Map road change detection update method based on three-party correlation protection state machine

Abstract

The application relates to the technical field of high-precision map data processing, and discloses a map road change detection updating method based on a three-party correlation protection state machine, which comprises the following steps: converting and normalizing first data, second data and third data to obtain first, second and third standardized objects; extracting geometric topological features to obtain first, second and third geometric topological features; performing candidate retrieval to obtain a candidate object set; performing multi-feature correlation matching and conflict resolution processing on the multi-feature correlation matching result to obtain a one-to-one matching set; performing change identification processing according to the one-to-one matching set to obtain an original change relation; and performing combination decision to generate an incremental updating instruction, performing incremental updating on the third data, and obtaining an incremental updating result. The application can avoid covering artificial editing results and reduce the risk of false correlation.

Description

Technical Field

[0001] This invention relates to the field of high-precision map data processing technology, and more specifically, to a map road change detection and update method based on a three-party association protection state machine. Background Technology

[0002] High-precision map road data is characterized by multiple feature levels, dense topological relationships, high geometric accuracy, frequent version updates, and high costs for manual refinement. Version updates typically require identifying additions, deletions, attribute changes, geometric changes, local changes, breaks and splits, merging, and directional changes, while ensuring the continuity of topological relationships between roads, lanes, and their associated facilities. Existing road network matching schemes generally use buffers, length, distance, angle, shape, or local network structures to match two versions of data, but their processing objects and update decisions still have different focuses.

[0003] For example, Chinese patent CN113032512B proposes a method, apparatus, device, storage medium, and product for updating road network data. It constructs new and old traffic road networks based on physical feature nodes, their locations, and ordered line segment connections, and establishes a mapping relationship between the two versions based on the similarity of feature nodes. This scheme can reduce the accumulated error of version mapping, but it mainly processes two sets of data: the original version and the updated version.

[0004] Chinese patent CN114064835B proposes a multi-source vector road network update method and electronic device based on change point detection. It determines change points based on the distance function and derivative between updated road segment nodes and historical road segments, trims road segments, and classifies trimmed segments as unmatched, replaceable, or requiring modification. This scheme is suitable for locating local change intervals, but it still replaces historical road segments with updated ones when similarity conditions are met, failing to consider manually edited results as an independent decision input.

[0005] Chinese patent CN110006439B discloses a method, apparatus, server, and storage medium for matching map trajectory data. Based on differential roads in the old and new road networks, historical trajectory matching results, and adjacent constrained trajectory segments, it performs incremental rematching on the differential trajectory segments to reduce redundant trajectory matching and maintain road sequence connectivity. The updated object of this scheme is the trajectory matching result, rather than the road geometry and manually edited results in the high-precision map production library.

[0006] Chinese patent CN115757672A discloses a method and system for updating high-precision map data. It obtains differences in roads and road facilities from old and new high-precision map road network data, establishes a difference list, and performs downloading, modification, or deletion based on point cloud, image, or GRB data. This solution provides a workflow for producing high-precision map difference items, but it does not establish a three-way relationship between the old original data, the new original data, and the manually edited database.

[0007] The above solutions address the issues of dual-version road network mapping, local change point location, incremental trajectory re-matching, or high-precision map difference list production, but none of them solve the problem of traceable updates when original version changes and manual refinement results coexist in a high-precision map production environment.

[0008] Specifically: When the original data is reshaped, split, or merged in the new version, directly replacing it with only the geometric similarity or change point results may overwrite the lane-level manual reshaping that has already been completed in the editing library. Relying solely on link_id or object_id inheritance will result in the loss of corresponding relationships when the identifier is recoded, broken, or merged. Therefore, it is necessary to calculate the original version change relationship and the manual editing state separately, and then use a state machine to combine and output the final geometric source, maintenance operations, and review requirements.

[0009] There are currently no effective solutions to the problems in the relevant technologies. Summary of the Invention

[0010] To address the problems in related technologies, this invention proposes a map road change detection and update method based on a three-party associated protection state machine, in order to overcome the aforementioned technical problems existing in the existing related technologies.

[0011] Therefore, the specific technical solution adopted by the present invention is as follows: A map road change detection and update method based on a three-party association protection state machine includes: S1. Perform projection coordinate transformation and normalization on the acquired first data, second data and third data to obtain a first normalized object, a second normalized object and a third normalized object in sequence; wherein, the first data is the updated original map road data, the second data is the original map road data, and the third data is the manually edited library road data; S2. Geometric features are extracted from the first standardized object, the second standardized object, and the third standardized object respectively to obtain the first geometric topological feature, the second geometric topological feature, and the third geometric topological feature in sequence; wherein, the geometric topological feature includes spatial geometric features and topological correlation features; S3. Based on the second standardized object, the third standardized object, the second geometric topology feature, and the third geometric topology feature, perform editing association processing to obtain the editing association relationship; according to the editing association relationship, calculate the geometric deviation, geometric coverage, and attribute difference between the mutually related second standardized object and the third standardized object, and determine the editing protection status based on the geometric deviation, geometric coverage, and attribute difference; S4. Construct a spatial index based on the second geometric topological feature, and use the first geometric topological feature to perform candidate retrieval in the spatial index to obtain a set of candidate objects corresponding to the first standardized object. S5. Based on the first geometric topological feature and the second geometric topological feature, perform multi-feature association matching on the first standardized object and the candidate objects in the candidate object set, and perform conflict resolution processing on the multi-feature association matching results to obtain a one-to-one matching set between the first standardized object and the second standardized object. S6. Based on the one-to-one matching set, perform consistency analysis and change identification processing on the first standardized object and the second standardized object to obtain the original change relationship; S7. Determine the compatibility results based on the one-to-one matching set and editing association, and make a combined decision based on the original change relationship, editing protection status and compatibility results to obtain the target geometry source, maintenance opcode and review mark; S8. Generate incremental update instructions based on the target geometry source, maintenance opcode, and verification mark, and perform incremental update on the third data according to the incremental update instructions to obtain the incremental update result.

[0012] The beneficial effects of this invention are as follows: 1. By calculating the original version association channel between the new version of the original data and the old version of the original data, and the editing protection channel between the old version of the original data and the manually edited database data separately, and then converging them in the combined decision-making stage, the original version changes and the manual editing status become independent decision-making bases. This helps to avoid overwriting the manual editing results by directly updating based solely on geometric similarity or object identifiers.

[0013] 2. By unifying projection coordinates, extracting geometric topological features, retrieving candidate spatial indexes, anchoring signs, topological gating, geometric scoring, and resolving deterministic conflicts, the candidate range can be narrowed down and a repeatable one-to-one matching set can be formed, reducing the risk of erroneous associations in intersections, parallel roads, and signage change scenarios.

[0014] 3. By combining the original change relationships, edit protection status, and compatibility results, the source of the target geometry, maintenance operation codes, and review markers are determined, so that compatible manually edited geometry is preserved, and incompatible cases are brought into the review process.

[0015] 4. By recording the target data layer and evidence identifier in the incremental update instruction, the candidate retrieval, association matching, editing protection status and combined decision path can be associated and traced, and it is convenient for the maintenance system to distinguish between automatically executed instructions and instructions pending review. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a map road change detection and update method based on a three-party association protection state machine according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the combined decision-making process of the original change relationship and the edit protection state in the map road change detection and update method based on the three-party association protection state machine according to an embodiment of the present invention. Detailed Implementation

[0018] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention. The components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0019] According to an embodiment of the present invention, a map road change detection and update method based on a three-party associated protection state machine is provided.

[0020] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figure 1 and Figure 2 As shown, the map road change detection and update method based on a three-party association protection state machine according to an embodiment of the present invention includes: S1. Perform projection coordinate transformation and normalization on the acquired first data, second data and third data to obtain a first normalized object, a second normalized object and a third normalized object in sequence; wherein, the first data is the updated original map road data, the second data is the original map road data, and the third data is the manually edited library road data; S2. Geometric features are extracted from the first standardized object, the second standardized object, and the third standardized object respectively to obtain the first geometric topological feature, the second geometric topological feature, and the third geometric topological feature in sequence; wherein, the geometric topological feature includes spatial geometric features and topological correlation features; S3. Based on the second standardized object, the third standardized object, the second geometric topology feature, and the third geometric topology feature, perform editing association processing to obtain the editing association relationship; according to the editing association relationship, calculate the geometric deviation, geometric coverage, and attribute difference between the mutually related second standardized object and the third standardized object, and determine the editing protection status based on the geometric deviation, geometric coverage, and attribute difference; S4. Construct a spatial index based on the second geometric topological feature, and use the first geometric topological feature to perform candidate retrieval in the spatial index to obtain a set of candidate objects corresponding to the first standardized object. S5. Based on the first geometric topological feature and the second geometric topological feature, perform multi-feature association matching on the first standardized object and the candidate objects in the candidate object set, and perform conflict resolution processing on the multi-feature association matching results to obtain a one-to-one matching set between the first standardized object and the second standardized object. S6. Based on the one-to-one matching set, perform consistency analysis and change identification processing on the first standardized object and the second standardized object to obtain the original change relationship; S7. Determine the compatibility results based on the one-to-one matching set and editing association, and make a combined decision based on the original change relationship, editing protection status and compatibility results to obtain the target geometry source, maintenance opcode and review mark; S8. Generate incremental update instructions based on the target geometry source, maintenance opcode, and verification mark, and perform incremental update on the third data according to the incremental update instructions to obtain the incremental update result.

[0021] In a preferred embodiment, S1 includes: S11. Read the road records in the first data, second data and third data respectively, obtain the geometric information of each road record, and keep the object identifier, feature type, traffic direction, topological association information, parent association identifier, version field and attribute field participating in attribute difference of each road record as reserved fields to obtain the road record set. S12. Read the source coordinate reference system corresponding to each road record in the road record set and verify it. Based on the verified source coordinate reference system and the target projection coordinate reference system, establish a coordinate transformation relationship. Transform the horizontal and vertical coordinates according to the original arrangement order of the geometric vertices, retain the original geometry and generate the projection working geometry. S13. Perform vertex deduplication on the projection working geometry, and verify the coordinate values, number of vertices and projection length. Road records that pass the verification are considered valid road records. S14. Based on the topological association information of the valid road records, determine the starting point connection direction and the ending point connection direction; for road records with unidirectional traffic and projection working geometry opposite to the connection direction, reverse the coordinate sequence to obtain the normalized projection working geometry; S15. Generate internal record identifiers for the valid road records corresponding to the first data, the second data, and the third data respectively, and generate standardized records based on the internal record identifiers, reserved fields, original geometry, and normalized projection working geometry to obtain the first standardized object, the second standardized object, and the third standardized object.

[0022] It should be noted that, in this embodiment, the first data is the updated original map road data, that is, the new version of the original high-precision map road data. The second data is the original map road data, that is, the old version of the original high-precision map road data. The third data source is manually edited road data. The first, second, and third standardized objects are road objects obtained by normalizing the corresponding data. A road object is a line element (Link) with stable topological meaning, which can correspond to equivalent objects such as road centerlines, lane centerlines, lane boundaries, or HERE Links. Each road record includes at least an object identifier code. or ), feature types ( Geometric information geometry ), traffic direction ( ), topological association information ( zlevels or ), parent association identifier ( or ), version field, and attribute fields participating in attribute difference.

[0023] Specifically, the source coordinate reference system is read from the .prj file of the Shaprfile corresponding to the first, second, and third data, the database metadata, or the configuration table, respectively. If the coordinate reference system is missing, the axis order conflicts, or the coordinate range verification fails, the corresponding record will be written to the exception record. Then it is transferred to manual review, without guessing the coordinate reference system; for data that passes the verification, it is based on the source coordinate reference system. Establish a coordinate transformation relationship with the target projection coordinate reference system in meters.

[0024] In this embodiment, the target projection coordinate reference system adopts EPSG:4547. During coordinate transformation, the x-coordinates are transformed according to the original vertex order. x and ordinate y While preserving the original geometry and projection working geometry .

[0025] After coordinate transformation, delete consecutive duplicate vertices with a spacing of less than 1 mm on the calculated copy of the projected working geometry, and verify whether the coordinates are finite values, whether the line object contains at least two different vertices, and whether the projected length is not less than 2 m. Records with empty geometry, failed transformation, insufficient vertices, or length less than 2 m are excluded from the current matching working set, while their source data is retained and the object identifier, original geometry, and reason for the exception are recorded.

[0026] The start and end connection directions of each road record are identified based on the start and end connection identifiers in the topological association information. When the traffic direction is unidirectional and the start and end point order of the projected working geometry is opposite to the direction indicated by the topological relationship, the coordinate sequence of the projected working geometry is reversed. When the traffic direction is bidirectional, the original geometric direction is retained, and an undirected angle is used in subsequent direction comparisons. Finally, a standardized record containing internal record identifiers, reserved fields, original geometry, and projected working geometry is generated. Record : .

[0027] The first standardized object, the second standardized object, and the third standardized object are formed respectively.

[0028] In a preferred embodiment, S2 includes: S21. Calculate the midpoint of the arc length, the object length, and the minimum bounding rectangle of the first, second, and third standardized objects respectively, and generate local direction vectors based on the object starting point and local sampling points determined along the object geometry according to the preset arc length to obtain spatial geometric features; S22. Obtain the travel direction, the set of associated objects at the starting point, and the set of associated objects at the ending point of the first, second, and third standardized objects. Determine the starting point connectivity and the ending point connectivity based on the set of associated objects at the starting point and the set of associated objects at the ending point to obtain the topological association features. S23. Combine the spatial geometric features and topological association features of the first standardized object, the second standardized object, and the third standardized object respectively to obtain the first geometric topological feature, the second geometric topological feature, and the third geometric topological feature.

[0029] It should be noted that the directional adjacency signature consists of the direction of travel, the starting point connectivity, the ending point connectivity, the set of associated objects at the starting point, and the set of associated objects at the ending point.

[0030] Specifically, geometric feature quantization extraction is performed on the first standardized object, the second standardized object, and the third standardized object, respectively. Let any road object... The projective working geometry is The starting point is The destination is The object length is .make Indicates from geometry The starting point moves along the arc length. The obtained point, let Representing geometry The smallest bounding rectangle.

[0031] The object length, arc midpoint, and local direction vector are determined by the following formulas: ; ; ; In the formula, for The cumulative arc length, To reach along the cumulative arc length Location point, For local directional sampling distances, 20m and 0.2 are used. The smaller value in Let be the unit direction vector pointing from the starting point to the local sampling point. The first i The x and y coordinates of the starting point of the road object. The first i The horizontal and vertical coordinates of the endpoint of the road object.

[0032] From topological association information ( zlevels or Retrieves the set of object identifiers connected to the object's origin. and the set of object identifiers connected to the endpoint And determine the starting connectivity. Connectivity with endpoint .

[0033] The travel direction, starting point connectivity, ending point connectivity, set of associated objects at the starting point, and set of associated objects at the ending point are combined into a topological association feature. Then, together with the object length, arc length midpoint, minimum bounding rectangle, and local direction vector, they form the corresponding first geometric topological feature, second geometric topological feature, and third geometric topological feature.

[0034] In a preferred embodiment, the editing association processing based on the second standardized object, the third standardized object, the second geometric topological feature, and the third geometric topological feature to obtain the editing association relationship includes: S31. Based on the object identifier, query the object with the same identifier corresponding to the third standardized object from the second standardized object, and verify the association between the object with the same identifier and the third standardized object based on the element type, direction of travel and endpoint adjacency. Establish a direct editing association between the unique object with the same identifier that passes the association verification and the third standardized object. S32. For the third standardized object that has not established a direct editing association, based on the third geometric topology features, a candidate retrieval is performed from the second standardized object to obtain a set of editing candidate objects; based on the second and third geometric topology features, directional adjacency gating and geometric matching scoring are performed on each editing candidate object in the third standardized object and the set of editing candidate objects, and supplementary editing associations are established based on the geometric matching scoring results; S33. Determine the editing relationship between the second standardized object and the third standardized object based on the direct editing association and the supplementary editing association.

[0035] Specifically, for each third standardized object First, search for second standardized objects with the same object identifier in the second standardized object. When there exists a unique object with the same identifier, and both have the same element type ( When the travel direction and endpoint adjacency pass the verification, a direct edit association is established. When objects share the same identifier but exhibit reuse, inconsistent feature types, inconsistent one-way traffic directions, or inconsistent endpoint adjacency relationships, direct editing associations are not established directly.

[0036] For third-standardized objects without established direct edit associations, a retrieval region with a radius of 30m is constructed centered on the midpoint of its arc length. A set of editable candidate objects is obtained from the STRtree spatial index, which is pre-established or reused based on the second-standardized objects. Directional-adjacency gating and geometric constraints are applied to the candidate objects, requiring consistent feature types; for unidirectional objects, consistent direction encoding is also required, and the following conditions must be met. , ,and , , , as well as .

[0037] For candidate objects that pass the gating, geometric calculations are performed, and the expression is: ; in, , , , , and The expressions are as follows: ; ; ; ; ; And calculate: ; In the formula, Hausdorff distance is the geometric distance between two lines, which is the maximum value of the shortest distances from all points on one line to the other line. The proportion of the length of the third normalized object that falls within the 8m buffer of the second normalized object to the total length of the third normalized object. Scores obtained for geometric features For directional adjacency consistency.

[0038] In the edit-related scoring, the center point distance, direction angle, length change rate, endpoint distance, and Hausdorff distance were normalized using denominators of 30, 80, 0.50, 25, and 15, respectively. highest and Candidates are automatically linked for supplementary editing; if the difference between the highest and second-highest scores is less than 0.05, a review flag is set instead of automatically linking them for supplementary editing. review flag =1). By combining direct editing associations and supplementary editing associations, the editing association relationships between the second standardized object and the third standardized object are obtained.

[0039] In a preferred embodiment, the step of calculating the geometric deviation, geometric coverage, and attribute differences between the mutually related second and third standardized objects based on the editing association, and determining the editing protection status based on the geometric deviation, geometric coverage, and attribute differences, includes: S34. Calculate the Hausdorff distance between the second and third normalized objects that are related in the editing relationship. Use the Hausdorff distance as the geometric deviation. Determine the geometric coverage based on the proportion of the geometric length of the third normalized object within the buffer area of ​​the second normalized object to the total length of the third normalized object. S35. Compare the geometric deviation with a preset deviation threshold and the geometric coverage with a preset coverage threshold, and determine the geometric editing mark based on the comparison results; determine the attribute editing mark based on the attribute differences between the attribute fields participating in the attribute difference. S36. Based on the geometric editing mark and the attribute editing mark, determine the editing protection state as unedited state, geometric editing state, attribute editing state, or geometric and attribute combined editing state.

[0040] Specifically, for the second standardized object in the editing relationship. and the third standardized object Hausdorf distance As a geometric deviation, the geometric coverage is calculated as the proportion of the geometric length of the third normalized object falling into the buffer region of the second normalized object to the total length of the third normalized object. Its expression is: ; In the formula, For the first i The second standardized object and the first k Geometric coverage among third-normalized objects (⋅) represents the geometric length calculation. B (⋅,⋅) represents buffer operations. For the first i A second standardized object, For the first k A third standardized object.

[0041] when or At that time, geometry editing markup ;otherwise The pre-determined participating attribute difference fields are compared item by item; if any field differs, an attribute editing flag is set. ,otherwise .

[0042] according to The value determines the edit protection status ES: It is in an unedited state. In geometric editing mode, In property editing mode, This is the joint editing state for geometric attributes.

[0043] The attribute differences also record the field names where differences occurred, as well as the second and third normalized object values, for use in subsequent combination decisions and incremental update instructions.

[0044] In a preferred embodiment, S4 includes: S41. Obtain the minimum bounding rectangle of each second normalized object from the second geometric topological features, and construct a spatial index based on each minimum bounding rectangle; S42. Obtain the midpoint of the arc length of the first standardized object from the first geometric topological features, construct a search region with the midpoint of the arc length as the center and the preset search distance as the radius, and obtain the search bounding rectangle corresponding to the search region; in the spatial index, search for the second standardized object that intersects the minimum bounding rectangle and the search bounding rectangle to obtain the initial candidate object set; S43. Obtain the midpoint of the arc length of each second standardized object in the initial candidate object set from the second geometric topological features, calculate the distance between the midpoint of the arc length of the first standardized object and each second standardized object, and form a candidate object set by the second standardized objects whose midpoint distance is less than or equal to the preset retrieval distance.

[0045] Specifically, each second standardized object is obtained from the second geometric topological features. minimum bounding rectangle Based on this, spatial index records are established. IndexRecordi : ; The index record also stores the internal record identifier and the midpoint of the arc length. Local direction vectors and topological association features.

[0046] The spatial index uses a STRtree query-type R-tree: after sorting by the x-coordinate of the center of the smallest bounding rectangle, it is divided into vertical slices, and within each slice, it is sorted by the y-coordinate of the center. Each record forms a leaf node, and the process is recursively repeated until the root node is formed.

[0047] For the first standardized object The midpoint of its arc length Construct a search area with a radius of 25m centered on a point, and denote the bounding rectangle of this search area as . Starting from the root node, only visit the smallest bounding rectangle and... Intersecting child nodes yield an initial set of candidate objects; further filtering is performed based on the distance from the midpoint of the arc length to obtain the final set of candidate objects. : ; In the formula, For the set of candidate objects, For the first i The minimum bounding rectangle of the second standardized object. For the first j The bounded rectangle of the search region centered at the midpoint of the first standardized object's arc length. These are the midpoints of the arc lengths of the second and first standardized objects, respectively.

[0048] If the candidate object set is empty, the corresponding first standardized object will be marked as a new candidate and further classified in the change identification step; this step does not directly identify it as a new object.

[0049] In a preferred embodiment, S5 includes: S51. Form candidate combinations by combining the first standardized object with each candidate object in the corresponding candidate object set, anchor the identification according to the object identifier, and perform gating filtering based on the topological association features to obtain the filtered candidate combinations. S52. Based on spatial geometric features, calculate the geometric matching score of the candidate combinations and determine the comprehensive matching score by combining the topological association consistency to obtain the association matching result; S53. Initialize the first and second standardized objects corresponding to the association matching results to the state to be matched. Sort the candidate combinations according to the comprehensive matching score. Select the first standardized object and the candidate object as the candidate combinations to be matched in turn. Update the corresponding first standardized object and the candidate object to the matched state. Gather the selected candidate combinations to obtain a one-to-one matching set.

[0050] Specifically, based on the first geometric topological feature and the second geometric topological feature, the first standardized object is... The second normalized object in its candidate object set Candidate combinations are formed. If the two objects have the same identifier and there is no reuse of the same identifier, an identifier anchoring candidate is formed; the identifier anchoring candidate still needs to pass the direction consistency and endpoint adjacency consistency checks. Direction consistency is calculated for the remaining candidates. Consistency of starting point adjacency Adjacency consistency with endpoint and in accordance with Achieving topological consistency. For candidates with different identifiers, only retain... Candidate combinations with consistent element types are selected as candidate combinations for screening.

[0051] Calculate the distance between the center points of the selected candidate combinations. , direction angle Length change rate endpoint distance Hausdorf Distance and geometric coverage .

[0052] Among them, the distance from the center point The expression is: ; Direction angle The expression is: ; Rate of change of length The expression is: ; Endpoint distance The expression is: ; in: ; ; Hausdorf Distance The expression is: ; Geometric coverage The expression is: ; In the formula, i and j These represent the serial numbers of the second standardized object and the first standardized object, respectively. The average distance between paired endpoints in the same direction. The average distance between the cross endpoints. For length calculation, For buffer operations, For the first i A second standardized vector geometric object (baseline geometry). For the first j The first standardized vector geometry object (geometry to be matched).

[0053] The direction encoding of unidirectional objects should be consistent and the angle between the directions should not exceed [value missing]. Two-way objects adopt As the directional angle, the candidate combinations should also satisfy... , , , and Using the aforementioned gating formula, a geometric matching score is calculated for the candidate combinations that satisfy the above constraints. and overall matching score Their expressions are as follows: ; in, , , , , and The expressions are as follows: ; ; ; ; ; The expression for the comprehensive matching score is: ; In the formula, The score for the geometric center distance matching component is determined by the center distance of the feature plane. Obtained through normalized calculation; To match the angle of travel direction, a sub-score is calculated based on the angle between the elements. Obtained through normalized calculation; The score for matching line segment length differences is determined by the relative length difference coefficient. Obtained through normalized calculation; The average distance between endpoints is used to score the matching sub-item score, which is derived from the average distance between endpoints. Obtained through normalized calculation; The geometric contour shape matching sub-score is calculated based on the Hausdorff distance between elements. The result was obtained through normalization calculation.

[0054] Select candidate combinations by Sort from highest to lowest. When the difference in the comprehensive matching score between two candidate combinations is less than 0.01, compare the consistency of high-precision physical features, Hausdorff distance, available historical trajectory connectivity, center point distance, and object identifier lexicographical order in sequence to determine a unique ranking. Initialize each first-standardized object and second-standardized object to a matching state. Select candidate combinations with both ends in a matching state and add them to the one-to-one matching set Match, and update the corresponding ends to a matched state. Discard any other candidate combinations that conflict with matched objects, thus ensuring that each first-standardized object and each second-standardized object participates in at most one one-to-one matching relationship. ; In the formula, A set of one-to-one matches. For comprehensive matching score.

[0055] The above set is passed through The conflict resolution process was used to select the results, and they met the one-to-one matching constraint.

[0056] In a preferred embodiment, S6 includes: S61. Based on the one-to-one matching set, the first standardized object that did not participate in the one-to-one matching is identified as a new candidate object; the second standardized object that did not participate in the one-to-one matching is identified as a deletion candidate object. S62. Based on the first and second geometric topological features, perform geometric stability analysis on the first and second standardized objects that match each other in the one-to-one matching set to obtain the geometric stability results; based on the geometric stability results, perform attribute consistency analysis on the first and second standardized objects to determine attribute changes. S63. Based on the geometric stability results, determine the geometric changes between the first standardized object and the second standardized object, and determine the directional changes based on the consistency of the travel direction; S64. For the candidate objects to be deleted, determine the first standardized object group that is continuous from the newly added candidate objects, and determine the segmentation relationship based on the group coverage between the first standardized object group and the candidate objects to be deleted. S65. For newly added candidate objects, determine the second standardized object group that is consecutive from the deleted candidate objects, and determine the merging relationship based on the group coverage between the second standardized object group and the newly added candidate objects. S66. Based on geometric changes, sample the first and second standardized objects that meet the local change positioning conditions to determine the local change interval. S67. Obtain the road facility elements associated with the first and second standardized objects, determine the parent object association relationship corresponding to the road facility elements based on the one-to-one matching set, and determine the differences in road facilities based on the object identifier, element type and spatial distance; generate the original change relationship based on the addition of candidate objects, deletion of candidate objects, attribute changes, geometric changes, directional changes, segmentation relationship, merging relationship, local change interval and differences in road facilities.

[0057] Specifically, a set of associated object pairs is formed based on the one-to-one matching set, consisting of the first standardized object and the second standardized object. Geometric stability, attribute consistency, and traffic direction consistency analyses are then performed on each associated object pair. Among these, when... , and Geometric stability is determined at any time. Under the premise of geometric stability, if there are differences in the fields involved in attribute differencing, it is determined to be an attribute change. If the geometric stability condition is not met and it does not belong to a segmentation or merging relationship, it is determined to be a geometric change; if the travel directions of unidirectional objects are inconsistent, a direction change flag is added.

[0058] For unrelated second standardized objects Retrieve unassociated first-normalized objects whose center point distance is no greater than 25m; when there are 2 to 6 consecutive first-normalized objects, and the distance between adjacent endpoints is no greater than 12m, and the undirected angle between any first-normalized object and the second-normalized object is no greater than... and group coverage At that time, it was determined to be a segmentation relationship. For the unassociated first standardized object, a symmetrical method was used to retrieve the second standardized object group that was consecutive from beginning to end, and in... The relationship was determined to be merged at that time. Their group coverage rates were as follows: ; ; In the formula, The first standardized object group consists of consecutive first and last objects participating in the segmentation determination. The first and last consecutive standardized object group participating in the merger determination. The group coverage of the first and last consecutive standardized object group to the second standardized object group. The group coverage of the first standardized object to the second standardized object group, which is consecutive from beginning to end. For length calculation, This is for buffer operations.

[0059] In addition, newly added and deleted candidate objects are the first and second standardized objects that were not included in the one-to-one matching set. When generating the original change relationships, newly added and deleted candidate objects are classified according to segmentation and merging relationships: for newly added and deleted candidate objects involved in segmentation or merging relationships, the corresponding segmentation or merging relationship is recorded in the original change relationships, and they are no longer separately identified as newly added or deleted objects; newly added candidate objects not involved in segmentation or merging relationships are identified as newly added objects, and deleted candidate objects not involved in segmentation or merging relationships are identified as deleted objects. The original change relationships are generated by combining the identified newly added objects, deleted objects, attribute changes, geometric changes, directional changes, segmentation relationships, merging relationships, local change intervals, and differences in road facilities.

[0060] In a preferred embodiment, S7 includes: S71. Based on the one-to-one matching set and the edit association relationship, determine the first and third standardized objects that are related to each other; S72. Calculate the Hausdorff distance and geometric coverage between the first standardized object and the third standardized object, compare the Hausdorff distance with a preset deviation threshold, and compare the geometric coverage with a preset coverage threshold. Determine the compatibility result between the first standardized object and the third standardized object based on the comparison results. S73. Based on the original change relationship, editing protection status and compatibility results, make a combined decision to determine the target geometric source, maintenance opcode and review mark.

[0061] Specifically, based on the one-to-one matching set and edit association, using the same second standardized object. To act as a mediator, identify the first standardized object that is mutually related. With the third standardized object Calculate the Hausdorff distance and geometric coverage between the interrelated first-normalized objects and third-normalized objects; where geometric coverage represents the proportion of the geometric length of the first-normalized object within the buffer region of the third-normalized object to the total length of the first-normalized object. The compatibility condition is: ; In the formula, For the first k The third standardized object and the first j Compatibility results among the first standardized objects; Let the Hausdorff distance between the two be _____. The length of the first standardized object falling into the buffer region of the third standardized object is the proportion of the total length of the first standardized object.

[0062] For newly added or deleted objects where the first and third standardized objects cannot be determined to be related, the compatibility calculation between the first and third standardized objects is not performed. Instead, a combination decision is made based on the original change relationship and the existing edit protection status.

[0063] It should be noted that, Figure 2 The geometric source in the data is used to summarize the state machine's decision on the target geometric source and the verification state. Specifically, the new version's original geometric representation uses the geometry of the first standardized object; the "edit library geometry" uses or retains the geometry of the third standardized object; and manual verification indicates that when the target geometric source that can be automatically executed cannot be directly determined based on the original change relationship, edit protection state, and compatibility results, a verification flag is set and the process is transferred to manual verification. Before manual verification is completed, the manually edited geometry in the third data is kept from being automatically overwritten; after manual verification is completed, the target geometric source and corresponding maintenance operation code are determined based on the verification results.

[0064] Therefore, manual review is used to indicate a state to be reviewed, and is not used as an independent source of entity geometry data.

[0065] As shown in Table 1, when the original change relationship is newly added, the first standardized object geometry is used and the INSERT (add new feature to the database) operation code is generated; when deleted and in an unedited state, the DELETE (delete existing old feature) operation code is generated; when the attribute changes and is in an unedited state or a geometry editing state, the UPDATE_ATTRIBUTE (update only the attribute, the geometry remains unchanged) operation code is generated, where the third standardized object geometry is retained when in a geometry editing state.

[0066] When the original change relationship is a geometric change and is in an unedited state, the first standardized object geometry is used and the REPLACE_GEOMETRY (completely replaces the original geometry of the entire feature) opcode is generated; when it is in a geometric editing state or a geometric attribute joint editing state and the first standardized object is compatible with the third standardized object, the third standardized object geometry is used and the KEEP_EDIT_GEOMETRY (retains the manually modified geometry in the editing library, does not use the original new version geometry) opcode is generated; when they are incompatible, the third standardized object geometry is retained, the KEEP_EDIT_AND_REVIEW (uses the geometry in the editing library, adds a review mark, and requires a second manual review) opcode is generated and the review mark is set.

[0067] When the original change relationship is segmentation and the object is in an unedited state, the segmentation geometry corresponding to the first standardized object is used, and the SPLIT_RELINK (feature segmentation and topological reassociation (automatically reconstructing adjacent topological relationships after segmentation)) opcode is generated. When the object is in geometric editing state or geometric attribute joint editing state, the geometry of the third standardized object is retained first, and the SPLIT_RELINK_AND_KEEP_EDIT (segmentation and topological reassociation, prioritizing the retention of manually edited geometry) opcode is generated. When the original change relationship is merging, the MERGE_RELINK (merging multiple features and reconstructing the overall topological association) opcode is generated. When there is a change in direction, the determined geometric source remains unchanged, and the UPDATE_DIRECTION (only updating the travel / driving direction of linear features, with geometric coordinates unchanged) opcode is generated.

[0068] Table 1. Operation Code Correspondence Table

[0069] In a preferred embodiment, S8 includes: S81. Generate an incremental update instruction based on the object identifiers, original change relationships, editing protection status, target geometric source, maintenance operation codes, review markers, and attribute differences of the first standardized object, the second standardized object, and the third standardized object; wherein, the incremental update instruction includes the target data layer and the evidence identifier; S82. Classify incremental update instructions according to the original change relationship, feature type, target data layer and verification mark to obtain a map difference list; S83. Based on the verification marks in the map difference list, perform incremental updates on the third data to obtain the incremental update results.

[0070] Specifically, based on the target geometric source, maintenance operation code, and verification mark, an incremental update instruction is generated for each original change relationship. As shown in Table 2, the incremental update instruction records at least the instruction identifier, the object identifiers of the first, second, and third standardized objects, the feature type, the parent association identifier, the original change relationship, the edit protection status, the target geometric source, the maintenance operation code, the verification mark, the attribute difference, the geometric matching score, the topological association consistency, the comprehensive matching score, the target data layer, and the evidence identifier. The target data layer can correspond to the result layers of addition, deletion, attribute change, geometric change, segmentation, merging, and direction change; the evidence identifier is associated with candidate search results, gating results, geometric matching scores, edit protection status, and combined decision paths.

[0071] Table 2 Incremental Update Command Field Definition Table instruction_id Incremental instruction unique identifier string / UUID Used to uniquely identify a maintenance command old_link_id Old version of the original object identifier String Pointing to the road or lane object in O new_link_id New version of the original object identifier String Pointer to the road or lane object in N edit_link_id Edit library object identifier String Pointing to the human editing object in E feature_type High-precision map feature types enumerate Road centerline, lane centerline, lane boundaries, road facilities, etc. parent_object_id Parent level road / lane markings String Maintaining the hierarchical relationship between roads, lanes, and facilities. raw_relation Original version change relationship enumerate Add, delete, attribute, geometry, local geometry, split, merge, direction change edit_state Edit protection status ES0-ES3 This indicates no editing, manual shaping, manual attribute modification, or combined editing. geometry_source Ultimate Geometry Source enumerate OLD, NEW, EDIT or REVIEW_PENDING operation_code Maintenance operation codes enumerate INSERT, DELETE, REPLACE, KEEP_EDIT, RELINK, etc. review_flag Manual review mark 0 / 1 Where 1 indicates that automatic submission is not allowed and manual confirmation is required. attribute_diff Attribute differences JSON / key-value pairs Record field names, old values, new values, and edit values. score_geo Geometric scoring 0-1 floating-point numbers <![CDATA[record s geo or s oe geo > score_topology Topology / Structure Scoring 0-1 floating-point numbers <![CDATA[record T and available q hd > score_total Final correlation score 0-1 floating-point numbers Record the total score used for conflict resolution change_extent Local variation range Start and end mileage / geometric segment Record [s_a,s_b] or corresponding subgeometry evidence_id Internal chain of evidence identification String Association candidate set, gating, scoring, state machine path source_evidence_id External evidence identification String list Associate point clouds, images, trajectories, or acquisition tasks data_version Data version information String / Timestamp Record old versions, new versions, and library versions. target_layer Target Result Layer / Table String Specify the location to write to the maintenance system When the review flag is 0, the corresponding incremental update instruction is classified as an automatically executed instruction; when the review flag is 1, the corresponding incremental update instruction is classified as an instruction pending review. The target geometry source, maintenance opcode, and target data layer are read from the automatically executed instruction: if the target geometry source is a first standardized object, the first standardized object geometry is written; if the target geometry source is a third standardized object, the third standardized object geometry is retained or written, and insertion, deletion, geometry replacement, attribute update, segmentation association reconstruction, merging association reconstruction, or direction update are performed according to the maintenance opcode.

[0072] The pending review instruction remains in the pending review state and is transferred to the review process. Manually edited results in the third-party data are not automatically overwritten until the review is completed. After automatic execution or review, the instruction execution status, executor, execution time, and result version are associated and saved with the evidence identifier, forming an incremental update result.

[0073] According to another embodiment of the present invention, a map road change detection and update system based on a three-party association protection state machine is provided, comprising: The data normalization module is used to perform projection coordinate transformation and normalization processing on the acquired first data, second data and third data, to obtain a first normalized object, a second normalized object and a third normalized object in sequence; wherein, the first data is the updated original map road data, the second data is the original map road data, and the third data is the manually edited library road data; The feature extraction module is used to perform geometric feature quantization extraction on the first standardized object, the second standardized object, and the third standardized object respectively, to obtain the first geometric topological feature, the second geometric topological feature, and the third geometric topological feature in sequence; wherein, the geometric topological feature includes spatial geometric features and topological correlation features; The edit protection status module is used to perform edit association processing based on the second standardized object, the third standardized object, the second geometric topology feature, and the third geometric topology feature to obtain the edit association relationship; based on the edit association relationship, it calculates the geometric deviation, geometric coverage, and attribute difference between the mutually related second standardized object and the third standardized object, and determines the edit protection status based on the geometric deviation, geometric coverage, and attribute difference. The candidate retrieval module is used to construct a spatial index based on the second geometric topological feature, and to perform candidate retrieval in the spatial index using the first geometric topological feature to obtain a set of candidate objects corresponding to the first standardized object. The association matching module is used to perform multi-feature association matching between the first standardized object and the candidate objects in the candidate object set based on the first geometric topological features and the second geometric topological features, and to perform conflict resolution processing on the multi-feature association matching results to obtain a one-to-one matching set between the first standardized object and the second standardized object. The change identification module is used to perform consistency analysis and change identification processing on the first standardized object and the second standardized object based on the one-to-one matching set to obtain the original change relationship; The combined decision module is used to determine the compatibility result based on the one-to-one matching set and the editing association, and to make a combined decision based on the original change relationship, the editing protection status and the compatibility result to obtain the target geometric source, maintenance opcode and review mark; The incremental update module is used to generate incremental update instructions based on the target geometry source, maintenance opcode, and verification mark, and to perform incremental updates on the third data according to the incremental update instructions to obtain the incremental update result.

[0074] To facilitate understanding of the above technical solutions of the present invention, the working principle or operation method of the present invention in actual process will be described in detail below.

[0075] In practical applications, this method reads road data from the new version of the original high-precision map, road data from the old version of the original high-precision map, and road data from the manually edited library within the same administrative region. The system follows steps S1 to unify fields, coordinate reference systems, and geometric directions; step S2 to generate spatial geometric features and topological association features; step S3 to establish the editing association relationship between the old version of the original data and the manually edited library data and determine the editing protection status; steps S4 and S5 to complete candidate retrieval, multi-feature association matching, and conflict resolution to obtain a one-to-one matching set; step S6 to perform consistency analysis and change identification to form the original change relationship; step S7 to determine the target geometric source, maintenance operation code, and verification mark; and step S8 to output the incremental update command and incremental update result.

[0076] When verifying the output, samples of newly added, deleted, attribute-changed, geometrically changed, segmented, merged, directionally changed, and stable objects can be extracted from the implementation area. These samples are then verified by operators to form reference results, which are then compared with the algorithm output. The expressions for precision, recall, and F1 score are: ; ; ; In the formula, Precision is the accuracy rate, TP is the number of correctly identified items, Recall is the recall rate, FP is the number of false positives, FN is the number of missed items, and F1 is the harmonic mean of precision and recall. The protection provided by the edit protection state machine to manually edited results can be verified by statistically analyzing the number of instructions marked as 1 for review, the number of retained edit geometry operations, and the number of retained edit geometry and review operations. Related test conclusions should be based on actual operation logs and should not be replaced by unverified simulated data.

[0077] In summary, by utilizing the above-mentioned technical solution of the present invention, the original version association channel between the new version original data and the old version original data, and the editing protection channel between the old version original data and the manually edited database data are calculated separately and then combined in the combined decision-making stage, so that the original version change and the manual editing status become independent decision-making bases. This helps to avoid directly updating based solely on geometric similarity or object identification and thus overwriting the results of manual editing.

[0078] By unifying projected coordinates, extracting geometric topological features, retrieving candidate spatial indexes, anchoring signs, topological gating, geometric scoring, and resolving deterministic conflicts, the candidate range can be narrowed down and a repeatable one-to-one matching set can be formed, reducing the risk of erroneous associations in intersections, parallel roads, and signage change scenarios.

[0079] By combining the original change relationships, edit protection status, and compatibility results, the source of the target geometry, maintenance opcodes, and review markers are determined, allowing compatible manually edited geometry to be preserved and incompatible cases to enter the review process.

[0080] By recording the target data layer and evidence identifier in the incremental update instruction, candidate retrieval, correlation matching, editing protection status and combined decision path can be associated and traced, and it is convenient for the maintenance system to distinguish between automatically executed instructions and instructions pending review.

[0081] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A map road change detection and update method based on a three-party association protection state machine, characterized in that, include: S1. Perform projection coordinate transformation and normalization on the acquired first data, second data and third data to obtain a first normalized object, a second normalized object and a third normalized object in sequence; wherein, the first data is the updated original map road data, the second data is the original map road data, and the third data is the manually edited library road data; S2. Geometric features are extracted from the first standardized object, the second standardized object, and the third standardized object respectively to obtain the first geometric topological feature, the second geometric topological feature, and the third geometric topological feature in sequence; wherein, the geometric topological feature includes spatial geometric features and topological correlation features; S3. Based on the second standardized object, the third standardized object, the second geometric topology feature, and the third geometric topology feature, perform editing association processing to obtain the editing association relationship; according to the editing association relationship, calculate the geometric deviation, geometric coverage, and attribute difference between the mutually related second standardized object and the third standardized object, and determine the editing protection status based on the geometric deviation, geometric coverage, and attribute difference; S4. Construct a spatial index based on the second geometric topological feature, and use the first geometric topological feature to perform candidate retrieval in the spatial index to obtain a set of candidate objects corresponding to the first standardized object. S5. Based on the first geometric topological feature and the second geometric topological feature, perform multi-feature association matching on the first standardized object and the candidate objects in the candidate object set, and perform conflict resolution processing on the multi-feature association matching results to obtain a one-to-one matching set between the first standardized object and the second standardized object. S6. Based on the one-to-one matching set, perform consistency analysis and change identification processing on the first standardized object and the second standardized object to obtain the original change relationship; S7. Determine the compatibility results based on the one-to-one matching set and editing association, and make a combined decision based on the original change relationship, editing protection status and compatibility results to obtain the target geometry source, maintenance opcode and review mark; S8. Generate incremental update instructions based on the target geometry source, maintenance opcode, and verification mark, and perform incremental update on the third data according to the incremental update instructions to obtain the incremental update result.

2. The map road change detection and update method based on a three-party association protection state machine according to claim 1, characterized in that, S1 includes: S11. Read the road records in the first data, second data and third data respectively, obtain the geometric information of each road record, and keep the object identifier, feature type, traffic direction, topological association information, parent association identifier, version field and attribute field participating in attribute difference of each road record as reserved fields to obtain the road record set. S12. Read the source coordinate reference system corresponding to each road record in the road record set and verify it. Based on the verified source coordinate reference system and the target projection coordinate reference system, establish a coordinate transformation relationship. Transform the horizontal and vertical coordinates according to the original arrangement order of the geometric vertices, retain the original geometry and generate the projection working geometry. S13. Perform vertex deduplication on the projection working geometry, and verify the coordinate values, number of vertices and projection length. Road records that pass the verification are considered valid road records. S14. Based on the topological association information of the valid road records, determine the starting point connection direction and the ending point connection direction; for road records with unidirectional traffic and projection working geometry opposite to the connection direction, reverse the coordinate sequence to obtain the normalized projection working geometry; S15. Generate internal record identifiers for the valid road records corresponding to the first data, the second data, and the third data respectively, and generate standardized records based on the internal record identifiers, reserved fields, original geometry, and normalized projection working geometry to obtain the first standardized object, the second standardized object, and the third standardized object.

3. The map road change detection and update method based on a three-party association protection state machine according to claim 1, characterized in that, S2 includes: S21. Calculate the midpoint of the arc length, the object length, and the minimum bounding rectangle of the first, second, and third standardized objects respectively, and generate local direction vectors based on the object starting point and local sampling points determined along the object geometry according to the preset arc length to obtain spatial geometric features; S22. Obtain the travel direction, the set of associated objects at the starting point, and the set of associated objects at the ending point of the first, second, and third standardized objects. Determine the starting point connectivity and the ending point connectivity based on the set of associated objects at the starting point and the set of associated objects at the ending point to obtain the topological association features. S23. Combine the spatial geometric features and topological association features of the first standardized object, the second standardized object, and the third standardized object respectively to obtain the first geometric topological feature, the second geometric topological feature, and the third geometric topological feature.

4. The map road change detection and update method based on a three-party association protection state machine according to claim 1, characterized in that, The editing association processing based on the second standardized object, the third standardized object, the second geometric topological feature, and the third geometric topological feature yields the following editing association relationships: S31. Based on the object identifier, query the object with the same identifier corresponding to the third standardized object from the second standardized object, and verify the association between the object with the same identifier and the third standardized object based on the element type, direction of travel and endpoint adjacency. Establish a direct editing association between the unique object with the same identifier that passes the association verification and the third standardized object. S32. For the third standardized object that has not established a direct editing association, based on the third geometric topology features, a candidate retrieval is performed from the second standardized object to obtain a set of editing candidate objects; based on the second and third geometric topology features, directional adjacency gating and geometric matching scoring are performed on each editing candidate object in the third standardized object and the set of editing candidate objects, and supplementary editing associations are established based on the geometric matching scoring results; S33. Determine the editing relationship between the second standardized object and the third standardized object based on the direct editing association and the supplementary editing association.

5. The map road change detection and update method based on a three-party association protection state machine according to claim 4, characterized in that, The step of calculating the geometric deviation, geometric coverage, and attribute differences between the interrelated second and third standardized objects based on the editing association, and determining the editing protection status based on the geometric deviation, geometric coverage, and attribute differences, includes: S34. Calculate the Hausdorff distance between the second and third normalized objects that are related in the editing relationship. Use the Hausdorff distance as the geometric deviation. Determine the geometric coverage based on the proportion of the geometric length of the third normalized object within the buffer area of ​​the second normalized object to the total length of the third normalized object. S35. Compare the geometric deviation with a preset deviation threshold and the geometric coverage with a preset coverage threshold, and determine the geometric editing mark based on the comparison results; determine the attribute editing mark based on the attribute differences between the attribute fields participating in the attribute difference. S36. Based on the geometric editing mark and the attribute editing mark, determine the editing protection state as unedited state, geometric editing state, attribute editing state, or geometric and attribute combined editing state.

6. The map road change detection and update method based on a three-party association protection state machine according to claim 1, characterized in that, S4 includes: S41. Obtain the minimum bounding rectangle of each second normalized object from the second geometric topological features, and construct a spatial index based on each minimum bounding rectangle; S42. Obtain the midpoint of the arc length of the first standardized object from the first geometric topological features, construct a search region with the midpoint of the arc length as the center and the preset search distance as the radius, and obtain the search bounding rectangle corresponding to the search region; in the spatial index, search for the second standardized object that intersects the minimum bounding rectangle and the search bounding rectangle to obtain the initial candidate object set; S43. Obtain the midpoint of the arc length of each second standardized object in the initial candidate object set from the second geometric topological features, calculate the distance between the midpoint of the arc length of the first standardized object and each second standardized object, and form a candidate object set by the second standardized objects whose midpoint distance is less than or equal to the preset retrieval distance.

7. The map road change detection and update method based on a three-party association protection state machine according to claim 1, characterized in that, S5 includes: S51. Form candidate combinations by combining the first standardized object with each candidate object in the corresponding candidate object set, anchor the identification according to the object identifier, and perform gating filtering based on the topological association features to obtain the filtered candidate combinations. S52. Based on spatial geometric features, calculate the geometric matching score of the candidate combinations and determine the comprehensive matching score by combining the topological association consistency to obtain the association matching result; S53. Initialize the first and second standardized objects corresponding to the association matching results to the state to be matched. Sort the candidate combinations according to the comprehensive matching score. Select the first standardized object and the candidate object as the candidate combinations to be matched in turn. Update the corresponding first standardized object and the candidate object to the matched state. Gather the selected candidate combinations to obtain a one-to-one matching set.

8. The map road change detection and update method based on a three-party association protection state machine according to claim 1, characterized in that, S6 includes: S61. Based on the one-to-one matching set, the first standardized object that did not participate in the one-to-one matching is identified as a new candidate object; the second standardized object that did not participate in the one-to-one matching is identified as a deletion candidate object. S62. Based on the first and second geometric topological features, perform geometric stability analysis on the first and second standardized objects that match each other in the one-to-one matching set to obtain the geometric stability results; based on the geometric stability results, perform attribute consistency analysis on the first and second standardized objects to determine attribute changes. S63. Based on the geometric stability results, determine the geometric changes between the first standardized object and the second standardized object, and determine the directional changes based on the consistency of the travel direction; S64. For the candidate objects to be deleted, determine the first standardized object group that is continuous from the newly added candidate objects, and determine the segmentation relationship based on the group coverage between the first standardized object group and the candidate objects to be deleted. S65. For newly added candidate objects, determine the second standardized object group that is consecutive from the deleted candidate objects, and determine the merging relationship based on the group coverage between the second standardized object group and the newly added candidate objects. S66. Based on geometric changes, sample the first and second standardized objects that meet the local change positioning conditions to determine the local change interval. S67. Obtain the road facility elements associated with the first and second standardized objects, determine the parent object association relationship corresponding to the road facility elements based on the one-to-one matching set, and determine the differences in road facilities based on the object identifier, element type and spatial distance; generate the original change relationship based on the addition of candidate objects, deletion of candidate objects, attribute changes, geometric changes, directional changes, segmentation relationship, merging relationship, local change interval and differences in road facilities.

9. The map road change detection and update method based on a three-party association protection state machine according to claim 1, characterized in that, S7 includes: S71. Based on the one-to-one matching set and the edit association relationship, determine the first and third standardized objects that are related to each other; S72. Calculate the Hausdorff distance and geometric coverage between the first standardized object and the third standardized object, compare the Hausdorff distance with a preset deviation threshold, and compare the geometric coverage with a preset coverage threshold. Determine the compatibility result between the first standardized object and the third standardized object based on the comparison results. S73. Based on the original change relationship, editing protection status and compatibility results, make a combined decision to determine the target geometric source, maintenance opcode and review mark.

10. The map road change detection and update method based on a three-party association protection state machine according to claim 9, characterized in that, S8 includes: S81. Generate an incremental update instruction based on the object identifiers, original change relationships, editing protection status, target geometric source, maintenance operation codes, review markers, and attribute differences of the first standardized object, the second standardized object, and the third standardized object; wherein, the incremental update instruction includes the target data layer and the evidence identifier; S82. Classify incremental update instructions according to the original change relationship, feature type, target data layer and verification mark to obtain a map difference list; S83. Based on the verification marks in the map difference list, perform incremental updates on the third data to obtain the incremental update results.

Images