Geographic information framework element data updating method and device

Abstract

The invention provides a geographic information framework element data updating method and device, and the method comprises the steps: obtaining remote sensing data for a target region from different data sources, and carrying out the fusion of the remote sensing data, and obtaining fusion data; interpreting the fusion data based on a double-branch interpretation model to obtain an interpretation result corresponding to the fusion data; according to the fusion data, the interpretation result and historical reference data corresponding to the target area, determining each target change element in the target area; and updating the geographic information framework element data corresponding to the target area based on each target change element and the updating priority of each target change element. Through fusion of multi-source remote sensing data and a double-branch interpretation model, construction of a full-process automatic framework and fusion of a multi-modal technology, manual field investigation and vectorization operation are replaced, manual intervention is eliminated, and efficient and accurate updating of geographic information framework element data is achieved.

Description

Technical Field

[0001] The present invention relates to the field of geographic data processing technology, and in particular to a method and device for updating geographic information framework element data. Background Art

[0002] Traditionally, updating feature data within the geographic information framework relies primarily on manual processes and semi-automated tools. Currently, the industry generally adopts three models: The first is a combination of manual field surveys and in-house vectorization. This involves field personnel using Global Positioning System (GPS) devices or mobile devices to collect data on-site, and then manually drawing vector layers using Geographic Information System (GIS) software such as ArcGIS. The second is semi-automated remote sensing interpretation, which uses remote sensing image processing tools such as the Environment for Visualizing Images (ENVI) platform to extract surface feature outlines, but still requires manual correction of the interpretation results to match geographic information standards. The third is automated updates based on rule engines, which extract features using preset rules such as spectral thresholds and texture features. However, due to environmental changes and sensor differences, frequent manual adjustments to the rule parameters are required.

[0003] Although the above method is feasible in specific scenarios, its core process is still mainly based on manual operation, with a low degree of automation, and it is difficult to adapt to high-frequency and large-scale data update needs. Summary of the Invention

[0004] To solve the above problems, the present invention provides a method and device for updating geographic information framework element data.

[0005] The present invention provides a method for updating geographic information framework element data, comprising:

[0006] Acquire remote sensing data for a target area from different data sources, and fuse the remote sensing data to obtain fused data, wherein the remote sensing data includes at least two of satellite image data, lidar point cloud data, multispectral data, and satellite radar data;

[0007] Interpreting the fused data based on a dual-branch interpretation model to obtain an interpretation result corresponding to the fused data;

[0008] determining target change elements in the target area based on the fused data, the interpretation result, and historical benchmark data corresponding to the target area;

[0009] Based on each of the target change elements and the update priority of each of the target change elements, the geographic information framework element data corresponding to the target area is updated.

[0010] According to a method for updating geographic information framework element data provided by the present invention, fusing the remote sensing data to obtain fused data includes:

[0011] In view of the differences in data sources of the remote sensing data, an adaptive correction algorithm is used to dynamically adjust the response curves corresponding to different data sources to generate a standardized raster dataset corresponding to each remote sensing data;

[0012] Based on a spatiotemporal registration algorithm, obtaining aligned data corresponding to each remote sensing data by comparing the spatial reference and time series of the standardized raster dataset corresponding to each remote sensing data;

[0013] The aligned data corresponding to each of the remote sensing data are fused to obtain the fused data.

[0014] According to a method for updating geographic information framework element data provided by the present invention, the interpretation result includes vector data with labels corresponding to each element in the fused data;

[0015] The interpreting the fused data based on the dual-branch interpretation model to obtain an interpretation result corresponding to the fused data includes:

[0016] Based on the geometric contour branch of the dual-branch interpretation model, the fused data is sequentially encoded, key point focused, and decoded to obtain the contour of each element in the fused data;

[0017] Based on the semantic attribute branch of the dual-branch interpretation model, position encoding and global semantic feature extraction are sequentially performed on the fused data to obtain the semantics of each element in the fused data;

[0018] For each of the elements in the fused data, based on the fusion layer of the dual-branch interpretation model, the contour and the semantics of the element are associated to obtain the vector data with the label corresponding to the element.

[0019] According to a method for updating geographic information framework element data provided by the present invention, determining each target change element in the target area based on the fused data, the interpretation result, and the historical benchmark data corresponding to the target area includes:

[0020] Comparing the fused data with historical benchmark data corresponding to the target area to determine initial change elements in the target area;

[0021] According to the interpretation result, pseudo-change elements in each of the initial change elements are eliminated to obtain each target change element in the target area, where the pseudo-change elements are elements that change based on illumination and / or seasonal changes.

[0022] According to a geographic information framework element data updating method provided by the present invention, the geographic information framework element data corresponding to the target area is updated based on each target change element and the update priority of each target change element, including:

[0023] Arrange the target change elements in descending order of the update priority to obtain an update element sequence;

[0024] Get the number of features updated in a batch;

[0025] According to the number of elements, a plurality of target change elements are sequentially extracted from the sequence head of the update element sequence, and the geographic information framework element data corresponding to the target area is updated until the update element sequence is empty.

[0026] According to a method for updating geographic information framework element data provided by the present invention, after interpreting the fused data based on a dual-branch interpretation model and obtaining an interpretation result corresponding to the fused data, the method further includes:

[0027] Based on the construction rules of the knowledge graph corresponding to the geographic information framework element data, the interpretation results and existing elements in the existing geographic database are modeled with a graph neural network to obtain the knowledge graph corresponding to the geographic information framework element data;

[0028] The construction rules include spatial topology rules, semantic association rules and temporal constraint rules.

[0029] According to a method for updating geographic information framework element data provided by the present invention, the method further includes:

[0030] Performing at least one of spatial topology verification, attribute logic verification, and visual comparison verification on the knowledge graph and / or the geographic information framework element data;

[0031] If the verification is passed, the incremental update package corresponding to the knowledge graph and / or the geographic information framework element data is pushed to the geographic information platform and metadata is generated.

[0032] The present invention also provides a device for updating geographic information framework element data, comprising:

[0033] a fusion module configured to obtain remote sensing data for a target area from different data sources and fuse the remote sensing data to obtain fused data, wherein the remote sensing data includes at least two of satellite image data, lidar point cloud data, multispectral data, and satellite radar data;

[0034] an interpretation module, configured to interpret the fused data based on a dual-branch interpretation model to obtain an interpretation result corresponding to the fused data;

[0035] a determination module configured to determine target change elements in the target area based on the fused data, the interpretation result, and historical benchmark data corresponding to the target area;

[0036] The updating module is configured to update the geographic information framework element data corresponding to the target area based on each of the target change elements and the update priority of each of the target change elements.

[0037] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for updating geographic information framework element data as described above is implemented.

[0038] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for updating geographic information framework element data.

[0039] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned methods for updating geographic information framework element data.

[0040] The method and device for updating geographic information framework element data provided by the present invention obtain remote sensing data for a target area from different data sources, and fuse each of the remote sensing data to obtain fused data, wherein each of the remote sensing data includes at least two of satellite image data, lidar point cloud data, multispectral data, and satellite radar data; interpret the fused data based on a dual-branch interpretation model to obtain an interpretation result corresponding to the fused data; determine each target change element in the target area based on the fused data, the interpretation result, and the historical benchmark data corresponding to the target area; and update the geographic information framework element data corresponding to the target area based on each of the target change elements and the update priority of each of the target change elements. By fusing multi-source remote sensing data with a dual-branch interpretation model, replacing manual field surveys and vectorization operations, eliminating manual intervention, and integrating a full-process automation architecture with multimodal technology, efficient and accurate updating of geographic information framework element data is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 It is a flowchart of a method for updating geographic information framework element data provided by the prior art.

[0043] Figure 2 This is one of the flow charts of the geographic information framework element data updating method provided by the present invention.

[0044] Figure 3 It is a structural diagram of the dual-branch interpretation model provided by the present invention.

[0045] Figure 4 It is a schematic diagram of the knowledge graph matching provided by the present invention.

[0046] Figure 5 This is a comparison chart of the incremental update effect provided by the present invention.

[0047] Figure 6 This is the second flow chart of the geographic information framework element data updating method provided by the present invention.

[0048] Figure 7 It is a structural diagram of the geographic information framework element data updating device provided by the present invention.

[0049] Figure 8 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0050] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0051] First, the related technologies provided by the present invention are briefly described.

[0052] See also Figure 1 , Figure 1 It is a flowchart of a method for updating geographic information framework element data provided by the prior art.

[0053] The traditional method of updating geographic information data adopts a linear process and relies mainly on manual intervention. The process starts with field data collection, where field personnel use GPS devices or tablets to collect geographic information data on site, that is, field personnel use handheld devices (GPS, tablets) to collect field data. This is followed by in-house vectorization, where operators use GIS software (such as ArcGIS or QGIS) to convert the collected raw data into a vector format, that is, manually use GIS software (such as ArcGIS) to draw the vector format. In this process, the operator needs to compare the remote sensing image with the existing database, manually adjust the coordinates (offset) and attributes (errors), and form a manual correction link. The corrected data needs to undergo quality review, and professionals will check for topological errors (such as broken roads, unclosed waters, etc.). If problems are found, they need to be repeatedly returned for correction until the requirements are met, that is, the topological errors are manually checked and repeatedly corrected until they pass. Finally, the review is completed, the updated data is imported into the geographic information platform, and the release is completed.

[0054] However, existing technologies have four key flaws: First, efficiency bottlenecks are prominent. Manual field surveys take up to several weeks, semi-automated tools need to repeatedly correct interpretation errors, and the mismatch rate of the rule engine is as high as over 30%, making it difficult to compress the overall update cycle. Second, the labor cost is high. From data collection to topology checking, it relies on professional operations, especially in complex terrain areas (such as mountainous areas and water areas), where labor costs account for more than 70% of the total budget. Third, the dynamic update capability is insufficient. Traditional methods require full processing of geographic data and cannot quickly respond to local changes (such as road damage and new buildings), making it difficult to meet real-time needs such as disaster emergency response and urban planning. Fourth, the data matching accuracy is low. There are differences in the coordinate system and attribute structure of remote sensing interpretation results and the geographic information framework. Manual corrections can easily introduce secondary errors, resulting in frequent problems such as broken roads and water system topology errors.

[0055] In response to the shortcomings of traditional methods, the present invention provides a fully automated, high-precision, high-frequency geographic information framework element updating method and device.

[0056] The following combination Figure 2-Figure 8 The present invention describes a method and device for updating geographic information framework element data.

[0057] Figure 2 This is one of the flow charts of the method for updating geographic information framework element data provided by the present invention, such as Figure 2 As shown, the method includes the following:

[0058] Step 201: remote sensing data for a target area is obtained from different data sources, and each remote sensing data is fused to obtain fused data, wherein each remote sensing data includes at least two of satellite image data, lidar point cloud data, multispectral data, and satellite radar data.

[0059] Specifically, the fused data can be a remote sensing image after data fusion. The target area refers to the area to which the geographic information framework feature data to be updated belongs. The area can be a village, township, county, city, province, country, etc. It can also be a certain set area, such as a desert, river basin, forest, etc.

[0060] In practical applications, in order to improve the reliability, integrity and accuracy of remote sensing data, remote sensing data for the target area can be obtained from different data sources, that is, remote sensing data for the target area can be obtained from at least two data sources.

[0061] Exemplarily, high-resolution satellite image data, lidar point cloud data, multispectral data, and satellite radar data may be received as multi-source remote sensing data.

[0062] Furthermore, multi-source remote sensing data, i.e., remote sensing data from different data sources, can be fused to generate fused data. This not only makes the fused data more comprehensive and complete, but also allows the fused data to incorporate the characteristics of different remote sensing data, thereby improving the accuracy of the fused data.

[0063] For example, taking urban areas as an example, the fused remote sensing data (fused data) can retain the texture details of 0.5-meter optical images, the building elevation information of lidar, and the vegetation coverage characteristics of multispectral data, which can provide high-precision input for subsequent interpretation.

[0064] Step 202: interpreting the fused data based on a dual-branch interpretation model to obtain an interpretation result corresponding to the fused data.

[0065] Specifically, the dual-branch interpretation model consists of two branches: the left branch for geometric contours and the right branch for semantic attributes. The interpretation result is the data obtained by associating contour geometry information with semantic labels, which is essentially vector data with attribute annotations.

[0066] In practical applications, in order to solve the problem that the traditional single model is difficult to balance the geometric accuracy of the contour and the accuracy of the semantic attributes, the embodiment of the present invention uses a dual-branch interpretation model of a dual-branch deep learning architecture to interpret the fused data and obtain the interpretation results.

[0067] Before using the dual-branch interpretation model, you need to train the dual-branch interpretation model. The dual-branch interpretation model uses a joint loss function during the training phase, such as the Dice loss function used in the geometric contour branch and the cross-entropy loss function used in the semantic attribute branch.

[0068] During the inference stage, the dual-branch interpretation model dynamically associates contour geometry information (the outline of the feature) with semantic labels (the semantics of the feature) through the feature fusion layer, and finally outputs vector data with attribute annotations (interpretation results), such as "road polygon + main road label".

[0069] Step 203: Determine target change elements in the target area based on the fused data, the interpretation result, and the historical benchmark data corresponding to the target area.

[0070] Specifically, the historical baseline data can be data related to the latest geographic information of the target area, such as the latest geographic information framework element data of the target area, or image data corresponding to the latest geographic information framework element data of the target area. The target change element refers to the element that has actually changed.

[0071] In practical applications, change detection algorithms, such as the improved Siamese network, can be used to identify areas that need to be updated based on fused data, interpretation results, and historical benchmark data corresponding to the target area, and further identify target change elements in the updated area.

[0072] Step 204: Based on each of the target change elements and the update priority of each of the target change elements, update the geographic information framework element data corresponding to the target area.

[0073] In actual applications, after determining each target change element, the update priority of each target change element can be obtained, such as the update priority of the transportation network > the update priority of public facilities > the update priority of the building complex > the update priority of natural elements, among which natural elements can be lakes, trees, grasslands, etc.

[0074] Furthermore, the geographic information framework element data corresponding to the target area is updated based on each target change element in the order of update priority from high to low.

[0075] The method for updating geographic information framework element data provided by the present invention obtains remote sensing data for a target area from different data sources, and fuses each of the remote sensing data to obtain fused data, wherein each of the remote sensing data includes at least two of satellite image data, lidar point cloud data, multispectral data, and satellite radar data; interprets the fused data based on a dual-branch interpretation model to obtain an interpretation result corresponding to the fused data; determines each target change element in the target area based on the fused data, the interpretation result, and the historical benchmark data corresponding to the target area; and updates the geographic information framework element data corresponding to the target area based on each of the target change elements and the update priority of each of the target change elements. By fusing multi-source remote sensing data with a dual-branch interpretation model, replacing manual field surveys and vectorization operations, eliminating manual intervention, and integrating a full-process automation architecture with multimodal technology, efficient and accurate updating of geographic information framework element data is achieved.

[0076] Optionally, fusing the remote sensing data to obtain fused data includes:

[0077] In view of the differences in data sources of the remote sensing data, an adaptive correction algorithm is used to dynamically adjust the response curves corresponding to different data sources to generate a standardized raster dataset corresponding to each remote sensing data;

[0078] Based on a spatiotemporal registration algorithm, obtaining aligned data corresponding to each remote sensing data by comparing the spatial reference and time series of the standardized raster dataset corresponding to each remote sensing data;

[0079] The aligned data corresponding to each of the remote sensing data are fused to obtain the fused data.

[0080] In practical applications, multi-source remote sensing data such as high-resolution satellite image data, lidar point cloud data, multispectral data and satellite radar data for the target area can be received simultaneously. In view of the differences in multi-source data, adaptive correction algorithms are used, such as combining transfer equations based on physical models with statistical normalization, to dynamically adjust the response curves of different data sources and generate standardized raster data sets corresponding to each remote sensing data.

[0081] Furthermore, because different remote sensing data are captured or acquired at different times and angles, data alignment is required before fusion. This can be achieved through a spatiotemporal registration algorithm, which aligns the spatial references and time series of each remote sensing data set, essentially aligning the standardized raster datasets to produce aligned data corresponding to each remote sensing data set. This resolves data offsets caused by differences in data sources and acquisition times, further improving the accuracy of the fused data. In other words, adaptive correction and registration algorithms can address information inconsistencies caused by differences in resolution and time phase across multiple remote sensing data sources.

[0082] Optionally, the interpretation result includes vector data with labels corresponding to each element in the fused data;

[0083] The interpreting the fused data based on the dual-branch interpretation model to obtain an interpretation result corresponding to the fused data includes:

[0084] Based on the geometric contour branch of the dual-branch interpretation model, the fused data is sequentially encoded, key point focused, and decoded to obtain the contour of each element in the fused data;

[0085] Based on the semantic attribute branch of the dual-branch interpretation model, position encoding and global semantic feature extraction are sequentially performed on the fused data to obtain the semantics of each element in the fused data;

[0086] For each of the elements in the fused data, based on the fusion layer of the dual-branch interpretation model, the contour and the semantics of the element are associated to obtain the vector data with the label corresponding to the element.

[0087] In practical applications, the geometric contour branch is based on an improved U-net network. Residual connections are introduced in the encoder to enhance gradient propagation, which is used to encode the input fused data to obtain encoded data. A channel attention mechanism (Channel Attention Module) is embedded in the decoding stage to prioritize high-frequency detail areas such as roads and building boundaries. That is, key points of the encoded data are focused to obtain key data. After decoding the key data, the pixel-level binary mask, i.e., the contour of each element, is output.

[0088] The semantic attribute branch uses the Vision Transformer (ViT) model to divide the input image (fused data) into blocks and embed them into position codes, that is, the fused data is divided into blocks and position-encoded to obtain position-encoded data; global semantic features are extracted through a multi-layer multi-head self-attention mechanism, that is, full semantic features are extracted from the position-encoded data, and combined with a transfer learning strategy, the model backbone network is pre-trained using open source geographic data such as Open Street Map (OSM), and then the classification head is fine-tuned using small sample target area data to achieve high-precision recognition of the semantics of factors such as road grade and building function, that is, the semantics of each factor are obtained.

[0089] Furthermore, the outputs of the two branches are cross-connected in the fusion layer, the contour geometry information is associated with the semantic attributes, and finally the labeled vector data is output. That is, the contours of each element are associated with the contours of each element through the output layer to obtain the labeled vector data corresponding to each element.

[0090] See also Figure 3 , Figure 3 It is a structural diagram of the dual-branch interpretation model provided by the present invention.

[0091] The dual-branch interpretation model is divided into two branches that process in parallel. The left branch is based on an improved U-Net network and extracts the contours of objects (elements) through an encoder-decoder architecture. The encoder part uses a residual network to enhance feature extraction capabilities, and the decoder part introduces an attention mechanism to focus on key areas and improve the boundary accuracy of elements such as roads and buildings. The left branch processes the input image (fused data) and obtains the contour output (the contour of the element). The right branch is based on the Transformer model. After pre-training, the input image is divided into blocks and embedded into position codes. Then, multi-head self-attention is used to parse the global semantic information of the remote sensing image (fused data) to identify the semantic attributes of the object (element) (such as road grade and building function), that is, to produce semantic output. Afterwards, the features of the two branches are cross-connected in the fusion layer to associate the contour geometry information with the semantic attributes, and finally output the labeled vector data.

[0092] The embodiment of the present invention designs a dual-branch deep learning architecture, integrates the detail capture capability of U-Net and the global semantic understanding advantage of Transformer, and simultaneously optimizes the geometric contours and semantic attribute extraction of land objects, thereby achieving a simultaneous improvement in geometric precision and attribute accuracy, thereby improving interpretation accuracy.

[0093] Optionally, determining each target change element in the target area according to the fused data, the interpretation result, and historical benchmark data corresponding to the target area includes:

[0094] Comparing the fused data with historical benchmark data corresponding to the target area to determine initial change elements in the target area;

[0095] According to the interpretation result, pseudo-change elements in each of the initial change elements are eliminated to obtain each target change element in the target area, where the pseudo-change elements are elements that change based on illumination and / or seasonal changes.

[0096] In practical applications, when identifying target change elements, pixel-level detection can be performed first: using a modified Siamese network to compare the latest imagery (fused data) with historical baseline data, a change probability map is generated to identify the initial change elements in the target area. Object-level verification is then performed: combining the vector outlines output by the interpretation model (the outlines of each element in the interpretation results) to eliminate spurious change elements caused by changes in lighting and / or seasons (such as temporary vegetation cover), thereby obtaining the target change elements. This ensures the accuracy and reliability of the target change elements, further improving the accuracy of feature data updates in the geographic information framework.

[0097] Optionally, updating the geographic information framework element data corresponding to the target area based on each target change element and its update priority includes:

[0098] Arrange the target change elements in descending order of the update priority to obtain an update element sequence;

[0099] Get the number of features updated in a batch;

[0100] According to the number of elements, a plurality of target change elements are sequentially extracted from the sequence head of the update element sequence, and the geographic information framework element data corresponding to the target area is updated until the update element sequence is empty.

[0101] Specifically, the number of elements refers to the maximum number of target change elements updated in each update.

[0102] In practice, after identifying all target change elements, the topology (element outlines) and semantics of the GIS feature data are updated in batches according to a preset priority queue (transportation network > public facilities > building complexes > natural elements), ensuring that key elements are processed first. This means that based on the number of elements and the order of update priority for each target change element, the topology and semantics of the GIS feature data are updated in batches.

[0103] For example, in disaster emergency scenarios, priority is given to updating damaged roads and shelters, and updating of secondary buildings is postponed.

[0104] It should be noted that during the update process of geographic information framework feature data, version control technology can be used to retain historical data snapshots to support backtracking and exception recovery.

[0105] Optionally, after interpreting the fused data based on the dual-branch interpretation model to obtain an interpretation result corresponding to the fused data, the method further includes:

[0106] Based on the construction rules of the knowledge graph corresponding to the geographic information framework element data, the interpretation results and existing elements in the existing geographic database are modeled with a graph neural network to obtain the knowledge graph corresponding to the geographic information framework element data;

[0107] The construction rules include spatial topology rules, semantic association rules and temporal constraint rules.

[0108] In practical applications, in order to achieve seamless integration of interpretation results with existing geographic databases, a geographic information knowledge graph can be constructed.

[0109] First, three types of construction rules are defined: spatial topology rules, such as roads must be connected to intersections or other roads, and building bases must not overlap with water areas; semantic association rules, such as road names must comply with administrative district naming standards, and roads near hospital buildings must be main roads; and temporal constraint rules, such as new features must be compatible with historical version data, and buildings cannot reappear at their old locations after demolition.

[0110] Based on these construction rules, a graph neural network (GNN) is used to model the structure of the interpreted results and existing features in the existing geographic database. Nodes represent geographic features and their attributes, and edges represent spatial or semantic relationships. During the matching process, the GNN calculates the topological similarity and semantic consistency between newly added nodes (such as interpreted road A) and existing nodes (such as road B and intersection C) at the same location. It automatically completes coordinate offset correction (such as extending the endpoint of road A to the nearest intersection) and attribute mapping (such as assigning standard names to unnamed roads based on the administrative district database), ultimately generating a knowledge graph.

[0111] Topological similarity is determined based on the size of the outlines of newly added and existing nodes, as well as the relationship between the images. Semantic consistency includes consistency in name, size, and width. The administrative district database is used for updating road names and assigning new names.

[0112] See also Figure 4 , Figure 4This is a schematic diagram of the knowledge graph matching provided by the present invention: the knowledge graph matching realizes the precise mapping of interpretation results and existing geographic elements through the graph structure. The nodes of the knowledge graph are divided into geographic elements (such as roads, buildings, water systems) and attributes (name, level, function), and the edge relationship defines the spatial topology (such as road connection, water area inclusion) and semantic association (such as the administrative district to which the building belongs). Taking the matching of the newly added road (Road_NEW) as an example, the Road_NEW in the interpretation result is not connected to the existing geographic database. The topological similarity of Road_NEW and the existing roads (such as Road_1 and Road_2) and intersections (Intersection_3) is calculated through the graph neural network, and automatically extended to the best connection position, that is, spatial connection is performed to avoid manual adjustments in traditional methods. Among them, Road_1 and Road_2 are connected through Intersection_3.

[0113] The embodiments of the present invention achieve minute-level local data updates and support dynamic updates of knowledge graphs by constructing a knowledge graph-driven factor matching mechanism. Through the dynamic matching mechanism of knowledge graph and GNN, the association rules of geographic factors are converted into graph structures, and the reasoning ability of GNN is used to achieve accurate factor-level mapping, reducing the amount of manual correction by 90%.

[0114] Optionally, the method further includes:

[0115] Performing at least one of spatial topology verification, attribute logic verification, and visual comparison verification on the knowledge graph and / or the geographic information framework element data;

[0116] If the verification is passed, the incremental update package corresponding to the knowledge graph and / or the geographic information framework element data is pushed to the geographic information platform and metadata is generated.

[0117] In practical applications, the updated target data (knowledge graph and / or geographic information framework feature data) can be subject to three levels of verification: spatial topology verification, which is used to check hard constraints such as road connectivity and water area closure based on the rule engine; attribute logic verification, which is used to check the compatibility of feature attributes with the knowledge graph (such as school buildings must not be marked as industrial land); visual comparison verification, which is used to automatically generate before-and-after comparison maps of the changed area, such as superimposing new and old road lines, for quick review by technical personnel.

[0118] After verification, the incremental update package of the target data can be pushed to the geographic information platform through the presentation state transfer application programming interface (RESTfulAPI), and metadata (such as update time, change area range, processing log) can be generated at the same time to ensure data traceability.

[0119] See also Figure 5 , Figure 5This is a comparison chart of the incremental update effect provided by the present invention: The present invention adopts an incremental update mechanism (using incremental update packages to push to the geographic information platform to update the existing geographic database), which greatly improves efficiency through local updates. Figure 5 The left and right columns are used to compare the effects of the traditional method and the present invention. The left column shows the full area update of the traditional method (the whole area is covered with gray shadow update), which requires reprocessing the entire map data and takes 7 days. It can be seen that the traditional method is time-consuming and wastes resources; the right column is the local incremental area update of the present invention, which is detected and updated, and the remaining areas remain as they are, that is, only the changed area is updated, which takes 2 hours; the update ratio is 100% vs. 5%.

[0120] The embodiment of the present invention supports priority-driven incremental updates through a closed loop of change detection and conflict resolution, updating only 5% to 20% of the changed areas, which is 5 to 10 times more efficient than traditional full updates and supports high-concurrency data processing.

[0121] Alternatively, for complex conflicts, such as where new buildings overlap with historical water system data, local reinterpretation can be initiated, combining multi-temporal remote sensing data with a rule engine to generate an optimal solution, replacing traditional manual intervention. Specifically, if topologically contradictory elements arise during the updating of GIS framework data and / or the construction of a knowledge graph, the local area within the target region containing these topologically contradictory elements can be reinterpreted using a dual-branch interpretation model. This updated interpretation result can then be used to update the GIS framework data and / or the constructed knowledge graph.

[0122] The embodiments of the present invention can resolve conflicts and contradictions by introducing local re-interpretation and automatically repairing spatial topology errors.

[0123] The following combination Figure 6 The method for updating geographic information framework element data provided by the present invention is further explained.

[0124] See also Figure 6 , Figure 6This is the second flow chart of the method for updating geographic information framework element data provided by the present invention: a fully closed-loop automated process can be adopted to significantly reduce manual intervention. First, multi-source remote sensing data is acquired, including high-resolution satellite image data, lidar point cloud data, multispectral data and satellite radar data, etc., and then input in parallel, that is, multi-source data input. The multi-source remote sensing data is processed by the intelligent interpretation link, and the contours and semantic attributes of the objects are simultaneously extracted through the dual-branch interpretation model, replacing the traditional manual vectorization step. The interpretation results then enter the dynamic matching link, based on the geographic information knowledge graph and graph neural network (GNN), to automatically correct the coordinate offset and resolve attribute conflicts, such as automatically connecting new roads to the existing road network, that is, conflict handling. The incremental update link monitors the changed area through a change detection algorithm, and only updates local data according to a preset priority queue (such as main roads take precedence over secondary buildings). At the same time, a local re-interpretation of the topological conflict area is initiated through a conflict handling mechanism, that is, the changed area is updated according to priority after detection, and conflicts are automatically handled. Finally, the automated verification phase uses the spatial topology rule engine to check data consistency. Once confirmed, it is directly published to the geographic information platform, that is, it is directly published after the rule engine checks.

[0125] The geographic information framework feature data update method provided by this embodiment of the present invention achieves multi-dimensional breakthroughs in the field of geographic information data updating by integrating a fully automated architecture with intelligent technologies. Compared to traditional methods, its core benefits include increased efficiency, cost optimization, breakthrough accuracy, and comprehensive improvements in scenario applicability.

[0126] In terms of efficiency, the traditional update process often takes several weeks to complete regional data updates due to its reliance on manual field surveys and repeated revisions. However, the geographic information framework element data update method provided by the embodiment of the present invention compresses the data processing cycle to the hourly level through intelligent interpretation and incremental update mechanisms. In particular, it can quickly respond to and generate update results for sudden changes in geographic information (such as road damage and building collapse after natural disasters), providing real-time data support for emergency decision-making. For example, in a disaster scenario, key damaged areas can be identified and updated first in a very short time, significantly shortening the delay from data acquisition to actual application.

[0127] In terms of cost control, manual data collection, vectorization, and correction processes in traditional models consume the vast majority of the budget. However, the GIS feature data updating method provided by this embodiment replaces manual operations with automated modules, significantly reducing reliance on manpower. The technical team only needs to review and confirm key points, reducing manpower investment to a fraction of that of traditional processes. This also reduces the frequency of field surveys and equipment wear and tear, resulting in orders of magnitude reduction in overall costs.

[0128] In terms of accuracy improvement, traditional semi-automatic tools are susceptible to environmental interference due to their reliance on manual experience and static rules, leading to element misjudgment or topological errors. The geographic information framework element data update method provided by the embodiment of the present invention, through the collaborative interpretation of the dual-branch interpretation model and the dynamic matching driven by the knowledge graph, can not only accurately capture the contour details of the object, but also deeply understand the semantic attributes of the elements (such as road grade, building function), ensuring that the geometric accuracy and attribute consistency are met simultaneously. For example, the break problem in the road network can be automatically repaired to avoid the secondary deviation that may be introduced by manual correction. At the same time, the attribute logic contradiction (such as mislabeling "school" as "industrial area") is verified through the semantic rule library to ensure data quality from the source.

[0129] In addition, the geographic information framework element data update method provided by the embodiment of the present invention demonstrates excellent scenario adaptability and scalability. Whether it is high-density urban areas, scattered rural settlements, or mountainous areas and waters with complex terrain, it can be quickly adapted through adaptive data fusion and transfer learning strategies, supporting the collaborative processing of multi-source remote sensing data (such as optical images, lidar, and radar data). In extended scenarios such as three-dimensional geographic information updates and underground pipe network modeling, the architecture can also be flexibly adjusted to reserve compatible interfaces for future technological evolution.

[0130] It can be seen that the geographic information framework element data updating method provided by the embodiment of the present invention not only reconstructs the technical path of geographic information updating, but also promotes the transformation of the industry from "labor-intensive" to "technology-driven" through full-process closed-loop control and intelligent decision-making, and provides efficient and reliable basic data guarantees for smart cities, natural resource management, public safety and other fields.

[0131] The geographic information framework element data updating device provided by the present invention is described below. The geographic information framework element data updating device described below and the geographic information framework element data updating method described above can be referenced to each other.

[0132] Figure 7 This is a schematic diagram of the structure of the geographic information framework element data updating device provided by the present invention. Figure 7 As shown, the geographic information framework element data updating device includes the following:

[0133] A fusion module 701 is configured to obtain remote sensing data for a target area from different data sources and fuse the remote sensing data to obtain fused data, wherein the remote sensing data includes at least two of satellite image data, lidar point cloud data, multispectral data, and satellite radar data;

[0134] An interpretation module 702 is configured to interpret the fused data based on a dual-branch interpretation model to obtain an interpretation result corresponding to the fused data;

[0135] A determination module 703 is configured to determine target change elements in the target area based on the fused data, the interpretation result, and historical benchmark data corresponding to the target area;

[0136] The updating module 704 is configured to update the geographic information framework element data corresponding to the target area based on each target change element and the update priority of each target change element.

[0137] Optionally, the fusion module 701 is specifically configured to:

[0138] In view of the differences in data sources of the remote sensing data, an adaptive correction algorithm is used to dynamically adjust the response curves corresponding to different data sources to generate a standardized raster dataset corresponding to each remote sensing data;

[0139] Based on a spatiotemporal registration algorithm, obtaining aligned data corresponding to each remote sensing data by comparing the spatial reference and time series of the standardized raster dataset corresponding to each remote sensing data;

[0140] The aligned data corresponding to each of the remote sensing data are fused to obtain the fused data.

[0141] Optionally, the interpretation result includes vector data with labels corresponding to each element in the fused data;

[0142] The interpretation module 702 is specifically configured to:

[0143] Based on the geometric contour branch of the dual-branch interpretation model, the fused data is sequentially encoded, key point focused, and decoded to obtain the contour of each element in the fused data;

[0144] Based on the semantic attribute branch of the dual-branch interpretation model, position encoding and global semantic feature extraction are sequentially performed on the fused data to obtain the semantics of each element in the fused data;

[0145] For each of the elements in the fused data, based on the fusion layer of the dual-branch interpretation model, the contour and the semantics of the element are associated to obtain the vector data with the label corresponding to the element.

[0146] Optionally, the determining module 703 is specifically configured to:

[0147] Comparing the fused data with historical benchmark data corresponding to the target area to determine initial change elements in the target area;

[0148] According to the interpretation result, pseudo-change elements in each of the initial change elements are eliminated to obtain each target change element in the target area, where the pseudo-change elements are elements that change based on illumination and / or seasonal changes.

[0149] Optionally, the updating module 704 is specifically configured to:

[0150] Arrange the target change elements in descending order of the update priority to obtain an update element sequence;

[0151] Get the number of features updated in a batch;

[0152] According to the number of elements, a plurality of target change elements are sequentially extracted from the sequence head of the update element sequence, and the geographic information framework element data corresponding to the target area is updated until the update element sequence is empty.

[0153] Optionally, the geographic information framework element data updating device further includes a construction module configured to:

[0154] Based on the construction rules of the knowledge graph corresponding to the geographic information framework element data, the interpretation results and existing elements in the existing geographic database are modeled with a graph neural network to obtain the knowledge graph corresponding to the geographic information framework element data;

[0155] The construction rules include spatial topology rules, semantic association rules and temporal constraint rules.

[0156] Optionally, the geographic information framework element data updating device further includes a verification module configured to:

[0157] Performing at least one of spatial topology verification, attribute logic verification, and visual comparison verification on the knowledge graph and / or the geographic information framework element data;

[0158] If the verification is passed, the incremental update package corresponding to the knowledge graph and / or the geographic information framework element data is pushed to the geographic information platform and metadata is generated.

[0159] Figure 8 An example of a physical structure diagram of an electronic device is shown below. Figure 8As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute a method for updating geographic information framework element data, the method comprising: acquiring remote sensing data for a target area from different data sources, and fusing each of the remote sensing data to obtain fused data, wherein each of the remote sensing data includes at least two of satellite image data, lidar point cloud data, multispectral data, and satellite radar data; interpreting the fused data based on a dual-branch interpretation model to obtain an interpretation result corresponding to the fused data; determining each target change element in the target area based on the fused data, the interpretation result, and the historical baseline data corresponding to the target area; and updating the geographic information framework element data corresponding to the target area based on each of the target change elements and the update priority of each of the target change elements.

[0160] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0161] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the geographic information framework element data updating method provided by the above methods, the method including: obtaining remote sensing data for the target area from different data sources, and fusing each of the remote sensing data to obtain fused data, each of the remote sensing data including at least two of satellite image data, lidar point cloud data, multispectral data and satellite radar data; interpreting the fused data based on a dual-branch interpretation model to obtain an interpretation result corresponding to the fused data; determining each target change element in the target area based on the fused data, the interpretation result and the historical benchmark data corresponding to the target area; updating the geographic information framework element data corresponding to the target area based on each of the target change elements and the update priority of each of the target change elements.

[0162] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the geographic information framework element data updating method provided by the above-mentioned methods, the method comprising: obtaining remote sensing data for a target area from different data sources, and fusing each of the remote sensing data to obtain fused data, each of the remote sensing data comprising at least two of satellite image data, lidar point cloud data, multispectral data and satellite radar data; interpreting the fused data based on a dual-branch interpretation model to obtain an interpretation result corresponding to the fused data; determining each target change element in the target area based on the fused data, the interpretation result and the historical benchmark data corresponding to the target area; and updating the geographic information framework element data corresponding to the target area based on each of the target change elements and the update priority of each of the target change elements.

[0163] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0164] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0165] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for updating geographic information framework element data, characterized in that: include: Acquire remote sensing data for a target area from different data sources, and fuse the remote sensing data to obtain fused data, wherein the remote sensing data includes at least two of satellite image data, lidar point cloud data, multispectral data, and satellite radar data; Interpreting the fused data based on a dual-branch interpretation model to obtain an interpretation result corresponding to the fused data; determining target change elements in the target area based on the fused data, the interpretation result, and historical benchmark data corresponding to the target area; Based on each of the target change elements and the update priority of each of the target change elements, the geographic information framework element data corresponding to the target area is updated.

2. The method for updating geographic information framework element data according to claim 1, characterized in that: The step of fusing the remote sensing data to obtain fused data includes: In view of the differences in data sources of the remote sensing data, an adaptive correction algorithm is used to dynamically adjust the response curves corresponding to different data sources to generate a standardized raster dataset corresponding to each remote sensing data; Based on a spatiotemporal registration algorithm, obtaining aligned data corresponding to each remote sensing data by comparing the spatial reference and time series of the standardized raster dataset corresponding to each remote sensing data; The aligned data corresponding to each of the remote sensing data are fused to obtain the fused data.

3. The method for updating geographic information framework element data according to claim 1, characterized in that: The interpretation result includes vector data with labels corresponding to each element in the fused data; The interpreting the fused data based on the dual-branch interpretation model to obtain an interpretation result corresponding to the fused data includes: Based on the geometric contour branch of the dual-branch interpretation model, the fused data is sequentially encoded, key point focused, and decoded to obtain the contour of each element in the fused data; Based on the semantic attribute branch of the dual-branch interpretation model, position encoding and global semantic feature extraction are sequentially performed on the fused data to obtain the semantics of each element in the fused data; For each of the elements in the fused data, based on the fusion layer of the dual-branch interpretation model, the contour and the semantics of the element are associated to obtain the vector data with the label corresponding to the element.

4. The method for updating geographic information framework element data according to claim 1, characterized in that: Determining target change elements in the target area based on the fused data, the interpretation result, and historical benchmark data corresponding to the target area includes: Comparing the fused data with historical benchmark data corresponding to the target area to determine initial change elements in the target area; According to the interpretation result, pseudo-change elements in each of the initial change elements are eliminated to obtain each target change element in the target area, where the pseudo-change elements are elements that change based on illumination and / or seasonal changes.

5. The method for updating geographic information framework element data according to claim 1, characterized in that: The updating of the geographic information framework element data corresponding to the target area based on the target change elements and the update priority of the target change elements includes: Arrange the target change elements in descending order of the update priority to obtain an update element sequence; Get the number of features updated in a batch; According to the number of elements, a plurality of target change elements are sequentially extracted from the sequence head of the update element sequence, and the geographic information framework element data corresponding to the target area is updated until the update element sequence is empty.

6. The method for updating geographic information framework element data according to any one of claims 1 to 5, characterized in that: After interpreting the fused data based on the dual-branch interpretation model to obtain an interpretation result corresponding to the fused data, the method further includes: Based on the construction rules of the knowledge graph corresponding to the geographic information framework element data, the interpretation results and existing elements in the existing geographic database are modeled with a graph neural network to obtain the knowledge graph corresponding to the geographic information framework element data; The construction rules include spatial topology rules, semantic association rules and temporal constraint rules.

7. The method for updating geographic information framework element data according to claim 6, characterized in that: The method further comprises: Performing at least one of spatial topology verification, attribute logic verification, and visual comparison verification on the knowledge graph and / or the geographic information framework element data; If the verification is passed, the incremental update package corresponding to the knowledge graph and / or the geographic information framework element data is pushed to the geographic information platform and metadata is generated.

8. A device for updating geographic information framework element data, characterized in that: include: a fusion module configured to obtain remote sensing data for a target area from different data sources and fuse the remote sensing data to obtain fused data, wherein the remote sensing data includes at least two of satellite image data, lidar point cloud data, multispectral data, and satellite radar data; an interpretation module, configured to interpret the fused data based on a dual-branch interpretation model to obtain an interpretation result corresponding to the fused data; a determination module configured to determine target change elements in the target area based on the fused data, the interpretation result, and historical benchmark data corresponding to the target area; The updating module is configured to update the geographic information framework element data corresponding to the target area based on each of the target change elements and the update priority of each of the target change elements.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the method for updating geographic information framework element data as described in any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for updating geographic information framework element data as described in any one of claims 1 to 6 is implemented.

Images