Block protection-oriented surveying and mapping data intelligent classification system

Abstract

The invention relates to the technical field of geographic space information retrieval, in particular to a block protection-oriented surveying and mapping data intelligent classification system, which comprises a ground feature structure analysis module, a semantic association modeling module, a path overlapping analysis module, a category attribution judgment module and a classification result integration module. According to the method, the attribute sequence is constructed through the surface feature elements, sequence analysis and semantic reduction are carried out, the expression hierarchy and the structural logic of the attribute tag are improved, the semantic context of the surface features is reduced through the directed connection relationship of the attribute paths, the spatial semantic association between the surface features is reflected, and the precise sorting of the attribute information and the analysis of the path connection relationship are realized; a path overlapping node end point frequency statistical mode is introduced, the stability and accuracy of ground feature category judgment are enhanced, the distribution and attribution precision of classification labels is optimized, classification results are integrated in combination with the mapping relation of path labels and classification nodes, and the automatic classification capacity of block protection area surveying and mapping data and the definition of category division are improved.

Description

Technical Field

[0001] This invention relates to the field of geospatial information retrieval technology, and in particular to an intelligent classification system for surveying and mapping data for street preservation. Background Technology

[0002] The field of geospatial information retrieval technology involves the acquisition, organization, management, and querying of geospatial data, including core aspects such as structured storage of spatial data, spatial relationship-based retrieval mechanisms, spatial semantic understanding, and user intent modeling. It aims to achieve accurate identification and efficient retrieval of geographic entities and is widely applied in urban planning, resource monitoring, public management, and emergency response. Especially when combined with data sources such as remote sensing imagery, geographic annotation, and 3D modeling, it enhances the automation and intelligence of spatial information processing. Among these, traditional intelligent classification systems for surveying and mapping data aimed at the protection of urban blocks refer to the extraction and classification of surveying and mapping data for areas with historical and cultural value or planning and control significance to assist in protection management. These systems primarily address the identification and classification of different land cover categories within protected urban blocks. Traditional methods rely on manual interpretation and rule-based classification methods based on map patch attributes. These methods depend on sub-categories of information such as historical imagery, cadastral data, and field survey records, combined with attributes such as map patch area, shape factors, and color characteristics for manual classification or semi-automatic assisted judgment.

[0003] Existing technologies that rely on manual interpretation and classification based on map feature attributes have significant limitations in complex terrain environments. Attributes such as map feature area and shape factor exhibit high similarity and ambiguous boundaries in actual surveying data, making it difficult to accurately distinguish between land feature categories. Furthermore, the formulation of manual rules is limited by the scope of experiential knowledge, which can easily lead to omissions and misjudgments when processing large-scale data. In scenarios involving a mixture of historical imagery and cadastral information, the inconsistent presentation of attribute information makes it difficult for classification standards to fully adapt to the diversity of land features. Field survey records rely on subjective judgment, resulting in insufficient data consistency, which in turn affects the uniformity and practicality of the final classification results. This is especially true in protected areas, where it is difficult to achieve high-precision identification and classification of land feature types with micro-differences. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing an intelligent classification system for surveying and mapping data for street protection.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent classification system for surveying and mapping data for street block preservation, comprising:

[0006] The feature structure analysis module acquires feature elements from the surveying and mapping data, extracts feature attribute tags, determines the first occurrence location and distribution frequency of key attributes, parses and restores the arrangement order of feature attributes, and constructs a feature semantic unit sequence template.

[0007] The semantic association modeling module sorts the attribute labels based on the semantic unit sequence template of the ground features, establishes a directed path of attributes from the top layer to the bottom layer according to the sorting result, collects the connection node numbers and adjacent node relationships of the edges in the path, and constructs the semantic association path structure.

[0008] The path overlap analysis module extracts the sequence of ground feature path nodes based on the semantic association path structure, compares the intersection nodes and counts the frequency of the endpoint node, filters the overlapping paths, and obtains a set of ground feature overlapping paths.

[0009] The category attribution determination module collects the category labels of the endpoint nodes in the set of overlapping ground features, sorts them by frequency of occurrence, and matches the path endpoint labels to obtain a group of ground feature category attribution labels.

[0010] As a further aspect of the present invention, the semantic unit sequence template for ground features includes an attribute label arrangement structure, a set of key attribute fragments, and an attribute frequency weight model; the semantic association path structure includes an attribute hierarchy mapping relationship, a directed path node chain, and a set of node connection relationships; the set of overlapping ground features paths includes a set of path intersection nodes, a frequency distribution of endpoint nodes, and a result of overlapping node filtering; and the ground feature category attribution label group includes a category label frequency ranking, a path endpoint category label mapping, and a result of attribution category label matching.

[0011] As a further aspect of the present invention, the ground feature structure analysis module includes:

[0012] The attribute extraction submodule acquires the land features in the surveying and mapping data, performs attribute-level annotation operations on them, records the index positions of key attributes, and classifies the land features according to attributes by comparing the relationship between the first occurrence position of the key attributes and the total number of attributes, thus obtaining the key attribute index distribution results.

[0013] The original sequence fragment construction submodule extracts attribute fragments from the original text of the land cover based on the key attribute index distribution results, extracts them based on the index range of the attribute in the text, constructs a set of attribute fragments, and reorganizes the fragments in combination with the land cover information to obtain a set of key attribute original sequence fragments.

[0014] The semantic template generation submodule rearranges the set of attribute fragments in the original order according to the set of key attribute fragments, splices multiple key attributes of the same land feature into a sequence, and integrates the semantic units of each land feature to obtain a land feature semantic unit sequence template.

[0015] As a further aspect of the present invention, the semantic association modeling module includes:

[0016] The hierarchical sorting submodule, based on the semantic unit sequence template of the land feature, and combined with the hierarchical labels of the land feature attribute nodes, compares and sorts the attribute nodes according to their hierarchical label priority values, and rearranges the land feature attributes in order from top to bottom to obtain the attribute sorting index value.

[0017] The path construction submodule sorts the index value of the attribute, obtains the set of adjacent nodes in the attribute rearrangement sequence, numbers and records the connection direction of each pair of adjacent nodes, and combines the node connection relationship in the path to integrate the structural edge information and generate attribute directed path graph data.

[0018] The node structure extraction submodule collects the node numbers and adjacent node pairs in the connecting edges based on the attribute-directed path graph data, constructs a node mapping table based on the adjacent structural relationship, stores the upstream and downstream relationship types and connection directions between each attribute, and obtains the semantic association path structure.

[0019] As a further aspect of the present invention, the path overlap analysis module includes:

[0020] Based on the semantic association path structure, the path extraction submodule collects the node sequence in any two feature paths, extracts the node number information under each path in turn, establishes a feature node mapping set, marks the feature identifier to which the path belongs and the path length parameter, and obtains the feature path node number value.

[0021] The intersection comparison submodule performs an intersection operation on the node number sequences of two feature paths based on the node number values ​​of the feature path, extracts the endpoint node number in the intersection, counts the number of times each node appears in the differentiated path, compares it with the path overlap benchmark value, filters out path pairs that meet the conditions, and establishes a set of path intersection numbers that meet the conditions.

[0022] The path filtering submodule, based on the set of path intersection numbers that meet the conditions, queries the original feature path identifiers according to the path combinations corresponding to the numbers, integrates the feature identifiers and path overlap node information, and generates a set of feature overlap paths.

[0023] As a further aspect of the present invention, the category attribution determination module includes:

[0024] The category label collection submodule collects the category label set to which each end node belongs based on the end node in the set of overlapping ground features paths, performs index mapping between ground feature paths and their end node labels, and generates path end point category label groups.

[0025] The tag frequency statistics submodule performs a frequency statistics operation on the category tags based on the path endpoint category tag group, records the number of times each category tag appears in the feature path set, and sorts them according to the number of occurrences to obtain a sorted category tag sequence.

[0026] The category determination submodule performs a matching judgment on the tag set corresponding to the endpoint node in the feature path according to the sorted category tag sequence, selects the tag item with the first position in the sorted sequence as the path's category, integrates the feature path's category tags, and obtains the feature category category tag group.

[0027] As a further aspect of the present invention, the system also includes a classification result integration module:

[0028] The classification result integration module assigns the label group to the land feature category, counts the classification nodes to which the labels belong, divides the land feature paths under the corresponding nodes, analyzes the classification correspondence between nodes and land feature paths, and generates a street block protection surveying data classification table.

[0029] The street block protection surveying and mapping data classification table includes classification node identifiers, feature path grouping results, and classification attribution relationship mapping results.

[0030] As a further aspect of the present invention, the classification result integration module includes:

[0031] The node extraction submodule collects the classification nodes corresponding to each tag according to the tag group to which the land feature category belongs, records the path numbers associated with the classification nodes and the number of corresponding land feature path sets, determines the matching index between the category tag and the classification node, and obtains the tag-assigned node number value.

[0032] The path classification submodule, based on the node number value of the tag, divides the corresponding feature paths into each category node according to the category tag, establishes a two-way correspondence structure between feature path numbers and node numbers, extracts the path number list to which each node belongs, and obtains the number of node paths.

[0033] The structure generation submodule integrates the classification nodes and their subordinate feature path numbers through structural mapping based on the number of node paths, outputs the classification node index, corresponding category label and total number of paths, determines the attribution of nodes and feature paths, and generates a street block protection surveying and mapping data classification table.

[0034] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0035] In this invention, after constructing attribute sequences from land feature elements in surveying and mapping data, sequential parsing and semantic restoration are performed. This enhances the expression level and structural logic of land feature attribute labels. By restoring the inherent semantic context of land feature attributes through the directed connection relationship of attribute paths, the spatial semantic association between land features is effectively reflected, thereby achieving accurate sorting of attribute information and in-depth analysis of path connectivity. When identifying land feature categories, the introduction of the endpoint frequency statistics method of overlapping path nodes helps to improve the stability and accuracy of land feature category determination. During the classification process, land feature category label groups are constructed based on the frequency of endpoint node labels, optimizing the overall distribution rationality and classification accuracy of classification labels. By integrating the classification table results through the mapping relationship between path labels and classification nodes, the automated classification capability and structural clarity of category division of surveying and mapping data of protected areas are effectively improved. Attached Figure Description

[0036] Figure 1 This is a system flowchart of the present invention;

[0037] Figure 2 This is a flowchart of the ground feature structure analysis module in this invention;

[0038] Figure 3 This is a flowchart of the semantic association modeling module in this invention;

[0039] Figure 4 This is a flowchart of the path overlap analysis module in this invention;

[0040] Figure 5 This is a flowchart of the category attribution determination module in this invention;

[0041] Figure 6 This is a flowchart of the classification result integration module in this invention. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0043] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0044] Please see Figure 1 A smart classification system for surveying and mapping data for neighborhood preservation includes:

[0045] The feature structure analysis module acquires feature elements from the surveying and mapping data, extracts feature attribute tags, determines the first occurrence location and distribution frequency of key attributes, parses and restores the arrangement order of feature attributes, and constructs a feature semantic unit sequence template.

[0046] The semantic association modeling module is based on the semantic unit sequence template of ground features. It sorts the attribute labels, establishes a directed path of attributes from the top layer to the bottom layer according to the sorting results, collects the connection node numbers and adjacent node relationships of the edges in the path, and constructs the semantic association path structure.

[0047] The path overlap analysis module extracts the sequence of ground feature path nodes based on the semantic association path structure, compares the intersection nodes and counts the frequency of the endpoint node, filters the overlapping paths, and obtains the set of ground feature overlapping paths.

[0048] The category attribution determination module collects the category labels of the endpoint nodes in the set of overlapping ground features paths, sorts them by frequency of occurrence, and matches the path endpoint labels to obtain the ground feature category attribution label group.

[0049] The classification result integration module assigns labels to land feature categories, counts the classification nodes to which the labels belong, divides land feature paths into corresponding nodes, analyzes the classification correspondence between nodes and land feature paths, and generates a classification table of street protection surveying data.

[0050] The semantic unit sequence template for ground features includes an attribute label arrangement structure, a set of key attribute fragments, and an attribute frequency weight model. The semantic association path structure includes attribute hierarchical mapping relationships, directed path node chains, and a set of node connection relationships. The set of overlapping ground features paths includes a set of path intersection nodes, a frequency distribution of endpoint nodes, and the results of overlapping node filtering. The label group for ground feature categories includes a ranking of category label frequencies, a mapping of path endpoint category labels, and a matching result of category labels. The classification table for street block protection surveying and mapping data includes classification node identifiers, ground feature path grouping results, and classification relationship mapping results.

[0051] Please see Figure 2 The feature structure analysis module includes:

[0052] The attribute extraction submodule acquires the land features in the surveying and mapping data, performs attribute-level annotation operations on them, records the index positions of key attributes, and classifies the land features according to attributes by comparing the relationship between the first occurrence position of the key attributes and the total number of attributes, thus obtaining the key attribute index distribution results.

[0053] The process involves acquiring geographic features from surveying data and performing attribute-level annotation on them, recording the index positions of key attributes. For example, in a surveying dataset about a "historical district central square," the identified geographic features include "central fountain," "square paving," and "surrounding green belt." Attribute-level annotation is performed manually or semi-automatically on the original text description of each geographic feature. This operation specifically identifies and records the start and end character indices of each key attribute (e.g., "fountain height," "paving material," "tree type") in the original text, forming a precise index range. Taking the "Central Fountain" as an example, its text description is "5 meters high, a classical fountain built in 1930, in good condition." "5 meters high" is labeled as the attribute "height," with an index range of [2, 6]. "1930" is labeled as the attribute "year of construction," with an index range of [10, 14]. "Classical" is labeled as the attribute "style," with an index range of [17, 20]. "Good condition" is labeled as the attribute "maintenance," with an index range of [27, 29]. This precise character index record forms the key attribute index location column for this feature. The table calculates a classification ratio by comparing the first occurrence position of the key attributes of each feature with the total number of key attributes contained in that feature. Taking the "Central Fountain" as an example, it has a total of 4 key attributes (height, construction year, style, and maintenance status). The first occurrence position is the starting index 2 of the attribute "height". Therefore, the classification ratio of this feature is calculated as 2 / 4 = 0.5. This ratio is compared with three preset threshold intervals for judgment. The preset intervals are: when the ratio is less than 0.2, it is judged as a "core feature type feature"; when the ratio is between 0.2 (inclusive) and 0.6 (inclusive), it is judged as... Features are classified as "auxiliary descriptive features" when the ratio is greater than 0.6, and as "supplementary information features" when the ratio is greater than 0.6. The interval values ​​(0.2 and 0.6) were determined after repeated cross-validation and expert evaluation of a large amount of urban historical block mapping data, ensuring an accurate reflection of the distribution characteristics of feature attributes. In this example, the classification ratio of "central fountain" is 0.5, which falls within the interval of 0.2 to 0.6, so it is classified as "auxiliary descriptive feature". This process is repeated for all feature elements, and finally the key attribute index distribution results containing the attribute index of each feature element and its corresponding classification identifier are obtained.

[0054] The original sequence fragment construction submodule extracts attribute fragments from the original text of the land cover based on the key attribute index distribution results, extracts them based on the index range of the attribute in the text, constructs a set of attribute fragments, and reassembles the fragments in combination with the land cover information to obtain a set of key attribute original sequence fragments.

[0055] Based on the key attribute index distribution results, attribute fragments are extracted from the original text of the land feature. For example, based on the key attribute index distribution results of "Central Fountain" obtained in the previous step, the original text of "Central Fountain" "5 meters high, classical fountain built in 1930, well maintained" is extracted based on the index interval of the attribute in the text. The "5 meters high" corresponding to the index interval [2, 6] is extracted as attribute fragment A, the "1930" corresponding to the index interval [10, 14] is extracted as attribute fragment B, the "classical" corresponding to the index interval [17, 20] is extracted as attribute fragment C, and the "well maintained" corresponding to the index interval [27, 29] is extracted as attribute fragment D. The extracted attribute fragments are... The segments together construct an initial set of attribute fragments. Combining the information of the land features, the fragments are recombined. The specific recombining process takes into account the type of land feature (e.g., "fountain"). Attribute fragments that are directly related, such as "5 meters high" and "built in 1930", are logically merged. For example, "5 meters high" and "built in 1930" are recombined into "5 meters high, built in 1930". This ensures that the recombined fragments can more completely express a certain aspect of the land feature's characteristics, while maintaining the local order of the original text and avoiding semantic breaks. By performing this kind of extraction, truncation, and recombining of attribute fragments for each land feature, a complete and ordered set of key attribute original sequence fragments containing each land feature is finally obtained.

[0056] The semantic template generation submodule rearranges the set of attribute fragments in the original order according to the set of key attribute fragments, splices multiple key attributes of the same land feature into a sequence, and integrates the semantic units of each land feature to obtain the land feature semantic unit sequence template.

[0057] Following the previous step's acquisition of the original set of key attribute fragments for the "Central Fountain," which includes fragments such as "5 meters high, built in 1930," "Classical style," and "well-maintained," the fragments are first logically rearranged according to their order of appearance in the original text. This ensures the arrangement of the fragments aligns with the original expression logic of the feature description. For example, if the original description is "5 meters high, a classical-style fountain built in 1930, well-maintained," the rearranged order will still be "5 meters high, built in 1930" first, followed by "Classical style," and finally "well-maintained." Multiple key attributes of the same feature are then sequentially pieced together, connecting the rearranged attribute fragments according to their logical relationships. For instance, "5 meters high, built in 1930" and "Classical style" are linked together. The sequence of descriptions for the "Central Fountain" is formed by piecing together the phrases "5 meters high, built in 1930, classical style" and then integrating "well-maintained" to form "5 meters high, built in 1930, classical style, well-maintained." This represents the complete descriptive sequence of the "Central Fountain." Subsequently, the semantic units of each feature are integrated. Specifically, the pieced feature description sequence is parsed into higher-level semantic units. For example, "5 meters high, built in 1930, classical style, well-maintained" is further abstracted into the semantic unit {Type: Fountain, Geometric Attribute: Height, Time Attribute: Year Built, Descriptive Attribute: Style, Status Attribute: Maintenance}. This semantic unit sequence template preserves the logical structure and semantic relationships of the feature attributes. By performing this rearrangement, piecing together, and semantic integration operation on all feature elements, the feature semantic unit sequence template is obtained.

[0058] Please see Figure 3 The semantic association modeling module includes:

[0059] The hierarchical sorting submodule, based on the semantic unit sequence template of land cover, combines the hierarchical labels of land cover attribute nodes to compare and sort the attribute nodes according to their hierarchical label priority values. It then rearranges the land cover attributes from top to bottom using the following formula:

[0060] ;

[0061] Get the attribute sorting index value;

[0062] in, Represents the attribute sorting index value. Representing the The weight of each land feature attribute, Representing the Hierarchical label values ​​of each land feature attribute, Representative and the The priority parameter of the attribute node associated with each attribute. The total number of attributes representing land features. It is the base of the natural logarithm;

[0063] The semantic unit sequence template for "Historical Building A" generated in the previous step contains attribute nodes such as "Building Name," "Construction Date," "Protection Level," and "Architectural Style." Each attribute node has a predefined hierarchical label; for example, "Building Name" is level 1, "Protection Level" is level 1, "Construction Date" is level 2, and "Architectural Style" is level 3. The value of the hierarchical label reflects the importance and level of abstraction of the attribute in the street block protection mapping data. Level 1 represents core identification information, level 2 represents basic descriptive information, and level 3 represents auxiliary descriptive information. Attribute nodes are compared and sorted according to their hierarchical label priority values. These priority values ​​are determined through the experience of historical street block experts and data importance assessment. The weight representing the k-th feature attribute indicates the importance of that attribute in the overall ranking. Its value ranges from 0 to 1 and is set according to the priority rules of the historical district protection plan. For example, the core identifier attribute weight is set to 0.9, the basic description attribute weight is set to 0.7, and the auxiliary description attribute weight is set to 0.5. This represents the hierarchical label value of the k-th land feature attribute, reflecting the hierarchical classification of the attribute. Its value range is a positive integer, such as 1, 2, 3. This represents the priority parameter of the attribute node associated with the k-th attribute, reflecting the relative importance of that attribute in the current land cover. Its value is a positive real number; for example, more important attributes have a higher priority. The value, n represents the total number of land feature attributes, e is the base of the natural logarithm, and the summation symbol in the formula. This represents the summation of the weighted hierarchical values ​​of all n land feature attributes, using a multiplication operation. Calculate the weighted hierarchical contribution of each attribute, in the form of an activation function. Then the attribute priority parameter Mapping to a range of 0 to 1, thus incorporating it non-linearly into the ranking value calculation, the advantage of this activation function design is that it makes the attribute priority parameter... The impact on the final sorted values ​​is no longer linear, especially in When the value is high, its influence on the result tends to saturate, effectively avoiding the excessive dominance of a few extremely high-priority attributes in the overall ranking result. This makes the ranking result more balanced and consistent with the actual semantic importance. The location of the land feature attributes is rearranged from top to bottom to ensure that important attributes are presented first. For example, for the four attributes of "Historical Building A", its , , The value settings are shown in Table 1;

[0064] Table 1: Land Feature Attribute Parameter Table

[0065] Attribute Name Weight Hierarchical label values Priority parameter Building Name 0.9 1 2.5 Construction period 0.7 2 1.8 Protection level 0.9 1 2.8 Architectural style 0.5 3 1.2

[0066] As shown in Table 1, It was set according to the importance rules of the historic district preservation plan. It is a predefined hierarchy of attributes. It is determined based on the completeness of the description of the attribute in a specific land feature and expert scoring, and the parameters are substituted into the formula for calculation;

[0067] For "building name" ( ): ;

[0068] Regarding the "construction date" ( ): ;

[0069] For "protection level" ( ): ;

[0070] Regarding "architectural style" ): ;

[0071] Total attribute sorting index value For each attribute The sum:

[0072] ;

[0073] The result This represents the comprehensive hierarchical ranking index value of all key attributes of "Historic Building A". A higher value indicates a higher comprehensive hierarchical priority for that feature attribute. This value is used to guide the subsequent rearrangement of feature attributes, calculated based on each attribute. Values, sorting attributes in descending order, such as "construction year" ( "Architectural style" "Protection Level" "Building Name" Therefore, the rearranged attribute order is: construction year, architectural style, protection level, building name, and finally the attribute sorting index value.

[0074] The path construction submodule sorts the index values ​​according to the attributes, obtains the set of adjacent nodes in the attribute rearrangement sequence, numbers and records the connection direction of each pair of adjacent nodes, combines the node connection relationship in the path, integrates the structural edge information, and generates attribute directed path graph data.

[0075] Based on the attribute sorting index value, for example, according to the attribute rearrangement sequence (construction year, architectural style, protection level, building name) of "Historical Building A" obtained in the previous step, obtain the set of adjacent nodes in the attribute rearrangement sequence. Specifically, identify consecutive attribute pairs in the sequence as adjacent node pairs. For example, "construction year" and "architectural style" form an adjacent node pair, "architectural style" and "protection level" form an adjacent node pair, and "protection level" and "building name" form an adjacent node pair. Number each pair of adjacent nodes and record their connection direction. For example, number "construction year" as N1, "architectural style" as N2, "protection level" as N3, and "building name" as N4. For N4, the connection direction from N1 to N2 is "sequential association", the connection direction from N2 to N3 is "derived association", and the connection direction from N3 to N4 is "refined association". The connection direction predefines the semantic relationship type between different attributes, ensuring the logical correctness of the path. Combining the node connection relationship in the path, the structural edge information is integrated, and the number, connection direction and semantic relationship type are encoded as structural edge information, for example, represented as {N1:N2, type: sequential association}, {N2:N3, type: derived association}, {N3:N4, type: refined association}. By performing this operation on all adjacent node pairs and integrating all structural edge information, the attribute directed path graph data is finally generated.

[0076] The node structure extraction submodule collects the node numbers and adjacent node pairs in the connecting edges based on the attribute-directed path graph data, constructs a node mapping table based on the adjacent structure relationship, stores the upstream and downstream relationship types and connection directions between each attribute, and obtains the semantic association path structure.

[0077] Following the previously generated directed path graph data for "Historical Building A", this step collects the node numbers and adjacent node pairs in the connecting edges. Specifically, it traverses each edge in the graph, extracting the starting and ending node numbers. For example, from {N1:N2, type: sequential association}, it extracts node numbers N1 and N2 and records their adjacent relationship N1→N2. Based on the adjacent structure relationship, it constructs a node mapping table, storing the collected node numbers and adjacent relationships in a node mapping table. For example, it constructs a dictionary or hash table where the key is the starting node number and the value is a list containing all nodes connected to that starting node. The table stores the adjacent nodes of the relationship and their connection types, as well as the upstream and downstream relationship types and connection directions between each attribute. For example, the node mapping table stores N1: {N2: sequential association}, indicating that N1 is the "upstream" node of N2 and the relationship type is "sequential association". N2: {N1: reverse sequential association, N3: derived association}, indicating that N2 has an "upstream" relationship from N1 with "reverse sequential association" and a "downstream" relationship pointing to N3 with "derived association". This table clearly defines the hierarchical and flow relationships between geographic attribute nodes. By performing this collection, construction and storage operation on all connection edges, the semantic association path structure is finally obtained.

[0078] Please see Figure 4 The path overlap analysis module includes:

[0079] The path extraction submodule, based on a semantically related path structure, collects node sequences from any two feature paths, sequentially extracts node number information for each path, establishes a feature node mapping set, and marks the feature identifier and path length parameter of the path, using the following formula:

[0080] ;

[0081] Obtain the path node number value of the ground feature;

[0082] in, Represents the path node number value of the ground feature. This represents the z-th node number in path i. This represents the z-th node number in path j. Represents the length of path i. This represents the length of path j. Represents the total number of path nodes;

[0083] From the semantic association path structure obtained in the previous step, there are two feature paths: path i (representing the attribute sequence of "historical building A") is [101, 102, 103, 104], and path j (representing the attribute sequence of "historical building B") is [201, 202, 103, 204]. Node numbers 101 represent "building name", 102 represent "construction year", 103 represent "protection level", 104 represent "architectural style", 201 represent "name", and 202 represent "historical features". "Period" and "204" represent "Structure Type". Node number information is extracted sequentially for each path, and a feature node mapping set is established. For example, the node number mapping set for path i is PathI_Nodes={101, 102, 103, 104}, and the node number mapping set for path j is PathJ_Nodes={201, 202, 103, 204}. The feature identifier and path length parameter are then labeled. For example, if the feature identifier for path i is "Historical Building A", its length... (Number of nodes), the feature identifier of path j is "Historical Building B", and its length is... ,in, This represents the node number value of a ground feature path, used to measure the degree of difference between two path node sequences. The smaller the value, the more similar the two paths are. Representative path The first in Each node number Representative path The first in Each node number Calculating the absolute difference between the node numbers at corresponding positions reflects the differences in node numbers. Indicates all The differences between corresponding nodes are summed to reflect the overall cumulative differences along the path. The denominator is... Normalization is performed by the total path length, so that the differences between paths of different lengths can be compared.

[0084] The advantage of this formula is that by calculating the sum of the absolute differences in the path node numbers and dividing by the total length of the two paths, it can quantify the degree of difference between the two paths in terms of node sequences. Normalization ensures a fair comparison of differences between paths of different lengths, thus providing a unified standard for measuring path similarity. In this example, the common length of the two paths is used. (Length of the shorter path) If the path lengths are inconsistent, align the shorter path lengths for comparison, or pad the shorter path with virtual node numbers (e.g., 0) to achieve the same length. ;

[0085] ;

[0086] The result This represents the difference in node numbering sequence between the two feature paths, "Historical Building A" and "Historical Building B". The value of 37.5 indicates that there is a significant difference in node numbering between the two paths, even though there are shared nodes at some locations (such as node 103). The higher the value, the greater the difference between the paths. This feature path node numbering value is the basic data for subsequent intersection comparison. The feature path node numbering value is obtained by calculation.

[0087] The intersection comparison submodule performs an intersection operation on the node number sequences of two feature paths based on the node number values ​​of the feature paths, extracts the endpoint node number in the intersection, counts the number of times each node appears in the differentiated paths, compares it with the path overlap benchmark value, filters out path pairs that meet the conditions, and establishes a set of path intersection numbers that meet the conditions.

[0088] Based on the path node numbers, path i, continuing from the previous step, is [101, 102, 103, 104], and path j is [201, 202, 103, 204]. First, an intersection operation is performed to identify the common node numbers in the two paths. In this example, the intersection result is {103}. The endpoint node number is extracted from the intersection. If the intersection contains multiple nodes, the node that is the endpoint (i.e., the last node in the sequence) on any path is selected. In this example, the intersection only contains 103, which is not the endpoint of any path; therefore, no endpoint node is extracted. Then, the number of times each node appears in the differentiated paths is counted. A differentiated path refers to the path portion remaining after removing the intersection nodes. For example, the differentiated path for path i is [101, 102, 104], and the differentiated path for path j is [201, 202, 204]. In path i... In the differentiation section, 101 appears once, 102 appears once, and 104 appears once. These are compared with the path overlap benchmark value, which is set at 0.3. This benchmark value is determined by statistical analysis of the path overlap characteristics of different historical blocks and expert review. It represents the minimum similarity threshold for determining that two paths have substantial overlap. If the calculated path similarity (e.g., calculated in reverse from the G value calculated in the previous module, or based on the ratio of the number of intersection nodes to the total path length) is higher than the benchmark value, it is considered to meet the condition, and path pairs that meet the condition are selected. In this example, the ratio of the number of intersection nodes of path i and path j to the total number of nodes is 1 / 7≈0.14, which is lower than the benchmark value of 0.3. Therefore, it is not selected as a path pair that meets the condition. If there are path pairs that meet the condition, a set of path intersection numbers that meet the condition is established.

[0089] The path filtering submodule queries the original feature path identifiers of the path intersection number set that meets the conditions, and integrates the feature identifiers and path overlap node information to generate a set of feature overlap paths.

[0090] The set of path intersection numbers that meet the conditions is selected as {Path_AB_overlap: {103, 105}}, where Path_AB_overlap represents the intersection of path A and path B, and the intersection nodes are 103 (protection level) and 105 (building height). Based on the path combinations (Path_A and Path_B) in this set, their original feature path identifiers are queried, that is, Path_A corresponds to "historical building A", and Path_B corresponds to "historical building C". The feature identifiers and path overlap node information are integrated and merged to form a structured data entry, such as {feature identifier A: "historical building A", feature identifier B: "historical building C", overlap nodes: [103, 105]}. This data entry clearly shows which feature paths overlap and the specific overlapping nodes. By performing this query and integration operation on all path intersection number sets that meet the conditions, a set of overlapping feature paths is generated.

[0091] Please see Figure 5 The category attribution determination module includes:

[0092] The category label collection submodule collects the category label set to which each endpoint node belongs based on the endpoint node in the set of overlapping ground features paths, performs index mapping between ground feature paths and their endpoint labels, and generates path endpoint category label groups.

[0093] Based on the endpoint nodes in the set of overlapping ground features, assuming the paths "Historical Building A_Path 1" and "Historical Building B_Path 2" are included, and the endpoint nodes are "Protection Level" and "Building Age", predefined category labels are collected for each endpoint node. For example, the "Protection Level" node belongs to categories such as "Laws and Regulations" and "Management", while the "Building Age" node belongs to categories such as "Time Information" and "Historical Value". The category labels are standard labels in a pre-constructed classification system. An index mapping is performed between the ground feature path and its endpoint label. Then, an association mapping is established between the ground feature path and the collected endpoint label. For example, "Historical Building A_Path 1" is mapped to {"Laws and Regulations", "Management"}, and "Historical Building B_Path 2" is mapped to {"Time Information", "Historical Value"}. This ensures that each overlapping path can be traced back to the semantic category of its endpoint node. By performing this collection and mapping operation on the endpoint nodes in all the set of overlapping ground features, a path endpoint category label group is generated.

[0094] The tag frequency statistics submodule performs a frequency statistics operation on the category tags based on the path endpoint category tag group, records the number of times each category tag appears in the feature path set, and sorts them according to the number of occurrences to obtain a sorted category tag sequence;

[0095] Based on the path endpoint category label group, which contains the following entries: {Path1: {A, B, C}, Path2: {A, D}, Path3: {B, D, E}}, where A, B, C, D, and E represent specific category labels (such as "laws and regulations", "historical value", etc.). First, all the category labels are traversed, and the number of times each category label appears in the set of geographic paths is recorded. For example, the statistical results show that: A appears 2 times, B appears 2 times, C appears 1 time, D appears 2 times, and E appears 1 time. The labels are then sorted according to the number of occurrences. Subsequently, the category labels are sorted in descending order of their number of occurrences. When the number of occurrences is the same, they are sorted in secondary order according to the alphabetical order of the label names. In this example, the sorting result is: {A: 2, B: 2, D: 2, C: 1, E: 1}, finally obtaining the sorted category label sequence.

[0096] The category determination submodule performs a matching judgment on the set of labels corresponding to the endpoint nodes in the feature path according to the sorted category label sequence, selects the label item with the first position in the sorted sequence as the path's category, integrates the feature path's category labels, and obtains the feature category category label group.

[0097] Based on the sorted category label sequence [A, B, D, C, E] (where A, B, and D appear the same number of times and are arranged in alphabetical order), and the label set corresponding to the endpoint node of the feature path "Historical Building X Path" being {B, E, F}, a matching judgment is performed. The label set {B, E, F} of "Historical Building X Path" is traversed and compared with the labels in the sorted sequence [A, B, D, C, E]. The label item that appears first in the sorted sequence for each path is selected as the path's corresponding category. Specifically, it finds which label in {B, E, F} appears first in the sorted sequence [A, B, D, C, E]. In this example, A is not... In {B, E, F}, B is in {B, E, F} and its position in the sorted sequence is 2. D is not in {B, E, F}, C is not in {B, E, F}, and E is in {B, E, F} and its position in the sorted sequence is 5. Therefore, the one that is first in the sorted sequence and exists in the tag set is B. So the category of "Historical Building X Path" is determined to be B. Integrate the category tags of the feature paths and bind "Historical Building X Path" with tag B to form a category mapping relationship of {Historical Building X Path: B}. By performing this matching and filtering operation on the tag set of the endpoint nodes of all feature paths, the feature category category tag group is finally obtained.

[0098] Please see Figure 6 The classification result integration module includes:

[0099] The node extraction submodule assigns a tag group to each land feature category, collects the classification node corresponding to each tag, records the path number associated with the classification node and the number of corresponding land feature path sets, determines the matching index between the category tag and the classification node, and obtains the tag's node number value.

[0100] Based on the feature category attribution label group {Path1: A, Path2: A, Path3: B, Path4: A, Path5: B}, where Path1-5 are feature paths and A and B are category labels, the system first collects the category node corresponding to each category label. For example, category A corresponds to category node NA, and category B corresponds to category node NB. The system records the associated path numbers and the number of corresponding feature path sets for each category node. For instance, for category node NA, the associated path numbers are recorded as {Path1, Path2, Path4}, and the number of corresponding feature path sets is 3. For category node NB, the associated path numbers are recorded as {Path3, Path5}, and the number of corresponding feature path sets is 2. The system then determines the matching index between category labels and category nodes, establishing a mapping index between category labels and category node numbers, for example, {A: NA, B: NB}. This index allows for quick location of the corresponding category node using the category label. By performing this collection, recording, and determination operation on all labels in the feature category attribution label group, the label attribution node number value is obtained.

[0101] The path classification submodule divides the corresponding feature paths into each category node based on the node number value of the tag and the category tag as the classification basis. It establishes a two-way correspondence structure between feature path numbers and node numbers, extracts the path number list of each node, and obtains the number of paths belonging to each node.

[0102] Based on the tag-assigned node ID values ​​{A:NA, B:NB}, all feature paths (Path1, Path2, Path4) with category label A are assigned to category node NA, and all feature paths (Path3, Path5) with category label B are assigned to category node NB. A bidirectional correspondence is established between feature path IDs and node IDs, ensuring that each feature path can be queried to its own category node using its own ID, and also to all feature paths it contains using the category node. For example, the following structure is established: {Path1:NA, Path2:NA, Path3:NB, Path4:NA, Path5:NA, Path6:NA, Path7:NA, Path8:NA, Path9:NA, Path9:NA, Path1:NA, Path2:NA, Path3:NA, Path4:NA, Path5:NA, Path6:NA, Path7:NA, Path8:NA, Path9 ... The mapping relationship is as follows: 4:NA, Path5:NB}. At the same time, {Path1, Path2, Path4} is recorded under node NA, and {Path3, Path5} is recorded under node NB. The path number list to which each node belongs is extracted. From the bidirectional correspondence structure, the list of all feature path numbers contained in each category node is extracted. For example, the path list of NA is {Path1, Path2, Path4}, and the path list of NB is {Path3, Path5}. By performing this division, establishment and extraction operation on the node number values ​​to which all labels belong, the number of node path affiliations is obtained.

[0103] The structure generation submodule integrates the classification nodes and their subordinate feature path numbers through structural mapping based on the number of node paths, outputs the classification node index, corresponding category label and total number of paths, determines the attribution of nodes and feature paths, and generates a classification table of street protection surveying data.

[0104] Based on the number of node paths, where category node NA belongs to paths {Path1, Path2, Path4} and category node NB belongs to paths {Path3, Path5}, a structured mapping and integration is performed on each category node and its subordinate feature path numbers to form a clear hierarchical structure. The output includes the category node index, corresponding category label, and total number of paths. For example, the output index of NA is 1, the corresponding category label is A, and the total number of paths is 3; the output index of NB is 2, the corresponding category label is B, and the total number of paths is 2. This determines the attribution of nodes and feature paths. This output clarifies the category represented by each category node and the specific feature paths it contains. Finally, by performing this mapping, integration, and output operation on the number of all node paths, a street block protection surveying data classification table is generated.

[0105] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A block-oriented protection mapping data intelligent classification system, characterized in that, The system comprises: The ground object structure analysis module obtains ground object elements in surveying and mapping data, extracts ground object attribute labels, judges the first occurrence position and distribution frequency of key attributes, analyzes and restores the arrangement order of ground object attribute, and constructs a ground object semantic unit sequence template; The semantic association modeling module sorts the attribute labels based on the ground object semantic unit sequence template, establishes an attribute directed path from the top layer to the bottom layer according to the sorting result, collects the connection node number and adjacent node relationship of the edge in the path, and constructs a semantic association path structure; The path intersection analysis module extracts the ground object path node sequence according to the semantic association path structure, compares the intersection nodes and counts the end node frequency, filters the overlapping paths, and obtains a ground object overlapping path set; The category attribution determination module collects the category label to which the end node belongs based on the end node in the ground object overlapping path set, matches the path end label after sorting according to the occurrence frequency, and obtains a ground object category attribution label group.

2. The lot-oriented, protection-mapping, data-intelligent classification system of claim 1, wherein, The ground object semantic unit sequence template comprises an attribute label arrangement structure, a key attribute segment set and an attribute frequency weight model, the semantic association path structure comprises an attribute level mapping relationship, a directed path node chain and a node connection relationship set, the ground object overlapping path set comprises a path intersection node set, an end node occurrence frequency distribution and an overlapping node filtering result, and the ground object category attribution label group comprises a category label frequency ranking, a path end category label mapping and an attribution category label matching result.

3. The lot-oriented, protection-mapping, data-intelligent classification system of claim 1, wherein, The ground object structure analysis module comprises: The attribute extraction submodule obtains ground object elements in surveying and mapping data, and performs attribute-level labeling operation thereon, records the index position of the key attribute, classifies the ground object according to the attribute by comparing the first occurrence position of the key attribute with the total number of attributes, and obtains the key attribute index distribution result; The original sequence segment construction submodule extracts attribute segments in the original text of the ground object according to the key attribute index distribution result, intercepts the attribute segments based on the index interval of the attribute in the text, constructs an attribute segment set, and recombines the segments in combination with the ground object information to obtain a key attribute original sequence segment set; The semantic template generation submodule rearranges the attribute segment set in the original order according to the key attribute original sequence segment set, sequences and splices multiple key attributes of the same ground object, integrates the semantic units of each ground object, and obtains a ground object semantic unit sequence template.

4. The lot-oriented, protection-focused, intelligent classification of survey data system of claim 3, wherein, The semantic association modeling module comprises: The hierarchical sorting submodule compares and sorts the attribute nodes according to the hierarchical label priority value of the attribute nodes based on the ground object semantic unit sequence template in combination with the hierarchical label of the ground object attribute node, rearranges the positions of the ground object attributes in order from the top layer to the bottom layer, and obtains an attribute sorting index value; The path construction submodule obtains an adjacent node set in the attribute reordering column according to the attribute sorting index value, respectively numbers and records the connection direction of each pair of adjacent nodes, integrates the structure edge information in combination with the node connection relationship in the path, and generates attribute directed path atlas data; The node structure extraction submodule collects node numbers and adjacent node pair relationships in the connection edges according to the attribute directed path graph data, constructs a node mapping table according to the adjacent structure relationship, stores the upstream and downstream relationship types and connection directions between each attribute, and obtains a semantic association path structure.

5. The lot-oriented, protection-focused, intelligent classification of survey data system of claim 4, wherein, The path intersection analysis module comprises: The path extraction submodule collects node sequence in any two feature path based on the semantic association path structure, extracts node number information in each path in turn and establishes a feature node mapping set, marks the path length parameter and the path belonging to the feature identifier, and obtains the feature path node number value; The intersection comparison submodule performs intersection operation on the node number sequences of the two feature paths according to the feature path node number value, extracts the end node number in the intersection, and counts the number of occurrences of each node in the differentiated path, compares with the path intersection reference value, selects the path combination meeting the conditions, and establishes a path intersection number set meeting the conditions; The path screening submodule queries the original feature path identifier according to the path combination corresponding to the number according to the path intersection number set meeting the conditions, integrates the feature identifier and the path intersection node information, and generates a feature intersection path set.

6. The lot-oriented, protection-focused, intelligent classification of survey data system of claim 5, wherein, The category attribution determination module comprises: The category label collection submodule collects the category label set to which each end node belongs based on the end node in the feature intersection path set, performs index mapping between the feature path and the end label, and generates a path end category label group; The label frequency statistics submodule performs frequency statistics operation on the category label based on the path end category label group, records the number of occurrences of each category label in the feature path set, and sorts the category labels according to the number of occurrences to obtain a sorted category label sequence; The attribution category determination submodule performs matching judgment on the label set corresponding to the end node in the feature path according to the sorted category label sequence, selects the label item in the front position in the sorting sequence as the attribution category corresponding to each path, integrates the attribution label of the feature path, and obtains a feature category attribution label group.

7. The lot-oriented, protection-mapping, data-intelligent classification system of claim 1, wherein, The system further comprises a classification result integration module: The classification result integration module divides the feature path to the corresponding node according to the feature category attribution label group, analyzes the classification corresponding relationship between the node and the feature path, and generates a block protection surveying and mapping data classification table; The block protection surveying and mapping data classification table comprises a classification node identifier, a feature path grouping result and a classification attribution relationship mapping result.

8. The lot-oriented, protection-focused, intelligent classification of survey data system of claim 7, wherein, The classification result integration module comprises: The node extraction submodule collects the classification node corresponding to each label according to the feature category attribution label group, records the path number and the number of corresponding feature path sets in the classification node that have been associated, determines the matching index between the category label and the classification node, and obtains the label attribution node number value; The path classification submodule divides the corresponding ground object path into each classification node according to the classification label based on the label attribution node number value, establishes a bidirectional corresponding structure between the ground object path number and the node number, extracts the path number list attributed to each node, and obtains the node path attribution quantity value; The structure generation submodule integrates the classification node and the subordinate ground object path number according to the node path attribution quantity value, outputs the classification node index, the corresponding classification label and the path total number, determines the attribution of the node and the ground object path, and generates a block protection surveying and mapping data classification table.

Images