pairs and non-ground OBJ matching pairs. Ground OBJs can include guiding arrow matching pairs and zebra crossing matching pairs on the road surface, etc. Non-ground OBJ matching pairs can include traffic sign matching pairs and road pole matching pairs on the road, etc. By distinguishing between ground OBJs and non-ground OBJs, specific difference operators for different element types are implemented, such as expanding the three-dimensional space for non-ground OBJs and calculating IOU and other geometric operators by projecting onto the XOZ and YOZ planes respectively. One-to-one matching or one-to-many matching recall is performed according to different element characteristics to improve the accuracy of the evaluation index.

[0138] Taking the accuracy detection of ground OBJ matching pairs in the detection scenario of the vector map data to be delivered as an example, the geometric matching accuracy detection process of ground OBJ matching pairs can include the following steps:

[0139] 1. Obtain the ground OBJ matching pairs from the vector difference results: gtobj - hdobj, where gtobj represents the map feature to be measured and hdobj represents the reference map feature.

[0140] 2. Calculate the minimum circumscribed rotated rectangles of the map feature to be measured gtobj and the reference map feature hdobj respectively.

[0141] 3. Calculate the angle difference dangle of the long side directions of the minimum circumscribed rotated rectangles of the map feature to be measured gtobj and the reference map feature hdobj.

[0142] 4. Calculate the ratio deare of the intersection area of the map feature to be measured gtobj and the reference map feature hdobj.

[0143] 5. Detect whether there is mutual overlapping of the map feature to be measured gtobj, that is, multiple vector data may be recognized for the same map feature to be measured in the vector data to be measured, and detect whether there is mutual overlapping between these multiple vector data.

[0144] 6. If dangle < the angle difference threshold, deare < the area ratio threshold, and there is no mutual overlapping of the map feature to be measured gtobj, it is determined that the ground OBJ matching pair is geometrically accurately matched; otherwise, the geometric matching of the ground OBJ matching pair is inaccurate.

[0145] Taking the accuracy detection of non - ground OBJ matching pairs in the detection scenario of the vector map data to be delivered as an example, the geometric matching accuracy detection process of non - ground OBJ matching pairs may include the following steps:

[0146] 1. Obtain the non - ground OBJ matching pairs from the vector difference results: gtobj - hdobj, where gtobj represents the map feature to be measured and hdobj represents the reference map feature.

[0147] 2. Calculate the angle difference of the OBJ normal directions between the map feature to be measured gtobj and the reference map feature hdobj in three - dimensional space.

[0148] 3. Calculate the three - dimensional distance of the geometric centers between the map feature to be measured gtobj and the reference map feature hdobj.

[0149] 4. Calculate the distance between the center points after the map feature to be measured gtobj and the reference map feature hdobj are projected onto the XOY plane respectively.

[0150] 5. Calculate the elevation difference of the center points of the map feature to be measured gtobj and the reference map feature hdobj.

[0151] 6. Calculate the area difference ratio between the map element gtobj to be measured and the reference map element hdobj.

[0152] 7. Calculate the aspect ratios of the length and width of the map element gtobj to be measured and the reference map element hdobj respectively.

[0153] 8. Compare the calculated values obtained in steps 2 to 7 with the corresponding preset thresholds respectively. If all the calculated values are less than the corresponding preset thresholds, the geometric matching of this non-ground OBJ matching pair is accurate; otherwise, the geometric matching of this non-ground OBJ matching pair is inaccurate.

[0154] In an embodiment, after step 303, it further includes: detecting the accuracy of the attribute information of the matching pairs in the vector difference result.

[0155] In this embodiment, by checking the attribute information of the map element to be measured and the reference map element in the matching pair, such as road text information, road type, identification, direction, etc., the consistency at the semantic and attribute levels of the matching pair is ensured. It can not only identify and correct the possible attribute deviations in the matching process, but also improve the overall credibility and data quality of the vector difference result, providing more accurate and reliable data support for navigation systems, traffic management, and other applications based on map data, and ultimately enhancing the user experience and system performance.

[0156] Taking the detection scenario of the vector map data to be delivered as an example, the detection process of the attribute accuracy of the OBJ matching pair can include the following steps:

[0157] i. Take out the specified type matching pairs (gtobj - hdobj) that are geometrically matched and geometrically accurate from the vector difference result, where gtobj represents the map element to be measured and hdobj represents the reference map element. Here, the specified type matching pairs include but are not limited to the matching pairs of ground OCR and traffic signs.

[0158] ii. Compare whether the OCR texts expressed by the two OBJs in the specified type matching pair are the same?

[0159] iii. If they are the same, record that the attribute information of this specified type matching pair is correct; otherwise, record that the attribute information of this specified type matching pair is incorrect, and a prompt can be issued for timely correction.

[0160] Step 304: Detect the currency of the to-be-tested map data based on the vector difference result and the target real-scene image, and generate a detection result for the to-be-tested map data. In this step, by combining the target real-scene image corresponding to the to-be-tested map feature with the vector difference result, a detection result for the to-be-tested map data is generated. The target real-scene image provides the real-time environmental background and the actual road conditions, and the vector difference result reflects the difference between the to-be-tested map data and the map master database data, so that the final detection result can cope with the changes of the actual scene road signs, not only improving the accuracy and reliability of the map data quality detection, but also effectively identifying and correcting potential map data errors or deviations, thereby providing more accurate and credible basic data support for the applications based on the map data.

[0161] In an embodiment, step 304 may specifically include: for the target matching pairs formed by different road guiding signs in the vector difference result, determining the first semantic information of the first to-be-tested map feature forming the target matching pair in the corresponding target real-scene image; if the first semantic information is the same as the semantic information of the first to-be-tested map feature in the to-be-tested map data and different from the semantic information corresponding to the target matching pair in the map master database data, then determining the first to-be-tested map feature as accurate data in the detection result; and / or, determining the first to-be-tested map feature as a map feature with real-scene changes, and updating the attribute information of the target matching pair according to the first semantic information.

[0162] In this embodiment, in the actual scenario, the map elements may have currency changes. For example, in the real world, due to certain reasons, lane lines, ground markings, poles, signs, etc. may have new additions, deletions, or attribute changes, and these changes may not have been updated in the master database yet. Therefore, when the map data to be tested is matched with the master database, there may be incorrect matches. To detect the accuracy of the map data to be tested, the currency of the map data to be tested can be detected based on the vector difference result. Specifically, for the target matching pairs with different road guiding signs in the vector difference result, for example, the straight arrow B1 in the map data to be tested and the turning arrow B2 in the master map data, the arrow matching pair formed by the straight arrow B1 and the turning arrow B2 in the vector difference result is the target matching pair. Determine the first semantic information corresponding to the second map element to be tested (such as the straight arrow B1) according to the target actual image corresponding to the first map element to be tested. Taking the straight arrow B1 as an example of the first map element to be tested, by comparing this first semantic information with the semantic information of the straight arrow B1 in the map data to be tested and the semantic information corresponding to the turning arrow B2 in the master map data, the system can identify the accuracy and variability of the data. If the first semantic information is consistent with the semantic information of the straight arrow B1 in the map data to be tested but different from the semantic information of the turning arrow B2 in the master map data, it means that the straight arrow B1 is the map element used in the current actual scenario and has not been updated in the master map data yet. Then mark the straight arrow B1 as accurate data in the detection result, which belongs to the normal currency change of the map data to be tested. In addition, if the real scene information shows that the road has changed, update the attribute information of the target matching pair to reflect the latest road conditions. Here, the attribute information includes but is not limited to the text description features of the map elements. In this way, the sensitivity and response ability of the map data to detect its real scene changes are enhanced, and the accuracy of the map data detection is improved.

[0163] As Figure 5 shown, it is a schematic diagram of a currency semantic difference process provided by this application. Figure 5 In the currency attribute change semantic analysis module, taking the currency attribute change detection scenario of the vector map data to be delivered as an example, it may include the following contents:

[0164] i. Extract the arrow matching pairs with geometric matching but different guiding signs from the vector difference result of the map data to be tested.

[0165] ii. Query the HQ road (HQ is the road of the high-quality map, such as a certain lane) to which the current first map element to be tested gtobj belongs in the arrow matching pair (refer to the road vector difference result of the foregoing embodiment).

[0166] iii. Query the SD road (SD is an ordinary road, without distinguishing lanes) associated with the HQ road

[0167] iv. Query the trajectory of the first sampling point in the real-scene database associated with the SD road.

[0168] v. Project the center point of the first map element to be measured, gtobj, onto the trajectory of the first sampling point to obtain the first projection point.

[0169] vi. At a position 30 meters backward along the SD road from the first projection point, take the trajectory point (i.e., the sampling point) at the backward position as the first target sampling point, and obtain the real-scene image of the first target sampling point from the real-scene database.

[0170] vii. Query the OBJ semantic information (i.e., the first semantic information) of the real-scene image of the first target sampling point.

[0171] viii. Determine whether the OBJ semantic information is the same as the semantic information of the first map element to be measured, gtobj, in the map data to be measured and different from the true value of the semantic information of the first map element to be measured, gtobj, in the map master database?

[0172] ix. If so, the first map element to be measured, gtobj, belongs to the normal currency change update of the map data to be measured, and update the attribute information of the arrow matching pair.

[0173] In an embodiment, the vector difference result includes a second map element to be measured that does not form a matching pair with the reference map element in the map master database; step 304 may specifically include: for the second map element to be measured, determine the second semantic information of the second map element to be measured in the corresponding target real-scene image; if the second semantic information is the same as the semantic information of the second map element to be measured in the map data to be measured and the second semantic information does not exist in the map master database, then determine that the second map element to be measured is accurate data in the detection result; and / or, determine that the second map element to be measured is a road object map element newly added in the real scene.

[0174] In this embodiment, by analyzing the second map elements to be measured that have not formed matching pairs and combining with the target real-scene image, the detection result for the map data to be measured is generated, thereby improving the integrity and real-time performance of the evaluation of map data. First, the second semantic information corresponding to the second map elements to be measured is determined according to the target real-scene image. By comparing the second semantic information with the semantic information of the second map elements to be measured in the map data to be measured, if the second semantic information is consistent with the semantic information in the map data to be measured and the second semantic information does not exist in the map master database, then the second map elements to be measured are marked as accurate data in the detection result, indicating that they are valid in the real scene. In addition, if the second semantic information does not exist in the map master database, it may indicate that the second map elements to be measured are newly added map elements in the real scene. This process not only ensures the accuracy of the map data, but also can timely reflect the changes and newly added information in the real scene, improve the adaptability of the map data quality detection, and further provide more dynamic and accurate data support for the navigation system and geographic information applications to adapt to the changing road environment.

[0175] As Figure 5 shown in the currency new semantic analysis module of the vector map data to be delivered, taking the currency new semantic detection scenario of the vector map data to be delivered as an example, it may include the following contents:

[0176] i. Take out the second map elements to be measured gtobj that are not matched in the map data to be measured from the vector difference result, ensuring that the second map elements to be measured do not cover each other.

[0177] ii. Query the HQ road (HQ is the road of the high-quality map, such as a certain lane) to which the current second map elements to be measured gtobj belong (refer to the road vector difference result of the foregoing embodiment).

[0178] iii. Query the SD road (SD is an ordinary road, without distinguishing lanes) associated with the HQ road

[0179] iv. Query the second sampling point trajectory in the real-scene database associated with the SD road.

[0180] v. Project the center point of the second map elements to be measured gtobj onto the second sampling point trajectory to obtain the second projection point.

[0181] vi. At a position 30 meters backward along the SD road from the second projection point, take the trajectory point (i.e., the sampling point) at the backward position as the second target sampling point, and obtain the real-scene image of the second target sampling point from the real-scene database (i.e., the socol database), refer to the recommended result of the photo (real-scene image).

[0182] vii. Query the OBJ semantic information (i.e., the second semantic information) of the real-scene image of the second target sampling point.

[0183] viii. Determine whether the semantic information of the OBJ is the same as the semantic information of the second map element to be measured, gtobj, in the map data to be measured and whether the semantic information of the OBJ does not exist in the map master database data? If so, proceed to step ix.

[0184] ix. If so, determine that the second map element to be measured, gtobj, belongs to the map elements newly added normally in the map data to be measured, and update the attribute information of the geometric matching of the second map element to be measured, gtobj, in the vector difference result.

[0185] In one embodiment, the vector difference result includes a first reference map element that does not form a matching pair with the map element to be measured in the map master database data; step 304 may specifically include: for the first reference map element, determining the third semantic information of the first reference map element in the corresponding target real-scene image; if the third semantic information is different from the semantic information of the first reference map element in the map master database data, determining that the first reference map element has been deleted from the map data to be measured in the detection result; if the third semantic information is the same as the semantic information of the first reference map element in the map master database data, determining that the first reference map element has been omitted from the map data to be measured in the detection result.

[0186] In this embodiment, there are first reference map elements in the map master database that have not formed matching pairs, indicating that there are no map elements in the map data to be measured that geometrically match the first reference map elements. By analyzing the first reference map elements in the map master database that have not formed matching pairs and combining the target real-scene images, a detection result for the map data to be measured is generated, thereby improving the integrity and accuracy of the evaluation of the map data. First, determine the target road to which the first reference map element belongs, and find the third sampling point trajectory bound to the target road in the real-scene database. Then project the center point of the first reference map element onto the corresponding third sampling point trajectory to obtain a third projection point. By selecting a third target sampling point that is at a second preset distance from the third projection point, the selection method of the third target sampling point here can refer to the selection method of the aforementioned first target sampling point, which will not be elaborated here. The system obtains the current real-scene image corresponding to the third target sampling point in the real-scene database as the target real-scene image, and obtains the third semantic information of the target real-scene image. The third semantic information represents the current semantic information of the first reference map element in the real scene. By comparing the third semantic information with the semantic information of the first reference map element in the map master database, if the third semantic information is different from the semantic information of the first reference map element in the map master database, it indicates that the first reference map element exists in the real scene, but only the semantic information has changed, but the first reference map element does not exist in the map data to be measured. Then mark in the detection result that the map data to be measured has deleted the first reference map element, reflecting the update or deletion of the map data to be measured. Conversely, if the third semantic information is the same as the semantic information of the first reference map element in the map master database, it indicates that the first reference map element is missing in the map data to be measured. This process can timely reflect the omissions or updates in the map data to be measured and improve the accuracy of data detection.

[0187] As Figure 5 shown in the currency deletion semantic analysis module of the current situation, taking the currency deletion semantic detection scenario of the vector map data to be delivered as an example, it can include the following contents:

[0188] i. Extract the first reference map element hdobj in the map master database that has not been matched and does not have mutual overlapping from the vector difference result.

[0189] ii. Query the HQ road (HQ is the road of the high-quality map, such as a certain lane) to which the current first reference map element hdobj belongs (refer to the road vector difference result of the aforementioned embodiment).

[0190] iii. Query the SD road (SD is an ordinary road, without distinguishing lanes) associated with the HQ road

[0191] iv. Query the third sampling point trajectory in the real-scene database associated with the SD road.

[0192] v. Project the center point of the first reference map element hdobj onto the third sampling point trajectory to obtain a third projection point. vi. At a point 30 meters backward along the SD road from the third projection point, use the trajectory point (i.e., the sampling point) at the backward position as the third target sampling point, and obtain the real-scene image of the third target sampling point from the real-scene database.

[0193] vii. Query the OBJ semantic information (i.e., the third semantic information) of the real-scene image of the third target sampling point.

[0194] viii. Determine whether the OBJ semantic information is the same as the semantic information of the first reference map element hdobj in the map master database? If so, proceed to step ix.

[0195] ix. If so, it proves that the first reference map element hdobj exists in both the real scene and the map master database and has the same semantics, indicating that the first reference map element hdobj is missing from the map data to be measured. If not, the first reference map element hdobj belongs to the map elements deleted from the map data to be measured, and update the attribute information of the geometric matching relationship of the first reference map element hdobj in the vector difference result.

[0196] In the above Figure 5 In the illustrated embodiment, by combining the current real-scene information with the vector difference result, perform real-time attribute change semantic detection, real-time high-speed rail new change semantic detection, and real-time high-speed rail deletion semantic detection on the map data to be measured respectively, complete the real-time semantic difference processing of the map data to be measured, and correct and update the geometric difference index in the vector difference result caused by real-time problems according to the analysis result, forming the final detection result of the map data to be measured, improving the accuracy of the data detection result.

[0197] The above map data processing method also has the ability of real-time evaluation. Taking the map master database data as the geometric and attribute truth values, and the real-scene image and semantics as the real-time change semantic truth values, perform automatic vector difference with the map data to be measured based on these two truth values, and introduce the semantic information of the sampled real-scene photos collected in real time to make up for the problem that geometric difference cannot handle real-time changes. Correct the geometric difference index based on scenarios such as real-time changes, real-time additions, and real-time deletions. Efficiently realize the rapid evaluation output of recall, accurate road, coverage rate, real-time performance, etc. for batch map data to be measured, with large throughput, low cost, and high concurrency capabilities, improving the efficiency and accuracy of the detection of the map data to be measured.

[0198] Such as Figure 6As shown in the figure, it is a schematic diagram of a map data detection scenario provided by this application. Taking the detection scenario of vector map data to be delivered as an example, it may include processes such as evaluation sample data storage, photo matching sub-process (i.e., real-scene image matching process), high-order vector difference sub-process, and differential data task push. Among them:

[0199] Evaluation sample data storage: Based on batch granularity, the core map data of dozens of cities within a regularly delivered batch is stored as evaluation samples according to map sheet granularity, and downstream process nodes all flow according to map sheet granularity.

[0200] The photo matching sub-process is the real-scene image matching process. In an actual database, the real-scene image can be a photo captured by a camera. Refer to the description of step 302 above.

[0201] As Figure 7 shown in the figure, it is a schematic diagram of a real-scene image matching process provided by this application. Taking the detection scenario of vector map data to be delivered as an example, it includes the following contents:

[0202] a. Matching group data preparation: Perform data specification conversion to convert the original map data to be measured into the JSON specification required for road matching, and perform coordinate transformation. For example, transform the map data to be measured and the preset map road data to the 2019 coordinate system.

[0203] b. Road matching: Realize the road mapping (drawing) at the vector element level of the map data to be measured, that is, match the vector elements of the map data to be measured with the preset map road data to determine the first road to which each vector element in the map data to be measured belongs. Provide data basis for downstream semantic difference currency evaluation.

[0204] c. Photo (real-scene image) recommendation: For the first road within the map sheet range, grab the information in the real-scene database according to the road-level range, select the optimal multiple trips of real-scene information from it according to time and information type, and take out the real-scene image and the semantic information in the image in the real-scene information and bind them to the first road. Provide data basis for downstream semantic difference currency evaluation.

[0205] Optionally, Figure 6 in the high-order vector difference sub-process: Start the high-order vector difference sub-process respectively according to the preset map sheet granularity in a concurrent driving manner for the evaluation samples (map sheets) stored at the evaluation sample data storage node. The high-order vector difference sub-process includes but is not limited to core difference operator capabilities such as lane line vector line difference, OBJ vector difference, lateral relative accuracy difference, and longitudinal relative accuracy difference, and summarize each difference index into a total table and flow it downstream.

[0206] As Figure 8As shown in the figure, it is a schematic diagram of a high-order vector difference sub-process provided by an embodiment of the present application. Taking the detection scenario of vector map data to be delivered as an example, based on a parallel network management system, in a concurrent driving manner, the high-order vector difference sub-process is started respectively according to the map sheet level, including:

[0207] a) Divider (lane line) vector difference: Automatically calculate various evaluation indexes such as the coverage, geometric accuracy, lane consistency rate, and attribute accuracy of the Divider line elements.

[0208] b) OBJ vector difference: Automatically calculate various evaluation indexes such as the coverage, geometric accuracy, attribute accuracy, and recall of the OBJ elements. At the same time, based on the road matching and photo recommendation results, the OBJ semantic information is retrieved from the real scene database, and semantic difference is made with the OBJ in the map data to be measured for currency evaluation. For details, please refer to the descriptions of steps 303 and 304 in the above embodiment, which will not be elaborated here.

[0209] c) Lateral relative accuracy difference: Automatically evaluate the relative accuracy of the lane line in the lateral direction.

[0210] d) Longitudinal relative accuracy difference: Automatically evaluate the relative accuracy of the lane line in the longitudinal direction.

[0211] e) Vector difference summary: Summarize the evaluation indexes output by each vector difference operator to form an evaluation summary table and transfer it downstream.

[0212] For the details of each step of the above method, please refer to the relevant descriptions of the above embodiment, which will not be elaborated here.

[0213] Optionally, the solution of the above embodiment performs differential evaluation based on the map sheet range of high-speed rail metadata. In the actual scenario, a higher-precision differential update range result can also be introduced, and vector difference is only performed on the changed area, which can further reduce the differential data volume, lower the machine cost, and accelerate the output of evaluation indexes. This improved solution mainly changes the magnitude of the evaluation data, and its core algorithm logic is the same as that of the foregoing embodiment.

[0214] Please refer to Figure 9 , which is the OBJ vector difference method of an embodiment of the present application. This method can be executed by the electronic device 1 shown in Figure 1 . Compared with the foregoing embodiment, taking the OBJ vector difference of the map data to be measured as an example, this method includes the following steps:

[0215] Step 901: Data loading, mainly including loading the vector data of the map sheet to be evaluated, the true value vector data of the map master library, the external constraint file (recording the true value IDs of the vectors to be deleted within the map sheet), the road vector difference result (recording the mapping relationship between the vectors and the roads in the map data to be measured), and the photo recommendation result (recording the sampling point trajectory and the semantic information of the trajectory points in the actual database associated with the SD road).

[0216] In response to a map navigation request, a preset anthropomorphic virtual image is displayed on the map interaction interface, and the anthropomorphic virtual image is generated by using the map data processing method of any of the foregoing embodiments.

[0217] Step 902: Coordinate transformation to unify the coordinate system: The data such as the map data to be measured, the true value data of the map master library, and the sampling point trajectory recommended by the real-scene image are uniformly transformed into the 2019 coordinate system and then converted into the Cartesian rectangular coordinate system to facilitate the calculation of the vector relationship in the downstream vector geometric difference matching.

[0218] Step 903: Create a spatial index: The map data to be measured and the true value data of the map master library are classified by feature type, and a spatial index search tree STRtree is created respectively to facilitate the rapid search for spatially adjacent geometries during downstream geometric calculations.

[0219] Step 904: Evaluation range filtering & coverage calculation, mainly including:

[0220] i. Evaluation range filtering: Filter out the high-speed rail vectors that are not concerned by the evaluation type, delete the true value IDs recorded in the external file, and do not participate in the subsequent evaluation.

[0221] ii. Coverage calculation: Taking the HQ road as the granularity, calculate whether the map data to be measured and the map master library data of a specific type exist simultaneously within the range of the HQ road. If they exist, it is defined as the uniformly covered range, and the uniformly covered range vector ratio is calculated accordingly.

[0222] For details, please refer to the relevant descriptions of the corresponding foregoing embodiments.

[0223] Step 905: OBJ vector difference: This module completes the core OBJ vector difference algorithm and outputs indicators such as recall, omission, geometric accuracy, and attribute accuracy. The objects processed by the OBJ vector difference are map elements such as ground arrow OBJs, ground OCRs, guide strips, stop lines, zebra crossings, other ground markings, traffic lights, and traffic signs. For details, please refer to the relevant descriptions of the foregoing embodiments.

[0224] Step 906: Determine whether the currently measured OBJ vector in the map data to be measured needs to be processed for currency semantic difference. If so, go to Step 907; otherwise, go to Step 908.

[0225] Step 907: Perform semantic difference processing on the current OBJ vector to be measured and determine the confidence level. Based on the road vector difference result and the photo recommendation result, this module combines the latest real-scene image semantics to compare the to-be-measured map data and the OBJ semantics in the mother library, realizes the currency evaluation ability, corrects relevant indicators after semantic difference, and then proceeds to Step 908. For details, refer to the description of the foregoing embodiments.

[0226] Step 908: Statistically output indicators, and the output content includes but is not limited to information such as the coverage rate, recall, redundancy, omission, geometric accuracy, attribute accuracy, and currency evaluation of the to-be-measured map data.

[0227] For each step of the above method, details can be referred to the relevant description of the foregoing embodiments, and will not be elaborated here.

[0228] Please refer to Figure 10 , which is the map data processing device 1000 according to an embodiment of the present application. This device can be applied to a terminal and can be applied to Figure 2 the map data processing application scenarios shown in, so as to improve the accuracy and reliability of map data quality detection, thereby providing more accurate and credible basic data support for applications based on map data. The device includes: an acquisition module 1001, a determination module 1002, a difference module 1003, and a detection module 1004. The functional principles of each module are as follows:

[0229] The acquisition module 1001 is used to acquire to-be-measured map data, and the to-be-measured map data includes a plurality of to-be-measured map elements;

[0230] The determination module 1002 is used to determine the target real-scene image corresponding to the to-be-measured map data, where the target real-scene image includes at least one to-be-measured map element;

[0231] The difference module 1003 is used to perform vector difference processing on the to-be-measured map data based on the map mother library data to obtain a vector difference result, and the vector difference result includes the geometric matching relationship and / or attribute information between the map mother library data and the to-be-measured map data;

[0232] The detection module 1004 is used to detect the currency of the to-be-measured map data based on the vector difference result and the target real-scene image, and generate a detection result for the to-be-measured map data.

[0233] In one embodiment, the determination module 1002 is used to determine the target real-scene image corresponding to each to-be-measured map element in the to-be-measured map data according to a preset real-scene database, and the real-scene database includes the sampling point trajectories of multiple roads and the real-scene images corresponding to each sampling point in the sampling point trajectories.

[0234] In one embodiment, the determination module 1002 is specifically configured to determine the sampling point trajectory bound to the target road to which each map element to be measured belongs in the real-scene database; project the center point of the map element to be measured onto the corresponding sampling point trajectory to obtain a target projection point; select a target sampling point from the sampling point trajectory that is at a preset distance from the target projection point, and determine the current real-scene image corresponding to the target sampling point in the real-scene database as the target real-scene image.

[0235] In one embodiment, the differential module 1003 is configured to perform geometric matching on the map element to be measured and the reference map element in the map master database data, form a matching pair between the target map element with the largest overlap degree between the reference map element and the map element to be measured and the map element to be measured, obtain the geometric matching relationship between the map element to be measured and the reference map element, and record the attribute information of the matching pair. The vector difference result includes the geometric matching relationship and the attribute information.

[0236] In one embodiment, the detection module 1004 is configured to, for the target matching pair formed by different road guiding identifiers in the vector difference result, determine the first semantic information of the first map element to be measured in the corresponding target real-scene image; if the first semantic information is the same as the semantic information of the first map element to be measured in the map data to be measured and different from the semantic information corresponding to the target matching pair in the map master database data, then determine that the first map element to be measured is accurate data in the detection result; and / or, determine that the first map element to be measured is a map element with a changed real scene, and update the attribute information of the target matching pair according to the first semantic information.

[0237] In one embodiment, the vector difference result includes a second map element to be measured that does not form a matching pair with the reference map element in the map master database data; the detection module 1004 is specifically configured to, for the second map element to be measured, determine the second semantic information of the second map element to be measured in the corresponding target real-scene image; if the second semantic information is the same as the semantic information of the second map element to be measured in the map data to be measured and the second semantic information does not exist in the map master database data, then determine that the second map element to be measured is accurate data in the detection result; and / or, determine that the second map element to be measured is a newly added map element in the real scene.

[0238] In one embodiment, the vector difference result includes a first reference map element in the map master library data that does not form a matching pair with the map element to be measured; the detection module 1004 is specifically configured to determine, for the first reference map element, the third semantic information of the first reference map element in the corresponding target real-scene image; if the third semantic information is different from the semantic information of the first reference map element in the map master library data, it is determined in the detection result that the first reference map element has been deleted from the map data to be measured; if the third semantic information is the same as the semantic information of the first reference map element in the map master library data, it is determined in the detection result that the first reference map element is missing from the map data to be measured.

[0239] For a detailed description of the above map data processing device 1000, please refer to the description of the relevant method steps in the above embodiments. Their implementation principles and technical effects are similar, and will not be elaborated here in this embodiment.

[0240] Figure 11 The following is a schematic structural diagram of a cloud device 110 provided by an exemplary embodiment of the present application. The cloud device 110 can be used to run the method provided in any of the above embodiments. As Figure 11 shown, the cloud device 110 may include: a memory 1104 and at least one processor 1105, Figure 11 taking one processor as an example.

[0241] The memory 1104 is used to store computer programs and can be configured to store various other data to support operations on the cloud device 110. The memory 1104 may be an Object Storage Service (OSS).

[0242] The memory 1104 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disk.

[0243] The processor 1105 is coupled to the memory 1104 and is configured to execute the computer program in the memory 1104 to implement the solution provided in any of the above method embodiments. The specific functions and achievable technical effects will not be elaborated here.

[0244] Further, as Figure 11 shown, the cloud device further includes: other components such as a firewall 1101, a load balancer 1102, a communication component 1106, and a power supply component 1103. Figure 11 Only some components are schematically shown in Figure 11The components shown.

[0245] In one embodiment, the above Figure 11 The communication component 1106 is configured to facilitate communication, in a wired or wireless manner, between the device in which the communication component 1106 is located and other devices. The device in which the communication component 1106 is located can access a communication standard-based wireless network, such as WiFi, 2G, 3G, 4G, LTE (Long Term Evolution), 5G and other mobile communication networks, or a combination thereof. In an exemplary embodiment, the communication component 1106 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 1106 further includes a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on technologies such as Radio Frequency Identification (RFID) technology, Infrared Data Association (IrDA) technology, Ultra Wide Band (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0246] In one embodiment, the above Figure 11 The power supply component 1103 supplies power to various components of the device in which the power supply component 1103 is located. The power supply component 1103 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device in which the power supply component is located.

[0247] An embodiment of the present application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the method of any of the foregoing embodiments.

[0248] An embodiment of the present application also provides a computer program product including a computer program, which, when executed by a processor, implements the method of any of the foregoing embodiments.

[0249] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of modules is only a logical function division, and there may be other division methods in actual implementation. For example, multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0250] The integrated module implemented in the form of software functional modules can be stored in a computer-readable storage medium. The above-mentioned software functional modules are stored in a storage medium and include several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor to execute some steps of the methods in various embodiments of the present application.

[0251] It should be understood that the above-mentioned processor may be a central processing unit (CPU for short), or other general-purpose processors, digital signal processors (DSP for short), application specific integrated circuits (ASIC for short), etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the application can be directly implemented by the execution of the hardware processor, or can be implemented by the combination of hardware and software modules in the processor. The memory may include a high-speed RAM (Random Access Memory) memory, and may also include a non-volatile storage NVM (Nonvolatile memory for short), such as at least one disk memory, and may also be a USB flash drive, a mobile hard disk, a read-only memory, a magnetic disk or an optical disk, etc.

[0252] The above-mentioned storage medium 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 for short), an electrically erasable programmable read-only memory (EEPROM for short), an erasable programmable read-only memory (EPROM for short), a programmable read-only memory (PROM for short), a read-only memory (ROM for short), a magnetic memory, a flash memory, a magnetic disk or an optical disk. The storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0253] An exemplary storage medium is coupled to a processor, enabling the processor to read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an Application Specific Integrated Circuit (ASIC). Of course, the processor and the storage medium can also exist as discrete components in an electronic device or a master device.

[0254] It should be noted that in this text, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, clothing or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, clothing or device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, clothing or device including that element.

[0255] The serial numbers of the embodiments of the present application above are only for description and do not represent the superiority or inferiority of the embodiments.

[0256] Through the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions for causing a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods of the various embodiments of the present application.

[0257] In the technical solution of the present application, the processing of the collection, storage, use, processing, transmission, provision, and disclosure of user data and other information complies with the provisions of relevant laws and regulations and does not violate public order and good customs.

[0258] The above are only the preferred embodiments of the present application and do not limit the patent scope of the present application. Any equivalent structural or equivalent process transformation made by using the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present application.

Claims

1. A method for processing map data, characterized in that, Including: Obtain map data to be measured, where the map data to be measured includes multiple map elements to be measured; Determine a target real-scene image corresponding to the map data to be measured, where the target real-scene image includes at least one of the map elements to be measured; Perform vector difference processing on the map data to be measured based on map master database data to obtain a vector difference result, where the vector difference result includes a geometric matching relationship and / or attribute information between the map master database data and the map data to be measured; Detect the currency of the map data to be measured based on the vector difference result and the target real-scene image, and generate a detection result for the map data to be measured.

2. The method according to claim 1, wherein The determining the target real-scene image corresponding to the map data to be measured includes: Determine a target real-scene image corresponding to each map element to be measured in the map data to be measured according to a preset real-scene database, where the real-scene database includes sampling point trajectories of multiple roads and real-scene images corresponding to each sampling point in the sampling point trajectories.

3. The method according to claim 2, wherein The determining the target real-scene image corresponding to each map element to be measured in the map data to be measured according to the preset real-scene database includes: Determine the sampling point trajectory bound in the real-scene database for the target road to which each map element to be measured belongs; Project the center point of the map element to be measured onto the corresponding sampling point trajectory to obtain a target projection point; Select a target sampling point from the sampling point trajectory that is at a preset distance from the target projection point, and determine the current real-scene image corresponding to the target sampling point in the real-scene database as the target real-scene image.

4. The method according to claim 1, characterized in that, The performing vector difference processing on the map data to be measured based on map master database data to obtain a vector difference result includes: Perform geometric matching on the map element to be measured and a reference map element in the map master database data, form a matching pair by pairing the target map element with the largest overlap degree between the reference map element and the map element to be measured with the map element to be measured, obtain the geometric matching relationship between the map element to be measured and the reference map element, and record the attribute information of the matching pair, where the vector difference result includes the geometric matching relationship and the attribute information.

5. The method according to claim 4, characterized in that, The detecting the currency of the map data to be measured based on the vector difference result and the target real-scene image and generating a detection result for the map data to be measured includes: For a target matching pair formed by different road guiding identifiers in the vector difference result, determine the first semantic information of the first map element to be measured in the corresponding target real-scene image; If the first semantic information is the same as the semantic information of the first map element to be measured in the map data to be measured and is different from the semantic information corresponding to the target matching pair in the map master database data, then determine the first map element to be measured as accurate data in the detection result; and / or, determine the first map element to be measured as a map element with a changed real scene, and update the attribute information of the target matching pair according to the first semantic information.

6. The method according to claim 4, wherein The vector difference result includes second to-be-measured map elements that do not form the matching pairs with the reference map elements in the map master database data; Detecting the currency of the to-be-measured map data based on the vector difference result and the target real-scene image, and generating a detection result for the to-be-measured map data, including: For the second to-be-measured map element, determining second semantic information of the second to-be-measured map element in the corresponding target real-scene image; If the second semantic information is the same as the semantic information of the second to-be-measured map element in the to-be-measured map data, and the second semantic information does not exist in the map master database data, then determining the second to-be-measured map element as accurate data in the detection result; and / or, determining the second to-be-measured map element as a newly added map element in the real scene.

7. The method according to claim 4, characterized in that, The vector difference result includes first reference map elements in the map master database data that do not form the matching pairs with the to-be-measured map elements; detecting the currency of the to-be-measured map data based on the vector difference result and the target real-scene image, and generating a detection result for the to-be-measured map data, including: For the first reference map element, determining third semantic information of the first reference map element in the corresponding target real-scene image; If the third semantic information is different from the semantic information of the first reference map element in the map master database data, then determining that the to-be-measured map data has deleted the first reference map element in the detection result; If the third semantic information is the same as the semantic information of the first reference map element in the map master database data, then determining that the to-be-measured map data has omitted the first reference map element in the detection result.

8. A map data processing device, characterized in that, Including: An acquisition module, configured to acquire to-be-measured map data, where the to-be-measured map data includes multiple to-be-measured map elements; A determination module, configured to determine a target real-scene image corresponding to the to-be-measured map data, where the target real-scene image includes at least one of the to-be-measured map elements; A difference module, configured to perform vector difference processing on the to-be-measured map data based on the map master database data to obtain a vector difference result, where the vector difference result includes a geometric matching relationship and / or attribute information between the map master database data and the to-be-measured map data; A detection module, configured to detect the currency of the to-be-measured map data based on the vector difference result and the target real-scene image, and generate a detection result for the to-be-measured map data.

9. An electronic device, characterized in that, Including: At least one processor; And A memory communicatively connected to the at least one processor; Wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the electronic device to execute the method according to any one of claims 1-7.

10. A computer program product, characterized in that, Including a computer program, where when the computer program is executed by a processor, it implements the method according to any one of claims 1-7.

Images