Method and device for analyzing differences in high-precision maps, and storage medium
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- 合肥四维图新科技有限公司
- Filing Date
- 2021-09-30
- Publication Date
- 2026-06-26
Smart Images
Figure CN115908619B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of high-precision map technology, and in particular to a method, apparatus and storage medium for high-precision map difference analysis. Background Technology
[0002] In the field of autonomous driving, high-precision maps play a crucial role as providers of prior environmental information, particularly in high-precision positioning, assisted environmental perception, planning, and decision-making. In addition to the content of traditional maps, high-precision maps also need to include detailed road models, including lane models, road components, road attributes, and various other dynamic information.
[0003] Given the demands of autonomous driving for high-precision maps, these maps must possess higher absolute and relative coordinate accuracy, and contain richer and more detailed information elements. They must provide autonomous driving systems with high-freshness, high-precision, and multi-dimensional road and supplementary information. Consequently, high-precision maps present new challenges compared to traditional electronic maps in terms of manufacturing processes, quality, update cycles, and the richness of road attributes.
[0004] Currently, differential identifiers are generally used for differential analysis of multiple versions of high-precision maps. However, this method relies heavily on differential identifiers, has special requirements for map production processes, and has limitations. Therefore, existing technologies cannot fully and effectively achieve differential analysis of high-precision maps. Summary of the Invention
[0005] This application provides a method, apparatus, and storage medium for high-precision map difference analysis to overcome the problem that existing technologies cannot fully and effectively achieve high-precision map difference analysis.
[0006] In a first aspect, embodiments of this application provide a method for difference analysis of high-precision maps, including:
[0007] The corresponding map sheets in the first high-precision map and the second high-precision map are obtained respectively. The first high-precision map and the second high-precision map have been divided into multiple map sheets of the same size. Each map sheet includes road links and elements associated with the road links.
[0008] For each pair of corresponding map sheets, based on the road links in the map sheets and the elements associated with the road links, the map sheets corresponding to the first high-precision map and the map sheets corresponding to the second high-precision map are compared in sequence with high-precision map data of different levels and the element attributes of the elements to determine the difference results between the first high-precision map and the second high-precision map, which are used to correct or create high-precision maps.
[0009] Secondly, embodiments of this application provide a high-precision map difference analysis device, comprising:
[0010] The acquisition module is used to acquire corresponding map sheets in the first high-precision map and the second high-precision map respectively. The first high-precision map and the second high-precision map have been divided into multiple map sheets of the same size. Each map sheet includes road links and elements associated with the road links.
[0011] The difference analysis module is used to compare the map sheet corresponding to the first high-precision map and the map sheet corresponding to the second high-precision map with high-precision map data of different levels and the feature attributes of the features in turn, based on the road links in the map sheet and the features associated with the road links, for each pair of corresponding map sheets, to determine the difference results between the first high-precision map and the second high-precision map, so as to correct or create a high-precision map.
[0012] Thirdly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions. When a processor executes the computer-executable instructions, it implements the high-precision map difference analysis method described in the first aspect and various possible designs of the first aspect.
[0013] Fourthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the high-precision map difference analysis method as described in the first aspect and various possible designs of the first aspect.
[0014] The high-precision map difference analysis method, apparatus, and storage medium provided in this embodiment first acquire corresponding map sheets from a first high-precision map and a second high-precision map. The first and second high-precision maps have been divided into multiple map sheets of the same size. Each map sheet includes road links and elements associated with those road links. Then, for each pair of corresponding map sheets, based on the road links and associated elements within each map sheet, the map sheets corresponding to the first and second high-precision maps are compared sequentially with high-precision map data of different levels and the element attributes of the elements to determine the difference results between the first and second high-precision maps. This result is used to correct or create a high-precision map. By using a novel differential matching method for hierarchical matching, the special requirements of map production processes are eliminated. It does not rely heavily on differential identifiers for differential analysis. Furthermore, through hierarchical matching, such as element attribute comparison, changes can be proactively detected, enabling comprehensive and effective difference analysis of high-precision maps. This provides a more comprehensive error correction mechanism for high-precision maps, improving the accuracy of high-precision map production. Attached Figure Description
[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0016] Figure 1 A schematic diagram illustrating a scenario for the high-precision map difference analysis method provided in this application embodiment;
[0017] Figure 2 A flowchart illustrating the difference analysis method for high-precision maps provided in this application embodiment;
[0018] Figure 3 A flowchart illustrating a high-precision map difference analysis method provided in another embodiment of this application;
[0019] Figure 4 This is a schematic diagram of the drawings provided for the embodiments of this application;
[0020] Figure 5 A schematic diagram illustrating the association between LINK and elements provided in the embodiments of this application;
[0021] Figure 6 A schematic diagram of a scenario for a high-precision map difference analysis method provided in another embodiment of this application;
[0022] Figure 7 A schematic diagram of a scenario for a high-precision map difference analysis method provided in another embodiment of this application;
[0023] Figure 8 A schematic diagram of the structure of the high-precision map difference analysis device provided in the embodiments of this application;
[0024] Figure 9 A schematic diagram of the structure of the high-precision map difference analysis device provided in the embodiments of this application. Detailed Implementation
[0025] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0026] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0027] Currently, differential identifiers are generally used for differential analysis of multiple versions of high-precision maps. However, this method relies heavily on differential identifiers, has special requirements for map production processes, and has limitations. Therefore, existing technologies cannot fully and effectively achieve differential analysis of high-precision maps.
[0028] To address the problems of existing technologies, this application proposes a novel differential method for comparing two input maps, namely, hierarchical matching. The differential method employs attribute similarity, thus eliminating the need for special requirements on map production processes and eliminating reliance on differential identifiers. Instead, feature similarity replaces identifier-based differential analysis, enabling proactive detection of changes. Unlike existing methods that use differential identifiers to verify reliability, this application comprehensively and effectively achieves difference analysis of high-precision maps, thereby providing a more comprehensive error correction mechanism and improving the accuracy of high-precision map production.
[0029] In practical applications, see Figure 1 As shown, Figure 1 This is a schematic diagram illustrating a scenario for the high-precision map difference analysis method provided in this application embodiment. The executing entity in this embodiment can be a high-precision map difference analysis device, such as a terminal device, server, or processor. The terminal device here includes electronic devices with image and data analysis capabilities, such as fixed terminals, mobile terminals, and computer devices (e.g., all-in-one machines). Taking terminal device 10 as an example, two versions of maps, such as two high-precision maps (a first high-precision map and a second high-precision map), are acquired by acquisition device 20. Acquisition device 20 transmits the two high-precision maps to terminal device 10. Terminal device 10 performs a difference comparison on the two input maps (i.e., the high-precision maps) and outputs the difference result.
[0030] Specifically, firstly, the two high-precision maps are divided into multiple map sheets (MESH) of the same size (e.g., the same dimensions). Then, a new differential method is used to perform hierarchical matching of the two high-precision maps, combined with... Figure 2 The flowchart shown illustrates the difference analysis process for high-precision maps, including map sheet-level matching, road network matching, and lane-level matching. Lane-level matching further requires feature attribute matching. The final result is the difference between the two versions of the high-precision map. This difference analysis method (such as this differential method) does not heavily rely on differential identifiers for differentiation; instead, it uses feature attribute comparison to replace identifier differentiation. This proactively detects changes, unlike existing methods that use differential identifiers to verify the reliability of differential analysis. It can comprehensively and effectively achieve difference analysis of high-precision maps. Based on the difference results, it can provide a more comprehensive error correction mechanism for high-precision maps, improving the accuracy of high-precision map production.
[0031] The technical solutions of this application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.
[0032] Figure 3A flowchart illustrating a high-precision map difference analysis method provided in another embodiment of this application is shown. The method may include:
[0033] S101. Obtain the corresponding map sheets in the first high-precision map and the second high-precision map respectively. The first high-precision map and the second high-precision map have been divided into multiple map sheets of the same size.
[0034] Each map sheet includes road links and elements associated with those road links. The high-precision maps mentioned here, such as the first and / or second high-precision maps, can be finished maps or not, i.e., intermediate data, such as data used to create high-precision maps. For example, "same specifications" here can mean the same size; no specific limitation is made here.
[0035] In this embodiment, the execution entity can be a high-precision map difference analysis device. Two versions of high-precision maps, namely a first high-precision map and a second high-precision map, are acquired through a data acquisition device. The high-precision map difference analysis device then segments the two input high-precision maps, that is, the complete map is organized into multiple meshes (i.e., map sheets, see [link]) of the same size. Figure 4 (See the schematic diagram of the map sheet shown). Subsequent high-precision map difference analysis equipment performs differential processing on the map data in units of MESH to reduce the data processing scale.
[0036] Specifically, within a MESH, there are LINKs (referring to a road segment in the road network; a LINK is a marker representing a road during map creation, hereinafter referred to as a road link) and the geographical features (elements) associated with those LINKs. See [link to MESH]. Figure 5 The diagram illustrates the relationship between links and features. A link is associated with all point and line features extending outwards in both directions from the road area containing the link, such as... Figure 5 The example in the text shows a point feature and its associated link.
[0037] Within the same map version, the linkID (i.e., the identifier of the road link) of a LINK is permanent and unchanging, as is the ObjectID (i.e., the identifier of the feature). Each LINK and feature corresponds to a unique ID (i.e., an identity identifier). The corresponding map sheet, LINK, etc., can be selected between two versions of high-precision maps using the identity identifier.
[0038] S102. For each pair of corresponding map sheets, based on the road links in the map sheets and the elements associated with the road links, the map sheets corresponding to the first high-precision map and the map sheets corresponding to the second high-precision map are compared in sequence with high-precision map data of different levels and the element attributes of the elements to determine the difference results between the first high-precision map and the second high-precision map, which are used to correct or create high-precision maps.
[0039] In this embodiment, a novel differential method is used to perform hierarchical matching of two versions of high-precision maps combined with high-precision map data at different levels, such as map sheet level matching, road network matching, lane level matching, and feature attribute comparison. The final result is the differential result of the two versions of high-precision maps. This difference analysis method (such as this differential method) does not rely heavily on difference labels for differential analysis. Instead, it uses feature attribute comparison to replace label differential analysis, which can proactively detect changes. Unlike existing methods that use difference labels to verify the reliability of differential analysis, this method can comprehensively and effectively realize the difference analysis of high-precision maps. Based on the differential results, it can provide a more comprehensive error correction mechanism for high-precision maps, improving the accuracy of high-precision map production.
[0040] In one possible design, this embodiment provides a detailed description of S102 based on the above embodiment. According to the road links in the map sheet and the elements associated with those road links, the map sheet corresponding to the first high-precision map and the map sheet corresponding to the second high-precision map are sequentially compared with high-precision map data at different levels and the element attributes of the elements to determine the difference result between the first high-precision map and the second high-precision map. This can be achieved through the following steps:
[0041] Step a1: Based on the road links in the map sheet and the elements associated with the road links, perform map sheet matching comparison, road network matching comparison, and lane matching comparison sequentially on the map sheet corresponding to the first high-precision map and the map sheet corresponding to the second high-precision map to obtain the lane matching result.
[0042] Step a2: If the lane matching comparison is consistent, then perform attribute comparison on the elements that match the lane matching comparison. If there is an attribute comparison inconsistency, then output the attribute matching result. The attribute matching result is used to indicate the change of element attributes. If the attribute comparison is completely consistent, then output the attribute matching result as the element remains unchanged.
[0043] In this embodiment, the high-precision map data at different levels may include map sheet data, road network data, and lane data. Based on the road links in the map sheet and the elements associated with the road links, map sheet matching comparison, road network matching comparison, and lane matching comparison are performed sequentially by combining map sheet data, road network data, and lane data to obtain lane matching results.
[0044] Specifically, performing map sheet matching comparison, road network matching comparison, and lane matching comparison on the map sheet corresponding to the first high-precision map and the map sheet corresponding to the second high-precision map in sequence can be achieved through the following steps:
[0045] Step a11: Based on the road links in each map sheet and the elements associated with the road links, perform map sheet matching comparison between the map sheet corresponding to the first high-precision map and the map sheet corresponding to the second high-precision map. If the map sheet matching comparison is inconsistent, output the map sheet matching result. The map sheet matching result is used to indicate whether the map sheet is added or deleted.
[0046] Step a12: If the map sheet matching comparison is consistent, then perform road network matching comparison on the target map sheet that matches the map sheet matching comparison. If the road network matching comparison is inconsistent, then output the road network matching result. The road network matching result is used to indicate whether the road is added or deleted.
[0047] Step a13: If the road network matching comparison is consistent, then perform lane matching comparison on the road links that are consistent in the road network matching comparison. If the lane matching comparison is inconsistent, then output the lane matching result. The lane matching result is used to indicate whether the element is added or deleted.
[0048] In this embodiment, the map sheets divided from the two high-precision maps are first matched at the MESH level. If they are not completely consistent, the MESH-level difference results are output first; if they are consistent, the next step of road network matching is performed.
[0049] Specifically, the map is divided into meshes to reduce the data processing scale. Since differential processing is performed, two different versions of the map can be selected for comparison. The mesh lists of the first and second versions of the map are compared (see Table 1 below). If the two lists are not completely consistent, the difference results at the mesh level are output first. The difference results at the mesh level will output the results of all related elements within the entire mesh, indicating that all elements within the entire mesh do not match.
[0050] Table 1
[0051]
[0052]
[0053] If they match, proceed to the next step of road network matching. This involves selecting a MESH shared by the first map version (i.e., the first high-precision map) and the second map version (i.e., the second high-precision map), selecting a link of that MESH in the second map version (denoted as Link1), and performing road network matching with it in the first map version. If no match is found, the link and its associated features are deleted. Similarly, the link of the MESH in the first map version (denoted as Link2) is selected and matched with it in the second map version. If no match is found, the link and its associated features are added. Then, lane-level matching is performed on shared links. If a feature in the second map version matches successfully in the first map version, feature attribute comparison is performed. If any attribute is inconsistent, it indicates a change in the feature attribute, and the attribute change is output. If all attributes are consistent, it indicates the feature remains unchanged.
[0054] The high-precision map difference analysis method provided in this application acquires continuous frame images, extracts features from each frame to obtain image feature points for each frame, then detects dynamic feature points corresponding to the continuous frame images based on the image feature points of each frame, performs superpixel segmentation on each frame to determine dynamic target regions, and removes feature points of the dynamic regions from the image feature points of each frame based on the detected dynamic feature points and the dynamic target regions to obtain feature points of the static regions. These feature points of the static regions are used to provide a data source for camera localization and environmental mapping operations. By detecting and segmenting dynamic target regions and then removing them, the method is relatively simple to implement and can improve the localization accuracy of the binocular visual SLAM algorithm in dynamic environments, thereby enabling accurate localization and mapping results to be obtained using purely static environments.
[0055] In one possible design, based on the above embodiments, a detailed explanation of how to achieve road network matching is provided. The process of performing road network matching comparison on target map sheets that match the map sheet matching, and outputting the road network matching result if the road network matching comparison is inconsistent, can be achieved through the following steps:
[0056] Step b1: Obtain the map sheet that matches the map sheet matching comparison from the first high-precision map and the second high-precision map as the target map sheet, and obtain the map data corresponding to the target map sheet. The map data includes the location of each road link in the target map sheet.
[0057] Step b2: If the first high-precision map is used as a reference map, the target road link is obtained from the target map sheet in the second high-precision map, and according to the location of the target road link, all road links within the preset distance threshold range of the target road link are obtained in the first high-precision map to form a road link set.
[0058] Step b3: Based on the road link set and the target road link, determine whether the road network matching comparison is consistent. If they are inconsistent, determine that the road network matching result is that the target road link and the elements associated with the target road link are newly added roads.
[0059] Step b4: If the second high-precision map is used as a reference map, the target road link is obtained from the target map sheet in the first high-precision map, and according to the location of the target road link, all road links within the preset distance threshold range of the target road link are obtained in the second high-precision map to form a road link set.
[0060] Step b5: Based on the road link set and the target road link, determine whether the road network matching comparison is consistent. If they are inconsistent, determine that the road network matching result is that the target road link and the elements associated with the target road link are deleted.
[0061] The determination of whether the road network matching comparison is consistent based on the road link set and the target road link can be achieved through the following steps:
[0062] Step c1: Based on the topological relationship between the road links in the road link set, group the road links in the road link set to obtain multiple consecutive road link groups.
[0063] Step c2: For each road link group, calculate the intersection length and distance between the road link group and the target road link to obtain the optimal road link group.
[0064] Step c3: If the optimal road link group is empty, then the road network matching comparison is determined to be inconsistent.
[0065] In this embodiment, a MESH (as the target map sheet) shared by the first and second versions of the map is selected. If the first version of the map is used as the reference map, the Link of the MESH in the second version of the map is selected and denoted as Link1 (i.e., the target road link), and road network matching is performed with the first version of the map.
[0066] The matching method is as follows:
[0067] First, based on the absolute location of Link1, select all Links within a range of x meters (i.e., within a preset distance threshold) on the first version of the map, and denot this as LinkSet (i.e., road link set). Then, based on the topological relationships between Links, including direction and Link continuity, group the Links in LinkSet to obtain multiple consecutive Link groups.
[0068] Then, for multiple consecutive Link groups, calculate their intersection length and distance with Link1, and select the optimal Link group. For details, see [link to details]. Figure 6 , Figure 6 This is a schematic diagram illustrating a scenario for a high-precision map difference analysis method provided in another embodiment of this application. The analysis is as follows... Figure 6 Based on the permutations and combinations of all possible links within the vicinity of Link1 shown, and the continuity between link groups, a set of links can be found, for example:
[0069] {L1, L3, L4, L5, L6, L10, L11, L12}, {L2, L3, L4, L5, L6, L10, L13, L14}, etc. Based on the link groups in the above linkSet, the link group with the longest intersection with Link1, the shortest distance, and the closest topological result is selected as the final link group for road network matching output. Here, segment1, segment2, ..., segment6 represent segment 1, segment 2, ..., segment 6, respectively. The data structure of each segment can be represented as: <Yes, list<L1,L2> >; <None, L3>; <Yes, list<(L4, L5, L6), (L4, L5, L6)>>; <None, L10>; <None, list<(L11, L12), (L13, L14)>>. This data structure indicates whether each segment requires incremental selection in subsequent calculations. If it displays "Yes," the optimal option needs to be selected from the list; if it displays "None," incremental calculation is not required.
[0070] If the selected optimal Link group is empty, it means that Link1 and the land features associated with Link1 are newly added roads.
[0071] Conversely, the links from the first map version are matched against the links from the second map version, using the second map version as the reference map. If no match is found, it means that Link1 and its associated features have been deleted, indicating that the link and its associated features have been deleted. Finally, the road-level additions and deletions are output. For links that exist in both versions, lane-level matching is performed.
[0072] In one possible design, lane-level matching can be achieved through the following steps:
[0073] Step d1: Take any road link that matches the road network matching as a common road link, and extract feature attributes from the common road links in the first high-precision map and the second high-precision map respectively.
[0074] Step d2: Calculate the attribute similarity for each of the aforementioned feature attributes to obtain the similarity of the features corresponding to the first high-precision map and the second high-precision map in terms of the feature attributes.
[0075] Step d3: Compare the elements based on the similarity of each element in the element attributes, and calculate the similarity of the elements.
[0076] Step d4: Match the elements according to their similarity. If the preset conditions are not met, determine that the lane matching comparison is inconsistent and output the lane matching result.
[0077] In this embodiment, the feature attributes are first extracted: a common LINK is selected, and the associated land features (features) on the LINK of the two maps are selected, including but not limited to signs, markings, road edges, etc., as shown in Table 2. This embodiment uses signs and markings and the following attributes as examples for illustration, but the scope of protection of this application includes but is not limited to the following features and attributes.
[0078] Table 2
[0079]
[0080] Then, calculate the single-attribute similarity: In order to match two elements, a similarity function needs to be defined for each attribute (the scales of multiple attributes are different). This indicates the similarity of a certain element between two versions of the map in attribute i.
[0081] Specifically, for continuous attributes (where the attribute value is a continuous numerical value), such as line width, which can take values continuously within a certain range, a difference can be used to define it: in, This represents the value of attribute i for a specific element in the first map. This represents the value of attribute i for a certain feature in the second map. max represents the maximum value of this attribute for all features, mainly used to normalize the difference to 0 to 1.
[0082] For the attributes of three-dimensional coordinates, the coordinates can be split into three continuous attributes in the x, y, and z directions, and the attribute similarity can be calculated separately for each attribute.
[0083] For discrete attributes (where the possible values are finite, such as the shape of a sign, which can only be rectangle, circle, or triangle), a two-dimensional lookup table can be defined. Values in the lookup table are then assigned to different attributes. Similarity can be defined as follows:
[0084]
[0085] Here, x is a value between 0 and 0.5, and y is a value between 0.5 and 1. Different attributes have different values for x and y.
[0086] For character-type attributes, the similarity of strings is typically compared. Taking text on a sign as an example, the higher the similarity of the text on the sign, the higher the similarity of the character-type attribute; conversely, the lower the similarity of the text on the sign, the lower the similarity of the character-type attribute. Character-type attribute similarity can be defined using edit distance: edit distance represents the minimum number of edits required to transform one string into another. For two strings a and b, the formula for edit distance is defined as follows:
[0087]
[0088] Among them, lev a,b (i, j) represents the edit distance between two strings a and b; i and j represent the indices of strings a and b, respectively. Indices start from 1, for example, the edit distance between abel and adel is 1. Then, dividing the edit distance by the maximum length of the two strings reduces to 0-1, representing the string difference. Language differences may also affect the algorithm, as shown in the following formula:
[0089]
[0090] Here, `edit-dis(String1, String2)` refers to the edit distance between string1 and string2, and `max(LengthOfString1, LengthOfString2)` refers to the longest string length between string1 and string2. The above formula converts the edit distance between strings String1 and String2 into a matching probability value within the range of 0-1, used to measure the similarity between the two strings.
[0091] Then, feature similarity is calculated: for each feature in the second map, it is compared with the associated features in the same LINK in the first map, and a matching degree (attribute similarity) is calculated.
[0092]
[0093] Where i represents the attribute index. ω i This represents the weight of the attribute, which is defined based on the attribute's importance and precision. For example, an attribute with high importance, largely determining the similarity level, will be assigned a higher weight. Another example is a feature whose 3D coordinates have high precision in the x and y directions but poor precision in the z direction. To reduce the impact of the z-direction's precision, higher weights will be assigned to the x and y directions, while a lower weight will be given to the z-direction. This represents the similarity of features between the two map versions on attribute i. The algorithm is described in the single-attribute similarity function.
[0094] Finally, feature matching is performed: if the preset conditions are not met, feature matching is not performed, or the matching fails. These preset conditions can be no more than a distance threshold α and / or no more than a similarity threshold β.
[0095] Specifically, first, a threshold for absolute accuracy is set: a distance threshold α is set to represent the absolute GPS location accuracy of an element. Elements exceeding this threshold cannot be the same element, and no matching is performed. The absolute location accuracy is set based on statistical knowledge of map accuracy, which varies between different maps. Depending on the map's accuracy, three dimensions are considered: horizontal, vertical, and vertical. Furthermore, separate accuracy thresholds are set for point elements and linear elements. The calculation of absolute accuracy is as follows: Figure 7 As shown, the following examples illustrate point-like features and line-like features respectively:
[0096] like Figure 7 As shown, when calculating absolute precision, linear features only consider the horizontal distance, while point features require consideration of both horizontal and vertical distances. Because the precision of the horizontal and vertical distances differs, the calculated distance is not an exact, absolute point-to-point distance.
[0097] Then set a similarity threshold: set a similarity threshold β, which represents the minimum similarity between elements. If the similarity is less than this threshold, they cannot be the same element and no matching will be performed.
[0098] Feature matching: For each feature in the second version of the map, calculate the similarity with features within the absolute precision threshold range in the first version of the map. If the similarity is less than β, they are not the same feature and no matching is performed. Otherwise, select the feature with the highest similarity from the first version of the map, indicating a successful match.
[0099] If a feature in the second version of the map does not match in the first version, it is considered newly added. Conversely, if a feature in the first version of the map does not match in the second version, it is considered deleted. If other features match successfully, proceed to the next step: determine attribute changes (i.e., feature attribute comparison).
[0100] Specifically, if the lane matching comparison is consistent, attribute comparison is performed on the elements that match the lane matching comparison. This can be achieved through the following steps:
[0101] Step e1: If the feature in the second high-precision map is successfully matched in the first high-precision map, then it is determined that the lane matching comparison is consistent, and it is determined whether the attributes of the feature in the second high-precision map are consistent with those of the feature in the first high-precision map.
[0102] Step e2: If the features in the first high-precision map are successfully matched in the second high-precision map, then it is determined that the lane matching comparison is consistent, and it is determined whether the attributes of the features in the first high-precision map and the features in the second high-precision map are consistent.
[0103] In this embodiment, if a feature in the second version of the map matches successfully in the first version of the map, the attributes of the two versions are determined. If any attribute is inconsistent, it indicates that the feature attribute has changed, and the change in feature attribute is output; if all attributes are consistent, it indicates that the feature has not changed. Similarly, if a feature in the first version of the map matches successfully in the second version of the map, the attributes of the two versions are determined. If any attribute is inconsistent, it indicates that the feature attribute has changed, and the change in feature attribute is output; if all attributes are consistent, it indicates that the feature has not changed.
[0104] The difference results are output through hierarchical matching, as shown in Table 3.
[0105] Table 3
[0106]
[0107] This application primarily utilizes attribute similarity to perform differencing on elements with the same ID, thus reducing reliance on differencing identifiers and enabling differencing between any two versions of ordinary maps. Specifically, since the unchanged elements in the differencing between two maps maintain ID consistency, matching and differencing elements with the same ID can be performed by comparing attribute similarity. This approach has low requirements for map production technology, and differencing can be performed on any two maps without differencing identifiers, achieving comprehensive and effective difference analysis. Difference allows for checking whether these changes and invariants meet expectations, thereby improving map quality and verifying the accuracy of the tool.
[0108] Therefore, this application, by comparing the old and new versions of the map, can proactively identify potential problems in the entire high-precision map production process, revealing both unchanged and changed elements. For unchanged elements, ID consistency can be maintained; for changed elements, subsequent checks can be conducted to ensure these changes meet expectations. This proactive problem-solving approach improves map production accuracy, does not heavily rely on differential identifiers for differential mapping, and places lower demands on map processing technology.
[0109] To implement the aforementioned high-precision map difference analysis method, this embodiment provides a high-precision map difference analysis device. (See also...) Figure 8 , Figure 8 This is a schematic diagram of the structure of a high-precision map difference analysis device provided in an embodiment of this application. The high-precision map difference analysis device 80 includes: an acquisition module 801 and a difference analysis module 802. The acquisition module 801 is used to acquire corresponding map sheets in a first high-precision map and a second high-precision map, respectively. The first high-precision map and the second high-precision map have been divided into multiple map sheets of the same size. Each map sheet includes road links and elements associated with the road links. The difference analysis module 802 is used to compare the map sheets corresponding to the first high-precision map and the map sheets corresponding to the second high-precision map with high-precision map data of different levels and the element attributes of the elements in sequence, for each pair of corresponding map sheets, based on the road links and elements associated with the road links in the map sheets, to determine the difference results between the first high-precision map and the second high-precision map, so as to correct or create a high-precision map.
[0110] In this embodiment, an acquisition module 801 and a difference analysis module 802 are configured to first acquire corresponding map sheets from a first high-precision map and a second high-precision map, respectively. The first and second high-precision maps have been divided into multiple map sheets of the same size. Each map sheet includes road links and elements associated with those road links. Then, for each pair of corresponding map sheets, based on the road links and associated elements within each map sheet, the map sheets corresponding to the first and second high-precision maps are compared sequentially with high-precision map data from different levels and the element attributes of the elements to determine the difference results between the first and second high-precision maps. This result is used to correct or create a high-precision map. This novel difference method for hierarchical matching eliminates the special requirements of map production processes and does not rely heavily on difference identifiers. Furthermore, hierarchical matching, such as element attribute comparison, can proactively detect changes, enabling comprehensive and effective difference analysis of high-precision maps. This provides a more comprehensive error correction mechanism for high-precision maps, improving the accuracy of high-precision map production.
[0111] The apparatus provided in this embodiment can be used to execute the technical solutions of the above method embodiments. Its implementation principle and technical effects are similar, and will not be described again here.
[0112] In one possible design, the difference result includes attribute matching results; the difference analysis module includes a first analysis unit and a second analysis unit; the first analysis unit is used to perform map sheet matching comparison, road network matching comparison, and lane matching comparison sequentially on the map sheet corresponding to the first high-precision map and the map sheet corresponding to the second high-precision map, based on the road links in the map sheet and the elements associated with the road links, to obtain lane matching results; the second analysis unit is used to perform attribute comparison on the elements that match the lane matching comparison when the lane matching comparison is consistent; if there is an inconsistency in the attribute comparison, the attribute matching result is output, which is used to indicate the change of element attributes; if the attribute comparison is completely consistent, the attribute matching result is output as the element remains unchanged.
[0113] In one possible design, the first analysis unit includes a first analysis subunit, a second analysis subunit, and a third analysis subunit. The first analysis subunit is used to perform map sheet matching comparison between the map sheet corresponding to the first high-precision map and the map sheet corresponding to the second high-precision map, based on the road links and associated elements in each map sheet. If the map sheet matching comparison is inconsistent, a map sheet matching result is output, indicating whether the map sheet is newly added or deleted. The second analysis subunit is used to perform road network matching comparison on the target map sheets with consistent map sheet matching when the map sheet matching comparison is consistent. If the road network matching comparison is inconsistent, a road network matching result is output, indicating whether the road is newly added or deleted. The third analysis subunit is used to perform lane matching comparison on the road links with consistent road network matching when the road network matching comparison is consistent. If the lane matching comparison is inconsistent, a lane matching result is output, indicating whether the element is newly added or deleted.
[0114] In one possible design, the second analysis subunit is specifically used to: obtain a map sheet that matches the map sheet comparison from the first high-precision map and the second high-precision map as the target map sheet, and obtain map data corresponding to the target map sheet, wherein the map data includes the location of each road link in the target map sheet;
[0115] If the first high-precision map is used as a reference map, the target road link is obtained from the target map sheet in the second high-precision map, and according to the location of the target road link, all road links within the preset distance threshold range of the target road link are obtained in the first high-precision map to form a road link set;
[0116] Based on the road link set and the target road link, determine whether the road network matching comparison is consistent. If they are inconsistent, determine that the road network matching result is that the target road link and the elements associated with the target road link are newly added roads.
[0117] Accordingly, if the second high-precision map is used as a reference map, then based on the road link set and the target road link, it is determined whether the road network matching comparison is consistent. If they are inconsistent, then the road network matching result is determined to be the target road link and the elements associated with the target road link, which are road deletions.
[0118] In one possible design, the second analysis subunit is specifically used to: group the road links in the road link set according to the topological relationship between the road links in the road link set, to obtain multiple consecutive road link groups;
[0119] For each road link group, the intersection length and distance between the road link group and the target road link are calculated to obtain the optimal road link group;
[0120] If the optimal road link group is empty, then the road network matching comparison is determined to be inconsistent. In one possible design, the third analysis subunit is specifically used for:
[0121] Any road link that matches the road network is taken as a common road link, and feature attributes are extracted from the common road links in the first high-precision map and the second high-precision map respectively.
[0122] Attribute similarity calculation is performed on each of the aforementioned element attributes to obtain the similarity of the elements corresponding to the first high-precision map and the second high-precision map on the aforementioned element attributes;
[0123] Based on the similarity of each element in the element attributes, the elements are compared and the similarity of the elements is calculated.
[0124] Based on the similarity of the elements, the elements are matched. If the preset conditions are not met, the lane matching comparison is determined to be inconsistent, and the lane matching result is output.
[0125] In one possible design, the second analysis unit is specifically used for:
[0126] If the feature in the second high-precision map matches successfully in the first high-precision map, then it is determined that the lane matching comparison is consistent, and it is determined whether the attributes of the feature in the second high-precision map are consistent with those of the feature in the first high-precision map.
[0127] Correspondingly, if the features in the first high-precision map are successfully matched in the second high-precision map, it is determined that the lane matching comparison is consistent, and it is determined whether the attributes of the features in the first high-precision map and the features in the second high-precision map are consistent.
[0128] To implement the aforementioned high-precision map difference analysis method, this embodiment provides a high-precision map difference analysis device. Figure 9 A schematic diagram of the structure of the high-precision map difference analysis device provided in this application embodiment. Figure 9 As shown, the high-precision map difference analysis device 90 of this embodiment includes a processor 901 and a memory 902; wherein, the memory 902 is used to store computer execution instructions; the processor 901 is used to execute the computer execution instructions stored in the memory to implement the various steps performed in the above embodiment. For details, please refer to the relevant descriptions in the foregoing method embodiments.
[0129] This application also provides a computer-readable storage medium storing computer-executable instructions. When a processor executes the computer-executable instructions, it implements the high-precision map difference analysis method described above.
[0130] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the high-precision map difference analysis method described above.
[0131] In the several embodiments provided in this 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 instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms. Additionally, the functional modules in the various embodiments of this application may be integrated into one processing unit, or each module may exist physically separately, or two or more modules may be integrated into one unit. The above-mentioned modular units can be implemented in hardware or in the form of hardware plus software functional units.
[0132] The integrated modules implemented as software functional modules described above can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods described in the various embodiments of this application. It should be understood that the processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.
[0133] The memory may include high-speed RAM, and may also include non-volatile memory (NVM), such as at least one disk drive, and may also be a USB flash drive, external hard drive, read-only memory, disk, or optical disc. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses in the accompanying drawings are not limited to a single bus or a single type of bus. The aforementioned storage medium 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 storage, flash memory, disk, or optical disc. The storage medium can be any available medium accessible to general-purpose or special-purpose computers.
[0134] An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Alternatively, the storage medium can be an integral part of the processor. Both the processor and the storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and storage medium can exist as discrete components in an electronic device or host device.
[0135] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.
[0136] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.
Claims
1. A method for difference analysis of high-precision maps, characterized in that, include: The corresponding map sheets in the first high-precision map and the second high-precision map are obtained respectively. The first high-precision map and the second high-precision map have been divided into multiple map sheets of the same size. Each map sheet includes road links and elements associated with the road links. For each pair of corresponding map sheets, based on the road links in the map sheets and the elements associated with the road links, map sheet matching comparison, road network matching comparison, and lane matching comparison are performed sequentially on the map sheet corresponding to the first high-precision map and the map sheet corresponding to the second high-precision map. When the lane matching comparison is consistent, the lane matching result is obtained. If the lane matching comparison is consistent, attribute comparison is performed on the elements with consistent lane matching to determine the difference results between the first high-precision map and the second high-precision map, which are used to correct or create a high-precision map. Specifically, a map sheet with consistent map sheet matching is obtained from the first and second high-precision maps as the target map sheet. The road network matching comparison includes comparing the intersection length and distance of the target road link in the target map sheet of the second high-precision map with each road link group in the road link set of the first high-precision map to obtain the optimal road link group. If the optimal road link group is empty, it is determined that the road network matching comparison is inconsistent. The road link set is formed by obtaining all road links within a preset distance threshold range of the target road link in the first high-precision map based on the location of the target road link. The road link group is obtained by grouping the road links in the road link set according to the topological relationship between the road links in the road link set.
2. The method according to claim 1, characterized in that, The method of performing attribute comparison on elements that match lane matching includes: if there are inconsistencies in attribute comparison, outputting attribute matching results, which are used to indicate changes in element attributes; if the attribute comparisons are completely consistent, outputting attribute matching results indicating that the element has not changed.
3. The method according to claim 2, characterized in that, The differential results also include at least one of map sheet matching results, road network matching results, and lane matching results; Based on the road links in the map sheet and the elements associated with the road links, map sheet matching comparison, road network matching comparison, and lane matching comparison are performed sequentially on the map sheet corresponding to the first high-precision map and the map sheet corresponding to the second high-precision map to obtain lane matching results, including: Based on the road links in each map sheet and the elements associated with the road links, a map sheet matching comparison is performed between the map sheet corresponding to the first high-precision map and the map sheet corresponding to the second high-precision map. If the map sheet matching comparison is inconsistent, a map sheet matching result is output. The map sheet matching result is used to indicate whether the map sheet is added or deleted. If the map sheet matching comparison is consistent, then the road network matching comparison is performed on the target map sheet that matches the map sheet matching comparison. If the road network matching comparison is inconsistent, then the road network matching result is output. The road network matching result is used to indicate whether the road is added or deleted. If the road network matching comparison is consistent, then lane matching comparison is performed on the road links that are consistent with the road network matching comparison. If the lane matching comparison is inconsistent, then the lane matching result is output. The lane matching result is used to indicate whether the element is added or deleted.
4. The method according to claim 3, characterized in that, The method involves performing road network matching comparisons on target map sheets that match the map sheet matching. If the road network matching comparisons are inconsistent, the road network matching results are output, including: The target map sheet is obtained by matching and comparing the map sheet from the first high-precision map and the second high-precision map, and the map data corresponding to the target map sheet is obtained, including the location of each road link in the target map sheet; If the first high-precision map is used as a reference map, the target road link is obtained from the target map sheet in the second high-precision map, and according to the location of the target road link, all road links within the preset distance threshold range of the target road link are obtained in the first high-precision map to form a road link set; The road network matching comparison is determined based on the road link set and the target road link. If they are inconsistent, the road network matching result is determined to be that the target road link and the elements associated with the target road link are newly added roads. Accordingly, if the second high-precision map is used as a reference map, then based on the road link set and the target road link, it is determined whether the road network matching comparison is consistent. If they are inconsistent, then the road network matching result is determined to be the target road link and the elements associated with the target road link, which are road deletions.
5. The method according to claim 4, characterized in that, The intersection length and distance of the target road links in the target map sheet in the second high-precision map are compared with each road link group in the road link set in the first high-precision map to obtain the optimal road link group. If the optimal road link group is empty, then the road network matching comparison is determined to be inconsistent, including: Based on the topological relationship between the road links in the road link set, the road links in the road link set are grouped to obtain multiple consecutive road link groups; For each road link group, the intersection length and distance between the road link group and the target road link are calculated to obtain the optimal road link group; If the optimal road link group is empty, then the road network matching comparison is determined to be inconsistent.
6. The method according to any one of claims 3-5, characterized in that, For target road links that match the road network matching, perform lane matching comparison. If the lane matching comparison does not match, output the lane matching result, including: Any road link that matches the road network is taken as a common road link, and feature attributes are extracted from the common road links in the first high-precision map and the second high-precision map respectively. Attribute similarity calculation is performed on each of the aforementioned element attributes to obtain the similarity of the elements corresponding to the first high-precision map and the second high-precision map on the aforementioned element attributes; Based on the similarity of each element in the element attributes, the elements are compared and the similarity of the elements is calculated. Based on the similarity of the elements, the elements are matched. If the preset conditions are not met, the lane matching comparison is determined to be inconsistent, and the lane matching result is output.
7. The method according to any one of claims 2-5, characterized in that, If the lane matching matches, then attribute comparison is performed on the elements that match the lane matching, including: If the feature in the second high-precision map matches successfully in the first high-precision map, then it is determined that the lane matching comparison is consistent, and it is determined whether the attributes of the feature in the second high-precision map are consistent with those of the feature in the first high-precision map. Correspondingly, if the features in the first high-precision map are successfully matched in the second high-precision map, it is determined that the lane matching comparison is consistent, and it is determined whether the attributes of the features in the first high-precision map and the features in the second high-precision map are consistent.
8. A high-precision map difference analysis device, characterized in that, include: The acquisition module is used to acquire corresponding map sheets in the first high-precision map and the second high-precision map respectively. The first high-precision map and the second high-precision map have been divided into multiple map sheets of the same size. Each map sheet includes road links and elements associated with the road links. The difference analysis module is used to perform map sheet matching comparison, road network matching comparison, and lane matching comparison on the map sheet corresponding to the first high-precision map and the map sheet corresponding to the second high-precision map in sequence, based on the road links in the map sheet and the elements associated with the road links, for each pair of corresponding map sheets, and to obtain the lane matching result when the lane matching comparison is consistent; If the lane matching comparison is consistent, attribute comparison is performed on the elements with consistent lane matching to determine the difference results between the first high-precision map and the second high-precision map, which are used to correct or create a high-precision map. Specifically, a map sheet with consistent map sheet matching is obtained from the first and second high-precision maps as the target map sheet. The road network matching comparison includes comparing the intersection length and distance of the target road link in the target map sheet of the second high-precision map with each road link group in the road link set of the first high-precision map to obtain the optimal road link group. If the optimal road link group is empty, it is determined that the road network matching comparison is inconsistent. The road link set is formed by obtaining all road links within a preset distance threshold range of the target road link in the first high-precision map based on the location of the target road link. The road link group is obtained by grouping the road links in the road link set according to the topological relationship between the road links in the road link set.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by the processor, implement the high-precision map difference analysis method as described in any one of claims 1 to 7.
10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the difference analysis method for high-precision maps as described in any one of claims 1 to 7.