CN-122023981-A - Urban and rural fusion map construction method based on remote sensing data

CN122023981ACN 122023981 ACN122023981 ACN 122023981ACN-122023981-A

Abstract

The invention discloses a construction method of urban and rural fusion maps based on remote sensing data, which relates to the technical field of geographic information and is used for solving the problems of incomplete data integration and ambiguous driving mechanism in urban and rural fusion level evaluation and evolution analysis. The method comprises the steps of carrying out structural labeling on space units, including space-time evolution spectrum coding, functional attribute hierarchical division and space relation vector construction, analyzing a driving mechanism, screening effective driving factors through a geographic detector and a regression model, identifying causal paths, constructing urban and rural fusion knowledge spectrums, carrying out integrity, consistency and effective quality assessment, and obviously improving accuracy and interpretation capability of the spectrums.

Inventors

FENG HAN
MA FANGJIN
SHI JIKANG
ZHENG YUAN
QIU YIJING
LIU JINXUAN
QIN QIANQIAN

Assignees

四川省国土空间规划研究院

Dates

Publication Date: 20260512
Application Date: 20251231

Claims (7)

1. The urban and rural fusion map construction method based on the remote sensing data is characterized by comprising the following steps of: S1, data acquisition and fusion classification, namely acquiring remote sensing image data with at least three time phases, time spans not less than 10 years and spatial resolution greater than 30 meters of a target area, simultaneously acquiring a digital elevation model, socioeconomic statistics and policy text data, identifying the types of ground objects by adopting an object-oriented classification method and combining a convolutional neural network based on CNLUCC standard after preprocessing, coordinate unification and spatial registration, constructing a three-level urban and rural fusion level classification system, and distributing two-digit codes to generate standardized classification grid data; S2, unit feature structured labeling, namely performing space-time evolution spectrum coding, functional attribute hierarchy labeling and space relation vector construction on each space unit, and fusing to generate a multi-dimensional semantic label; s3, analyzing a driving mechanism, namely screening driving factors of urban and rural fusion level evolution, performing geographic detection and heterogeneity analysis, and identifying a time sequence causal path; And S4, constructing an urban and rural fusion knowledge graph based on the urban and rural fusion level classification data of the step S1, the multidimensional semantic tag generated in the step S2 and the obvious causal path identified in the step S3, and performing integrity, consistency and effectiveness verification and comprehensive quality assessment.
2. The urban and rural fusion map construction method based on remote sensing data according to claim 1, wherein the specific implementation process of the step S2 is as follows: Space-time evolution spectrum coding, namely performing space-time superposition operation on urban and rural fusion level classification results of different time phases through a grid calculator, generating an evolution spectrum code of each space unit, identifying the conversion type of the urban and rural fusion level based on the evolution spectrum code, and calculating a space function transfer matrix and a dynamic index; The method comprises the steps of establishing an urban and rural fusion function attribute knowledge base, dividing the function attribute into a primary leading function, a secondary auxiliary function and a tertiary potential function, endowing the space unit with an area smaller than 1 square kilometer with the primary leading function directly based on a classification result, and labeling the secondary auxiliary function and the tertiary potential function based on an area proportion and time sequence analysis; constructing a space relation vector, namely establishing a multi-scale buffer zone by taking a geometric center point of each space unit as a reference, and counting area occupation ratios of different function types to generate a multi-scale space relation vector; And (3) multi-dimensional semantic fusion and label generation, namely carrying out structural organization on space-time evolution spectrum codes, functional attribute level information and space relation vectors, constructing a multi-dimensional semantic label tree structure by adopting an extensible markup language (XML) format, and setting comprehensive confidence coefficient parameters for each attribute value.
3. The urban and rural fusion map construction method based on remote sensing data according to claim 1, wherein the specific implementation process of the space-time evolution map coding is as follows: For the classified data of the initial stage and the final stage of the research period, the evolution map code is obtained through the classified code of the space unit at the initial stage of the research and the classified code of the space unit corresponding to the final stage of the research; Identifying the conversion type of the urban and rural fusion level based on the evolution map coding; carrying out space superposition analysis based on data of at least three time phases to obtain a space function transfer matrix of each time period; for each urban and rural fusion function type, obtaining a single dynamic attitude based on the total area of the initial and final function types of the research period and the length of the research period; Based on the space function transfer matrix, combining the total area of all function types in the initial stage of research and the length of a research period to obtain the comprehensive dynamic attitude of the function type change of the whole research area; The evolution state of each space unit is divided into a rapid expansion type, a slow expansion type, a stable type, a slow shrinkage type and a rapid shrinkage type based on a single dynamic state, codes are allocated, a time sequence evolution state sequence is formed based on the evolution state codes, and a transition point is detected to identify an evolution period.
4. The urban and rural fusion map construction method based on remote sensing data according to claim 1, wherein the specific construction process of the multidimensional semantic tag is as follows: Based on evolution state coding, functional attribute hierarchy labeling and spatial relation vector, and adopting an extensible markup language (XML) format to construct a multi-dimensional semantic tag tree structure, the method comprises the following steps: The first layer is a unique identifier of a space unit and is generated by adopting a UUID (universal unique identifier) coding rule; the second layer is three dimension nodes of evolution state, functional attribute and spatial relationship; The third layer is a specific attribute value and a confidence coefficient parameter of each dimension; Setting a comprehensive confidence coefficient parameter for each attribute value based on the data source quality confidence coefficient, the model prediction confidence coefficient and the expert verification confidence coefficient; when the comprehensive confidence coefficient is lower than a preset threshold value, marking the current space unit as a unit to be verified, preferentially incorporating the unit to be verified into a manual auditing process, and generating a structured database comprising multidimensional semantic tags.
5. The urban and rural fusion map construction method based on remote sensing data according to claim 1, wherein the specific implementation process of the step S3 comprises: The driving factor screening and geographic detection comprises the steps of constructing a urban and rural fusion level evolution driving factor candidate index set comprising natural conditions, traffic location, socioeconomic and policy system, carrying out significance test and explanatory power evaluation on each candidate driving factor based on a geographic detector, carrying out multiple collinearity diagnosis on the driving factors passing the significance test to obtain a variance expansion factor VIF value, eliminating redundant factors with multiple collinearity problems, and judging an effective driving factor set; The geographic weighted regression heterogeneity analysis comprises the steps of adopting a geographic weighted regression model to analyze the space non-stationarity relation between urban and rural fusion state-of-development change and effective driving factors, determining the optimal bandwidth of the model based on a corrected red pool information criterion, selecting a double square kernel function as a space weight function, carrying out local regression on each space unit to obtain a local regression coefficient space distribution diagram of each driving factor, carrying out cluster analysis on regression coefficients, identifying space boundaries of high, medium and low value areas and inversion of regression coefficient signs, and judging the regional difference of the spatial heterogeneity and the action direction of the action intensity of the driving factors.
6. The urban and rural fusion map construction method based on remote sensing data according to claim 1, wherein the specific implementation process of the step S3 further comprises: based on the multidimensional semantic tag system and the driving factor action relation, constructing a time sequence causal path network which takes a function type space unit or a driving factor state as a node and takes a potential causal conduction relation as a side; Searching all paths from a cause unit to a result unit based on a path searching algorithm of time sequence constraint, and checking the consistency of the time sequence causal constraint and the time sequence of the paths to obtain a candidate causal path set; Dividing a research period and a region into a plurality of space-time subsets, respectively verifying causal relation in each subset, counting the proportion of causal establishment, normalizing the F statistic and the average treatment effect value of the Granges causal test, and obtaining causal strength through a weighting formula; And comparing the causal strength with a preset threshold, and only reserving causal paths with causal strength greater than or equal to the preset threshold and marking as obvious causal paths.
7. The urban and rural fusion map construction method based on remote sensing data according to claim 1, wherein the specific implementation process of the step S4 comprises: based on urban and rural fusion classification data, multidimensional semantic tags and obvious causal paths, constructing an urban and rural fusion knowledge map stored by adopting a graph database and relational database hybrid architecture; carrying out integrity check on the map, wherein the integrity check comprises space coverage rate, functional coverage rate and average node degree; consistency check is carried out, wherein the consistency check comprises space logic, time sequence logic and semantic logic consistency check; performing interpretation validity verification, namely comparing a graph reasoning result with an actual situation to obtain interpretation consistency scores by selecting typical phenomena as test cases; Based on the integrity, consistency and effectiveness of urban and rural fusion maps, a weighted fusion formula is combined after normalization processing to obtain comprehensive quality scores, quality grades are divided, and a problem diagnosis and correction mechanism is started for unqualified maps.

Description

Urban and rural fusion map construction method based on remote sensing data Technical Field The invention relates to the technical field of geographic information, in particular to a method for constructing urban and rural fusion maps based on remote sensing data. Background With the acceleration of the remote sensing technology to the multisource collaboration, high space-time resolution and multidimensional sensing direction, the application of urban and rural fusion assessment and evolution analysis is evolving from single-dimensional assessment of traditional statistical data and static maps to intelligent map analysis of multisource remote sensing data driving, space-time dynamic depiction and multidimensional coupling; Meanwhile, the urban and rural fusion map construction technology integrating multi-source remote sensing data cooperative processing, space-time dynamic association analysis and multi-dimensional element visualization has become a core direction for breaking through the traditional research bottleneck and realizing urban and rural fusion closed loops, however, the prior art still has a plurality of defects: In the traditional urban and rural fusion data processing, the analysis mechanism based on a single remote sensing data source and the quantitative fusion logic of multi-source remote sensing data are insufficient, the traditional land utilization classification only focuses on ground object type identification to carry out systematic division, the traditional research is mainly characterized by multi-emphasis macroscopic qualitative description or global linear analysis and space atlas mainly based on static visualization, and the structural semantic annotation on space unit characteristics is lacking, so that the atlas has insufficient interpretation capability and application reliability. In order to solve the above-mentioned defect, a technical scheme is provided. Disclosure of Invention The invention aims to solve the problems of incomplete data integration and unclear driving mechanism in urban and rural fusion level evaluation and evolution analysis, and provides an urban and rural fusion map construction method based on remote sensing data. The aim of the invention can be achieved by the following technical scheme: the urban and rural fusion map construction method based on the remote sensing data comprises the following steps: S1, data acquisition and fusion classification, namely acquiring remote sensing image data with at least three time phases, time spans not less than 10 years and spatial resolution greater than 30 meters of a target area, simultaneously acquiring a digital elevation model, socioeconomic statistics and policy text data, identifying the types of ground objects by adopting an object-oriented classification method and combining a convolutional neural network based on CNLUCC standard after preprocessing, coordinate unification and spatial registration, constructing a three-level urban and rural fusion level classification system, and distributing two-digit codes to generate standardized classification grid data; S2, unit feature structured labeling, namely performing space-time evolution spectrum coding, functional attribute hierarchy labeling and space relation vector construction on each space unit, and fusing to generate a multi-dimensional semantic label; s3, analyzing a driving mechanism, namely screening driving factors of urban and rural fusion level evolution, performing geographic detection and heterogeneity analysis, and identifying a time sequence causal path; And S4, constructing an urban and rural fusion knowledge graph based on the urban and rural fusion level classification data of the step S1, the multidimensional semantic tag generated in the step S2 and the obvious causal path identified in the step S3, and performing integrity, consistency and effectiveness verification and comprehensive quality assessment. As a further improvement of the invention, the specific implementation process of the step S2 is as follows: Space-time evolution spectrum coding, namely performing space-time superposition operation on urban and rural fusion level classification results of different time phases through a grid calculator, generating an evolution spectrum code of each space unit, identifying the conversion type of the urban and rural fusion level based on the evolution spectrum code, and calculating a space function transfer matrix and a dynamic index; The method comprises the steps of establishing an urban and rural fusion function attribute knowledge base, dividing the function attribute into a primary leading function, a secondary auxiliary function and a tertiary potential function, endowing the space unit with an area smaller than 1 square kilometer with the primary leading function directly based on a classification result, and labeling the secondary auxiliary function and the tertiary potential function based on an area proportion and time seque