CN-122019793-A - Cadastral data retrieval method and system based on knowledge graph

CN122019793ACN 122019793 ACN122019793 ACN 122019793ACN-122019793-A

Abstract

The embodiment of the invention provides a cadastral data retrieval method and a cadastral data retrieval system based on a knowledge graph, which belong to the technical field of cadastral data processing, and by carrying out structural analysis on historical religion semantics in a retrieval request, request contents such as a historical number, a historical right person, a historical sitting, a historical adjacent relation and the like can be converted into data constraint which can be directly called by subsequent graph retrieval. The method comprises the steps of constructing a time sequence cadastre knowledge graph by using land parcel state data, land parcel evolution event data and land parcel adjacency relation data in the same cadastre service database, generating an evolution corridor around seed state nodes, and executing condition association propagation and candidate scoring processing in the evolution corridor, so that the retrieval process is not limited to static field matching, but current situation continuation land corresponding to historical semantics can be gradually positioned along the land parcel evolution relation, and an output result not only has higher accuracy, but also has definite continuation basis and interpretability.

Inventors

ZHANG HU
WANG HUI
WANG QIAN
LI ZHE
YANG HONG
LI BIN
ZHANG LIANG
WEI SHAOGUANG

Assignees

成都市自然资源调查利用研究院(成都市卫星应用技术中心)

Dates

Publication Date: 20260512
Application Date: 20260409

Claims (10)

1. A cadastral data retrieval method based on a knowledge graph is characterized by comprising the following steps: obtaining retrieval request data, religion state data, religion evolution event data and religion adjacent relation data in the same cadastral service database, and carrying out structural analysis on the retrieval request data to obtain an entity seed set, a time constraint set and a relation constraint set; Constructing a time sequence cadastral knowledge graph based on the land parcel state data, the land parcel evolution event data and the land parcel adjacency relationship data, and determining a historical land parcel state corresponding to the entity seed set as a seed state node; Generating an evolution corridor along a religion evolution relation in the time sequence cadastral knowledge graph based on the entity seed set, the time constraint set and the relation constraint set, and extracting a candidate religion state set based on the evolution corridor; Based on the time constraint set and the relation constraint set, performing conditional relevance propagation and relevance calculation on each candidate parcel state in the candidate parcel state set to obtain a candidate ranking result; And backtracking a historical relay chain based on the candidate sequencing result, executing consistency judgment by combining the time constraint set and the relation constraint set, and outputting a target cadastral retrieval result and a corresponding evidence chain when judging that the preset condition is met.
2. The knowledge-graph-based cadastral data retrieval method according to claim 1, wherein the step of performing structural analysis on the retrieval request data to obtain an entity seed set, a time constraint set and a relationship constraint set includes: Extracting at least one of a historical land number, a historical right person identifier, a historical sitting description and a historical adjacent relation description in the search request data to form a request element set; Based on the request element set identification history entity identification, forming an entity seed set, identifying time description information, forming a time constraint set, identifying relation description information and forming a relation constraint set; And encoding the entity seed set, the time constraint set and the relation constraint set to form a query representation, and storing the query representation in association with the entity seed set, the time constraint set and the relation constraint set.
3. The knowledge-graph-based cadastral data retrieval method of claim 1, wherein constructing a time-series cadastral knowledge graph based on the religion state data, the religion evolution event data, and the religion adjacency relationship data, and determining a historical religion state corresponding to the entity seed set as a seed state node, comprises: According to the occurrence time of the religion evolution event in the religion state data, carrying out state expansion on each religion entity in the religion state data to form a plurality of religion state nodes, wherein each religion state node corresponds to an effective time interval; Configuring state numbers, land numbers, boundary geometries, areas, entitlement states and seating description attributes for each land state node; based on the religion evolution event data, establishing a state continuous edge, a segmentation edge, a merging edge, a renumbering edge and a boundary adjustment edge between each religion state node; Establishing adjacent edges between the land parcel state nodes which can correspond in time based on the land parcel adjacent relation data so as to form the time sequence cadastral knowledge graph; Mapping the historical entity identification in the entity seed set to a corresponding historical religion state node in the time sequence cadastral knowledge graph, and determining the religion state node obtained by mapping as the seed state node.
4. The knowledge-graph-based cadastral data retrieval method of claim 1, wherein generating an evolution corridor along a religious evolution relationship in the time-series cadastral knowledge graph based on the entity seed set, the time constraint set and the relationship constraint set comprises: taking the seed state node as a starting point, and executing constrained expansion along a state continuous edge, a segmentation edge, a merging edge, a renumbering edge and a boundary adjustment edge in the time sequence cadastral knowledge graph; calculating a succession mapping strength for each evolving edge pointing from the parent state node to the child state node; The time constraint set is used for eliminating evolution edges which are in semantic conflict with the request time, and the relation constraint set is used for restraining the expansion direction which does not meet the relation condition; and reserving nodes and edges with the relay mapping strength meeting preset conditions to form the evolution corridor.
5. The knowledge-based cadastral data retrieval method of claim 4, wherein extracting a set of candidate geodetic states based on the evolutionary corridor comprises: Calculating path inheritance strength for each candidate path from the seed state node to any reachable parcel state node; determining evolution inheritance values of the religion state nodes based on the inheritance mapping strength products of the candidate paths; determining a religion state node with an evolution inheritance value not lower than a preset threshold value as a candidate religion state node, and determining a corresponding node set as the candidate religion state set; and storing evolution inheritance values corresponding to the candidate religion state nodes in association with the evolution corridor.
6. The knowledge-graph-based cadastral data retrieval method of claim 2, wherein performing conditional associative propagation and relevance computation on each candidate geodetic state in the set of candidate geodetic states based on the set of time constraints and the set of relationship constraints to obtain a candidate ranking result, comprises: Based on the query representation, the time constraint set and the relation constraint set, performing conditional association propagation on nodes and associated edges thereof in the candidate religion state set to obtain a context representation of each candidate religion state under the current retrieval request condition; Respectively calculating semantic matching values, evolution inheritance values, adjacency relation satisfaction and entitlement state continuous satisfaction of each candidate land state; Calculating a candidate land state relevance score according to the semantic matching value, the evolution inheritance value, the adjacency satisfaction and the entitlement state continuous satisfaction; and sorting the candidate religion state sets according to the relevance scores from high to low to form the candidate sorting results.
7. The knowledge-based cadastral data retrieval method according to claim 1, wherein backtracking the history relay chain based on the candidate ranking result, comprising: selecting a preset number of high-score candidate land states in the candidate sorting result to form a high-score candidate set; for each candidate religion state in the high-score candidate set, backtracking all feasible paths between it and the seed state node from the evolution corridor; and calculating an evidence chain score for each feasible path, determining the feasible path with the largest evidence chain score as a historical relay chain corresponding to the candidate religion state, and outputting the historical relay chain in association with the candidate religion state.
8. The knowledge-graph-based cadastral data retrieval method of claim 7, wherein performing a consistency determination in conjunction with the set of time constraints and the set of relationship constraints comprises: Extracting an upstream state set, area information and adjacency relation information in a corresponding historical relay chain aiming at the candidate land states with highest scores in the candidate sorting result; calculating an area relay residual error based on the upstream state set and the area information, and calculating adjacency consistency based on the relation constraint set and the actual adjacency of the candidate land state; and under the condition that the area relay residual is not larger than a preset area threshold value and the adjacency relation consistency is not smaller than a preset relation threshold value, determining that the candidate land parcel state meets a consistency judging condition.
9. The cadastral data retrieval method based on the knowledge graph according to claim 8, wherein the target cadastral retrieval result and the corresponding evidence chain are output when the fact that the preset condition is met is judged, and the method specifically comprises: Outputting a current religion identifier, a historical relay chain, a key evolution event sequence, a relation matching result and a result confidence corresponding to the candidate religion state when the highest score candidate religion state in the candidate sequencing result meets the consistency judging condition; When the highest score candidate land parcel state does not meet the consistency judgment condition, selecting the next candidate land parcel state according to the candidate sequencing result sequence to repeatedly execute consistency judgment until the candidate land parcel state meeting the consistency judgment condition is obtained; And outputting a target land parcel result which does not search to meet the historical relay constraint when all the candidate land parcel states do not meet the consistency judging condition.
10. A knowledge-graph-based cadastral data retrieval system for performing the knowledge-graph-based cadastral data retrieval method of any one of claims 1 to 9, the system comprising: the analysis unit is used for acquiring retrieval request data, and religion state data, religion evolution event data and religion adjacent relation data in the same cadastral service database, and carrying out structural analysis on the retrieval request data to obtain an entity seed set, a time constraint set and a relation constraint set; A construction unit, configured to construct a time-series cadastral knowledge graph based on the land parcel state data, the land parcel evolution event data, and the land parcel adjacency relationship data, and determine a historical land parcel state corresponding to the entity seed set as a seed state node; the generation unit is used for generating an evolution corridor along a religion evolution relation in the time sequence cadastral knowledge graph based on the entity seed set, the time constraint set and the relation constraint set, and extracting a candidate religion state set based on the evolution corridor; The sorting unit is used for executing condition association propagation and relevance calculation on each candidate religion state in the candidate religion state set based on the time constraint set and the relation constraint set so as to obtain a candidate sorting result; And the output unit is used for backtracking the historical relay chain based on the candidate sorting result, executing consistency judgment by combining the time constraint set and the relation constraint set, and outputting a target cadastral retrieval result and a corresponding evidence chain when the judgment meets the preset condition.

Description

Cadastral data retrieval method and system based on knowledge graph Technical Field The invention relates to the technical field of cadastral data processing, in particular to a cadastral data retrieval method and system based on a knowledge graph. Background The cadastral data retrieval is a basic supporting link in natural resource management, real estate registration, land reclamation, planning verification and historical archival retrieval. In actual business, the search object is typically represented by land, rights, sitting descriptions, address relationships, and various types of registration status information. In the prior art, the common cadastral data retrieval mode mainly comprises a field retrieval mode based on land parcel number, a fuzzy matching mode based on rights or sitting description and a common map retrieval mode based on association relation. Under a general scene, when the search request is consistent with the religious state in the current cadastral library, the method can more directly return the target result. However, during cadastral management, the religion is not always in a static state. The same land may correspond to different land numbers, different boundary forms and different entitlement states at different times, affected by business events such as segmentation, merging, renumbering, boundary adjustment, entitlement change, etc. At this time, the service personnel may still describe the historical religion number, the historical rights person or the historical adjacent relation when initiating the search, and the target result is in the current religion state, thereby forming obvious dislocation between the search condition and the cadastral current state data. The existing field-based retrieval mode can only be used for matching the current storage field, and when the retrieval condition belongs to the history semantics and the result belongs to the current semantics, the situation of incapacity or mishit easily occurs. Although the common map retrieval mode can express the association relation between religions, the association relation between static entities is still biased in many implementations, continuous expression of the evolution of the religion state along with time is lacking, and particularly, the capability of restraining and judging the succession relation in complex evolution processes such as segmentation, merging, renumbering and the like is lacking. As such, in cases where a historical religion has been evolved multiple times into multiple current religions, it is often difficult for the system to accurately identify a true successor religion from among multiple candidates. Therefore, how to accurately search the corresponding current situation relay land by a search system aiming at the historical semantic input such as the historical land number, the historical right person description or the historical adjacent relation description under the condition that only the kernel data in the same cadastral service database are used, and meanwhile, a traceable relay basis is provided, and still the technical problem to be solved in the cadastral data search field is still needed. Disclosure of Invention The invention aims to provide a cadastral data retrieval method based on a knowledge graph, which at least solves the technical problem that the current situation of a religion can not be accurately positioned based on historical semantics after the religion is subjected to segmentation, merging, renumbering or boundary adjustment in the existing cadastral data retrieval mode. In order to achieve the above object, a first aspect of the present invention provides a cadastral data retrieval method based on a knowledge graph, including: obtaining retrieval request data, religion state data, religion evolution event data and religion adjacent relation data in the same cadastral service database, and carrying out structural analysis on the retrieval request data to obtain an entity seed set, a time constraint set and a relation constraint set; Constructing a time sequence cadastral knowledge graph based on the land parcel state data, the land parcel evolution event data and the land parcel adjacency relationship data, and determining a historical land parcel state corresponding to the entity seed set as a seed state node; Generating an evolution corridor along a religion evolution relation in the time sequence cadastral knowledge graph based on the entity seed set, the time constraint set and the relation constraint set, and extracting a candidate religion state set based on the evolution corridor; Based on the time constraint set and the relation constraint set, performing conditional relevance propagation and relevance calculation on each candidate parcel state in the candidate parcel state set to obtain a candidate ranking result; And backtracking a historical relay chain based on the candidate sequencing result, executing consistency judgment by combining