A Human-Machine Collaborative Remote Sensing Interpretation Method for Complex Urban Buildings

Abstract

This invention discloses a remote sensing interpretation method for complex urban buildings based on human-machine collaboration, belonging to the field of remote sensing information technology. The method includes: receiving remote sensing interpretation requirements input by a user; obtaining analytical results based on a pre-constructed professional domain-adapted interpretation model for complex urban buildings; detecting the user's selection of a specific interpretation mode option when inputting the remote sensing interpretation requirements; determining the current remote sensing interpretation type based on the selection result; inputting the analytical results into a professional domain interpretation resource library for complex urban buildings that is compatible with the current remote sensing interpretation type for feature matching; and outputting accurate remote sensing interpretation results associated with the analytical results. This invention analyzes the user's accuracy requirements through mode options, calls a professional domain interpretation resource library containing building planning standards and material spectral characteristics for feature matching, and achieves accurate interpretation in complex scenarios based on professional data.

Description

Technical Field

[0001] This invention relates to the technical field of remote sensing information technology, and in particular to a remote sensing interpretation method for complex urban buildings based on human-machine collaboration. Background Technology

[0002] With the rapid iteration of high-resolution remote sensing satellites and aerial mapping technologies, as well as continuous breakthroughs in artificial intelligence algorithms, remote sensing interpretation technology has been widely applied in fields such as urban governance and land planning. Remote sensing interpretation models have evolved from traditional rule-driven algorithms to data-driven deep learning models, and some general-purpose remote sensing interpretation models have already acquired certain ground feature recognition capabilities.

[0003] Currently, mainstream remote sensing interpretation technologies mainly focus on the identification of common land features such as roads, green spaces, and water bodies. However, significant shortcomings remain in the interpretation of complex urban buildings and structures (such as dense urban villages, irregularly shaped commercial buildings, and high-rise residential buildings with obstructions). On the one hand, general remote sensing interpretation models lack vertical knowledge support for urban buildings and structures (such as urban planning regulations, spectral characteristics of building materials, and three-dimensional contour constraints), resulting in insufficient accuracy in contour extraction, material identification, and height inversion for buildings and structures in complex scenarios. On the other hand, existing interpretation methods mostly rely on manually labeled samples to train models. When faced with dynamic updates of urban buildings and structures (such as new buildings and demolition areas), the model iteration efficiency is low, and it cannot adapt to real-time interpretation requirements.

[0004] Furthermore, remote sensing interpretation of urban buildings involves classified spatial data and high-value surveying and mapping results. Existing technologies lack sophisticated control mechanisms for interpretation permissions and data access, making it difficult to ensure data security and compliant use. At the same time, interpretation tasks of complex buildings require the combination of professional tools (such as professional domain LiDAR point cloud processing and multi-source data fusion) and vertical resource libraries. However, the existing systems lack sufficient tool invocation and resource matching capabilities, failing to achieve full-link collaboration in the professional domain, from "intent recognition of the professional domain to professional domain tool adaptation to the professional domain to professional domain resource support." This results in interpretation efficiency and accuracy failing to meet the actual needs of urban governance. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0006] In view of the problems existing in the prior art, the present invention is proposed.

[0007] To address the aforementioned technical problems, this invention provides the following technical solution: a remote sensing interpretation method for complex urban buildings based on human-machine collaboration, comprising: The system receives remote sensing interpretation request information input by the user, and performs interpretation intent analysis on the remote sensing interpretation request information based on a pre-built professional domain-adaptive interpretation model for complex urban buildings, and obtains the analysis result, which represents the user's remote sensing interpretation intent. If the analysis result contains a preset interpretation tool trigger identifier, then the remote sensing interpretation intention is to call a special dedicated interpretation tool, call the target dedicated interpretation tool that matches the interpretation tool trigger identifier to perform targeted interpretation processing on the input remote sensing data, and output targeted interpretation processing results. The targeted interpretation processing includes fine extraction of building outlines, material spectrum identification, height dimension inversion, change area detection and / or multi-source data fusion interpretation. If the analysis result does not contain a preset interpretation tool trigger identifier, then the remote sensing interpretation intention is determined to be basic remote sensing interpretation. The user selects the special interpretation mode option when inputting the remote sensing interpretation requirement information, and determines the current remote sensing interpretation type based on the selection result. The analysis result is then input into the urban complex building professional field interpretation resource library that is compatible with the current remote sensing interpretation type for feature matching, and the accurate remote sensing interpretation result associated with the analysis result is output.

[0008] As a preferred embodiment of the human-machine collaborative remote sensing interpretation method for complex urban buildings and structures described in this invention, the step of detecting the user's selection of a specific interpretation mode option when inputting the remote sensing interpretation requirement information, determining the current remote sensing interpretation type based on the selection result, and inputting the analysis result into a professional domain interpretation resource library for complex urban buildings and structures that matches the current remote sensing interpretation type for feature matching includes: The interface for selecting interpretation modes is displayed. The interface includes specialized interpretation mode options, which at least include high-precision professional interpretation options and fast conventional interpretation options. The system detects whether to check or confirm the specialized interpretation mode options when the user inputs the remote sensing interpretation requirement information. If so, based on the operation result, the current remote sensing interpretation type is determined to be the professional domain interpretation of the corresponding mode, and the analysis result is input into the professional domain interpretation resource library of urban complex buildings that is compatible with the special interpretation mode option for feature matching; The method further includes: If not, based on the operation result, the current remote sensing interpretation type is determined to be general basic interpretation. The parsing result is then input into the preset general remote sensing interpretation model for rapid interpretation and response, and the basic remote sensing interpretation result is output.

[0009] As a preferred embodiment of the remote sensing interpretation method for complex urban buildings based on human-machine collaboration described in this invention, the step of detecting whether to perform a selection or confirmation operation for the specialized interpretation mode option when the user inputs the remote sensing interpretation requirement information includes: When the user inputs the remote sensing interpretation request information, whether the user selects or confirms the resource database matching interpretation option and / or the real-time remote sensing retrieval interpretation option.

[0010] As a preferred embodiment of the remote sensing interpretation method for complex urban buildings based on human-machine collaboration described in this invention, the method further includes, before receiving the remote sensing interpretation request information input by the user: Acquire basic data in the professional field of complex urban buildings and structures, extract interpretation feature information from the basic data in the professional field, and store the interpretation feature information in the professional field interpretation resource library of complex urban buildings and structures; When the specialized interpretation mode option is the resource library matching interpretation option, the step of inputting the analysis result into the urban complex building structure professional field interpretation resource library adapted to the specialized interpretation mode option for feature matching, and outputting accurate remote sensing interpretation results associated with the analysis result, includes: The analysis results are input into the professional domain interpretation resource library of complex urban buildings and structures, and compared and matched with the interpretation feature information stored in the library. Based on the matching degree threshold between the analysis results and the interpretation feature information, the professional domain basic data that meet the correlation requirements are selected. The relevant professional domain basic data are integrated with the interpretation analysis results to form accurate remote sensing interpretation results, which are then returned to the user.

[0011] As a preferred embodiment of the human-machine collaborative remote sensing interpretation method for complex urban buildings and structures described in this invention, wherein: when the specialized interpretation mode option is a real-time remote sensing retrieval interpretation option, the step of inputting the analysis result into a specialized interpretation resource library for complex urban buildings and structures that is compatible with the specialized interpretation mode option for feature matching, and outputting accurate remote sensing interpretation results associated with the analysis result, includes: The analysis results are input into a preset third-party remote sensing data engine, and real-time remote sensing data retrieval is performed based on the core information such as the interpretation object and area range in the analysis results. Based on the relevance and timeliness of the search results, remote sensing data sources that meet the preset thresholds are selected. The selected remote sensing data sources are correlated with the interpretation and analysis results, and the data source's acquisition time, resolution, and key information of the data provider are labeled to form accurate remote sensing interpretation results, which are then returned to the user.

[0012] As a preferred embodiment of the remote sensing interpretation method for complex urban buildings based on human-machine collaboration described in this invention, the method further includes, before receiving the remote sensing interpretation request information input by the user: Acquire training data for the professional domain of complex urban buildings, and enhance the original remote sensing interpretation model based on the training data to obtain the pre-constructed professional domain-adapted interpretation model for complex urban buildings. After outputting targeted interpretation results or outputting accurate remote sensing interpretation results associated with the analysis results, the method further includes: Receive user feedback on the interpretation results evaluation information; based on the interpretation results evaluation information, perform incremental learning on the professional domain-adapted interpretation model for the complex urban buildings, and update the model's interpretation feature weights and professional domain adaptation parameters.

[0013] As a preferred embodiment of the remote sensing interpretation method for complex urban buildings based on human-machine collaboration described in this invention, the method further includes, after receiving remote sensing interpretation request information input by the user: Obtain the user's interpretation function permissions and remote sensing data permissions; From the set of specialized interpretation tools for complex urban buildings, obtain specialized interpretation tools that match the interpretation function permissions to obtain the target specialized interpretation tool; from the basic data resources of the professional field of complex urban buildings, obtain remote sensing data that matches the remote sensing data permissions, and construct the professional field interpretation resource library of complex urban buildings that is adapted to the current remote sensing interpretation type using the matched remote sensing data.

[0014] The apparatus for applying the above-mentioned remote sensing interpretation method for complex urban buildings based on human-machine collaboration includes: a demand receiving module for receiving remote sensing interpretation demand information input by the user; The intent parsing module is used to perform interpretation intent parsing on the remote sensing interpretation requirement information based on a pre-built professional domain-adaptive interpretation model for complex urban buildings, and to obtain the parsing result, which represents the user's remote sensing interpretation intent. The tool invocation module is used to determine that the remote sensing interpretation intention is to invoke a special dedicated interpretation tool if the parsing result contains a preset interpretation tool trigger identifier. The module then invokes a target dedicated interpretation tool that matches the interpretation tool trigger identifier to perform targeted interpretation processing on the input remote sensing data and outputs targeted interpretation processing results. The targeted interpretation processing includes fine extraction of building outlines, material spectrum identification, height dimension inversion, change area detection, and / or multi-source data fusion interpretation. The interpretation matching module is used to determine that the remote sensing interpretation intention is basic remote sensing interpretation if the parsing result does not contain a preset interpretation tool trigger identifier, detect the user's selection operation result of the special interpretation mode option when inputting the remote sensing interpretation requirement information, determine the current remote sensing interpretation type based on the selection operation result, input the parsing result into the urban complex building professional field interpretation resource library that is compatible with the current remote sensing interpretation type for feature matching, and output accurate remote sensing interpretation results associated with the parsing result.

[0015] The present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method for remote sensing interpretation of complex urban buildings based on human-computer collaboration.

[0016] The present invention also discloses a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method for remote sensing interpretation of complex urban buildings based on human-computer collaboration.

[0017] The beneficial effects of this invention are as follows: By analyzing the user's interpretation needs and intentions through a professional domain-adaptive interpretation model, corresponding interpretation resources are automatically allocated. If the intention is to call a tool, a high-precision interpretation tool specific to urban buildings and structures is matched; if the intention is to perform basic interpretation, the user's accuracy requirements are analyzed through mode options, and a professional domain interpretation resource library containing building and structure planning specifications and material spectral characteristics is called for feature matching. Accurate interpretation in complex scenarios is achieved based on professional data. In this way, this solution simultaneously realizes professional domain-specific tool interpretation and high-precision resource library-adaptive interpretation for complex urban buildings and structures, solving the core problems of insufficient adaptability and low interpretation accuracy of general remote sensing interpretation technology in scenarios such as dense urban villages and irregularly shaped commercial buildings. Furthermore, through automatic intention recognition and automatic allocation of professional domain resources / tools, the cost of manual processing of remote sensing data is reduced, and the efficiency of remote sensing interpretation in scenarios such as urban planning verification and land surveys is improved. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a schematic diagram of the overall process of the remote sensing interpretation method for complex urban buildings based on human-machine collaboration proposed in this invention. Figure 2This is a deployment architecture diagram of the remote sensing interpretation system for complex urban buildings based on human-machine collaboration proposed in this invention. Figure 3 This is a functional architecture diagram of the remote sensing interpretation system for complex urban buildings based on human-machine collaboration proposed in this invention. Detailed Implementation

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0020] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0021] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0022] Reference Figure 1 As an embodiment of the present invention, a remote sensing interpretation method for complex urban structures based on human-machine collaboration is provided. This method includes the following steps: Step 1: Receive remote sensing interpretation request information input by the user; wherein, the remote sensing interpretation request information is information input by the user through a remote sensing interpretation terminal, a geographic information system (GIS) platform, or a mobile surveying and mapping device, including text descriptions (such as "extract the outline of high-rise residential buildings in a certain area"), coordinates of the area, interpretation accuracy requirements, etc. In some embodiments, the request information can also be structured request information converted from the user's on-site voice commands or hand-drawn area sketches, that is, the user can submit interpretation requests by inputting text + coordinates; or by submitting interpretation requests by voice descriptions or hand-drawn sketches; based on a pre-built professional domain-adaptive interpretation model for complex urban buildings, the remote sensing interpretation request information is analyzed to obtain the analysis results, which represent the user's remote sensing interpretation intention.

[0023] The pre-built domain-specific adaptation interpretation model for complex urban buildings decomposes and analyzes the semantics of demand information based on domain-specific semantic rules (integrating urban planning terminology and building interpretation industry standards), thereby transforming unstructured demand information into structured parsing results. In this step, the domain-specific adaptation interpretation model identifies whether the user needs to invoke a specialized interpretation tool and, if so, which type of tool.

[0024] Specifically, specialized interpretation tools include: a fine-grained building outline extraction tool, a material spectral recognition tool, and a height dimension inversion tool. The fine-grained building outline extraction tool can extract the building edges in dense areas (such as urban villages) in remote sensing images at the sub-pixel level. The material spectral recognition tool can determine the building facade material (such as glass curtain walls / concrete) by matching the spectral curves of remote sensing images. The height dimension inversion tool can calculate the relative height of buildings by combining optical images with LiDAR point cloud data.

[0025] The functional keywords of these specialized interpretation tools can be used as preset interpretation tool trigger identifiers. A professional domain-adaptive interpretation model matches and identifies whether the remote sensing interpretation requirement information contains these trigger identifiers, obtaining the analysis results. The analysis results either contain the preset interpretation tool trigger identifiers or do not contain them. These preset interpretation tool trigger identifiers can be professional terms (such as "sub-pixel contour extraction"), functional keywords (such as "material spectral matching"), or symbolic markers (such as "[height inversion]").

[0026] For example, when a user inputs "to perform material spectrum identification on the glass curtain wall building in XX business district", the professional domain-adaptive interpretation model interprets the interpretation intent and obtains the interpretation result represented as "calling the material spectrum identification tool", where "material spectrum identification" is a preset interpretation tool trigger identifier, and the interpretation result contains this trigger identifier.

[0027] Step 2: If the analysis result contains a preset interpretation tool trigger identifier, then the remote sensing interpretation intent is to call a special dedicated interpretation tool. The target dedicated interpretation tool that matches the interpretation tool trigger identifier is called to perform targeted interpretation processing on the input remote sensing data and output the targeted interpretation processing results. The targeted interpretation processing includes fine extraction of building outlines, material spectrum identification, height dimension inversion, change area detection and / or multi-source data fusion interpretation. For example, a user uploads a 0.5-meter resolution optical remote sensing image of an urban village area and asks, "Extract the sub-pixel-level contours of all buildings with three or more stories in this area." The system calls a domain-specific interpretation model to analyze the interpretation intent. The model recognizes the trigger marker "sub-pixel-level contour extraction" and filters out the corresponding targets based on the requirement of "buildings with three or more stories." Finally, it uses a building contour refinement tool to perform edge detection and sub-pixel fitting on the uploaded image, returning to the user the vector contour data (in .shp format) of the buildings with three or more stories in the area. In this example, the preset interpretation tool trigger marker is "sub-pixel-level contour extraction," the target-specific interpretation tool is the building contour refinement tool, and the targeted interpretation processing refers to extracting the sub-pixel-level contours of buildings with three or more stories. The output of the targeted interpretation processing result is the corresponding building vector contour .shp data.

[0028] In one embodiment, after receiving remote sensing interpretation request information input by the user, the method further includes: Obtain user's interpretation function permissions and remote sensing data permissions; From a collection of specialized interpretation tools for complex urban buildings, dedicated interpretation tools matching the required interpretation function permissions are obtained, resulting in the target-specific interpretation tools. From basic data resources in the field of complex urban buildings, remote sensing data matching the required remote sensing data permissions is acquired, and a specialized interpretation resource library for complex urban buildings adapted to the current remote sensing interpretation type is constructed using the matched remote sensing data. This hierarchical access control system adapts to the unique characteristics of remote sensing scenarios involving urban buildings, ensuring secure access to high-value / confidential remote sensing data while preventing unprofessional users from accidentally operating high-precision interpretation tools.

[0029] The remote sensing data permissions include access to and use of remote sensing image data with a resolution of less than 0.5 meters in the urban core area, remote sensing data of buildings and structures around military management areas, user-owned unpublished building and structure surveying data, and remote sensing data open to public geographic information platforms; the interpretation function permissions include access to the sub-pixel contour extraction tool for buildings and structures, the multi-source data fusion interpretation tool, the height dimension inversion tool, and the change area time series detection tool.

[0030] Users are categorized into government planning and management users, research institution users, and general public users. Administrators allocate and adjust interpretation function permissions and remote sensing data permissions for different user types through the remote sensing interpretation permission management backend. Remote sensing data permissions include: user-owned surveying data can only be accessed and manipulated by the user with a unique ID; remote sensing data from classified areas can only be viewed by government planning and management users; publicly accessible remote sensing data has permission ranges divided according to resolution, with each user corresponding to a unique remote sensing data permission list, allowing for the configuration of specific area data access permissions for individual users. Interpretation function permissions include: breaking down specialized interpretation tools into their smallest functional units and forming a function permission list; each user has an independent function permission list, allowing for the granting of specific tool usage permissions to individual users; when the server receives a tool call request, it verifies the user's function permission list and then provides the corresponding tool service.

[0031] It should also be noted that the automatic mapping triggering rule for user qualifications and permissions is as follows: A one-to-one correspondence between qualifications and permissions is established by connecting the surveying and mapping qualification database and the user real-name authentication system. For government planning and management users: After completing the Class A surveying and mapping qualification verification and jurisdictional management authority filing, access to remote sensing data in classified areas and access to all specialized interpretation tools are automatically triggered, without manual approval. For research institution users: After completing the Class B surveying and mapping qualification verification and research project filing, access to the public planning data sub-database and access to high-dimensional inversion / multi-source data fusion interpretation tools are automatically triggered; access to classified data requires an additional application for project approval documents. For general public users: After completing real-name authentication, access to publicly available remote sensing data (resolution ≥ 1 meter) and access to the fast, conventional interpretation mode are automatically triggered; high-precision professional interpretation permissions require manual application. The triggering process for dynamic permission adjustments is as follows: Permission upgrade trigger: The user submits qualification upgrade proof / project approval documents, the system automatically verifies, the administrator reviews (within 1 working day), and the permissions are automatically updated and synchronized to the user permission list. Permission downgrade trigger: User qualification expires / Project ends → The system automatically freezes the corresponding permissions and sends a notification to remind the user. If the qualification is not updated within 30 days, the user will be permanently downgraded.

[0032] Step 3: If the analysis result does not contain the preset interpretation tool trigger identifier, then the remote sensing interpretation intention is determined to be basic remote sensing interpretation. The result of the user's selection operation of the special interpretation mode option when inputting remote sensing interpretation requirement information is detected, and the current remote sensing interpretation type is determined based on the selection operation result. The analysis result is input into the urban complex building professional field interpretation resource library that is compatible with the current remote sensing interpretation type for feature matching, and the accurate remote sensing interpretation results associated with the analysis result are output.

[0033] This step sets up specialized interpretation mode options and detects the user's interpretation accuracy requirements based on the operation results of these options: when the user selects the high-precision professional interpretation option, a precise interpretation service adapted to the professional domain resource library is provided; when the user selects the fast and general interpretation option, a lightweight basic interpretation service is provided. The high-precision professional interpretation service includes: adapting and training the original remote sensing interpretation model with professional domain feature data of complex urban buildings, and optimizing parameters to improve the model's recognition accuracy of professional domain features such as irregular building outlines and material spectra. Specifically, the model trained with professional domain features possesses exclusive knowledge such as urban building planning standards and material spectral responses, and can accurately identify the building outlines obstructing dense urban villages and the morphological features of irregularly shaped commercial buildings, avoiding confusion with features such as roads and green spaces.

[0034] The high-precision professional interpretation service also includes: matching corresponding interpretation feature information (such as high-rise residential building outline parameters and glass curtain wall spectral curves) in the professional field interpretation resource library, and integrating the matching results with the model interpretation data to return the results; the data in this resource library comes from urban planning filing materials, high-precision surveying samples, and building material standard data, which can improve the accuracy and compliance of the interpretation results by relying on professional data.

[0035] In one embodiment, the system detects the result of a user's selection of a specific interpretation mode option when inputting remote sensing interpretation requirements, determines the current remote sensing interpretation type based on the selection result, and inputs the parsing result into a professional interpretation resource library for urban complex buildings that matches the current remote sensing interpretation type for feature matching, including: The interpretation mode selection interface is displayed, which includes specialized interpretation mode options. These specialized interpretation mode options include at least a high-precision professional interpretation option and a fast general interpretation option. The system detects whether the user selects or confirms the specialized interpretation mode option when inputting remote sensing interpretation requirements. If so, the current remote sensing interpretation type is determined to be the professional domain interpretation of the corresponding mode based on the operation results, and the analysis results are input into the professional domain interpretation resource library of urban complex buildings that is compatible with the special interpretation mode options for feature matching; In this embodiment, when the professional domain-adapted interpretation model identifies that the user does not need to call a dedicated interpretation tool (i.e., the user needs to enable basic remote sensing interpretation), it will further detect whether the user needs high-precision professional interpretation or fast conventional interpretation. If the user needs high-precision professional interpretation, they can select the high-precision professional interpretation option in the professional domain interpretation mode by checking or clicking on the interface. Based on this selection, the system determines that the current interpretation type is high-precision professional interpretation and enables the precise interpretation service adapted to the professional domain resource library, that is, it calls the contour and material feature data in the urban building professional domain interpretation resource library and matches and integrates them with the model interpretation results.

[0036] In this embodiment, if not, the current remote sensing interpretation type is determined to be general basic interpretation based on the operation result. The parsing result is then input into a preset general remote sensing interpretation model for rapid interpretation and outputs a basic remote sensing interpretation result. That is, the parsing result is input into a preset general remote sensing interpretation model for rapid interpretation and outputs a basic building distribution vector map.

[0037] As an example, the specialized interpretation mode option can be set with multiple sub-options, such as resource library matching interpretation option and real-time remote sensing retrieval interpretation option. The above detection of whether the user performs the check or confirmation operation on the specialized interpretation mode option when inputting remote sensing interpretation requirement information includes detecting whether the user performs the check operation on the resource library matching interpretation option and / or the real-time remote sensing retrieval interpretation option.

[0038] In this embodiment, the sub-options of the professional domain interpretation mode include a resource library matching interpretation option and a real-time remote sensing retrieval interpretation option. The resource library matching interpretation option corresponds to a professional domain interpretation resource library for complex urban buildings (including building outline features and material spectral data), while the real-time remote sensing retrieval interpretation option corresponds to a third-party high-resolution remote sensing data engine (capable of acquiring building images updated within the last 30 days). The professional domain interpretation resource library can be further divided into a personal surveying data sub-library and a public planning data sub-library. The personal sub-library stores user-uploaded building surveying data and extracted features, while the public sub-library stores building data filed with urban planning authorities. The public sub-library can be categorized by administrative division and building type, supporting users to select resources from specific regions or building types for matching.

[0039] The following section uses an interpretation resource database for complex urban buildings as an example to further illustrate the high-precision professional interpretation service. Before receiving the remote sensing interpretation request information input by the user, the method also includes: Acquire basic data in the professional field of complex urban buildings and structures, extract interpretation feature information from the basic data in the professional field, and store the interpretation feature information in the professional field interpretation resource library of complex urban buildings and structures; When the specialized interpretation mode option is set to resource library matching interpretation, the analysis results are input into the urban complex building and structure professional domain interpretation resource library that is compatible with the specialized interpretation mode option for feature matching. The output is accurate remote sensing interpretation results associated with the analysis results, including: The analysis results are input into the interpretation resource library for complex urban buildings and structures, and compared and matched with the interpretation feature information stored in the library. Based on the matching degree threshold between the analysis results and the interpretation feature information (such as similarity ≥ 85%), the basic data of the professional field that meet the correlation requirements are selected. The basic data of the related professional field are integrated with the interpretation analysis results to form accurate remote sensing interpretation results, which are then returned to the user.

[0040] In the above embodiments, the basic data of the professional field consists of professional data provided by the city surveying and mapping institute and the planning bureau and verified for accuracy. It is deployed in the remote sensing interpretation server in the form of spatial index storage. When an interpretation request comes in, the feature information in the library can be quickly called to perform professional field adaptation and enhancement on the model interpretation results, thereby improving the recognition accuracy of complex buildings and structures.

[0041] This section uses a third-party remote sensing data engine as an example to further illustrate the high-precision professional interpretation service. When the specialized interpretation mode option is set to real-time remote sensing retrieval and interpretation, the analysis results are input into a specialized interpretation resource library for urban complex buildings that is compatible with the specialized interpretation mode option for feature matching. The output is accurate remote sensing interpretation results associated with the analysis results, including: The analysis results are input into a preset third-party remote sensing data engine, and real-time remote sensing data retrieval is performed based on the core information such as the interpretation object and area range in the analysis results.

[0042] Based on the relevance of the search results (the degree of matching with the shape / material of the object being interpreted) and timeliness (data collection time ≤ 30 days from the present), remote sensing data sources that meet the preset threshold are selected.

[0043] The selected remote sensing data sources are correlated with the interpretation and analysis results, and the data source's acquisition time, resolution, and key information of the data provider are labeled to form accurate remote sensing interpretation results, which are then returned to the user.

[0044] In some embodiments, the method further includes, prior to receiving remote sensing interpretation request information input by the user: Acquire training data for the professional domain of complex urban buildings, and enhance the original remote sensing interpretation model based on the professional domain training data to obtain a pre-constructed professional domain-adapted interpretation model for complex urban buildings. After outputting targeted interpretation results or precise remote sensing interpretation results correlated with the analysis results, the method also includes: Receive user feedback on interpretation results evaluation information; based on the interpretation results evaluation information, perform incremental learning on the professional domain-adapted interpretation model for complex urban buildings and update the model's interpretation feature weights and professional domain adaptation parameters.

[0045] Specifically, the interpretation result evaluation information represents the user's satisfaction with the accuracy of the interpretation results (such as the compliance rate of building outline deviation ≤ 0.5 meters, and the accuracy of material spectrum recognition). If the user reports that the interpretation accuracy of a certain type of building is not up to standard, the model will automatically adjust the interpretation feature weights corresponding to that type of building (such as increasing the edge detection weight of irregular outlines) to improve the interpretation accuracy of subsequent similar complex building scenarios.

[0046] In summary, in the embodiments described in steps one to three above, the remote sensing interpretation request information input by the user is received; the interpretation intent of the request information is analyzed based on a pre-built professional domain-adaptive interpretation model for complex urban buildings, and the analysis result is obtained, which represents the user's remote sensing interpretation intent; if the analysis result contains a preset interpretation tool trigger identifier, it is determined that the interpretation intent is to call a special-purpose interpretation tool, and the target special-purpose interpretation tool matching the trigger identifier is called to perform targeted interpretation processing on the input remote sensing data, and the processing result is output. This processing includes fine extraction of building outlines, material spectrum identification, height dimension inversion, etc.; if the analysis result does not contain a preset interpretation tool trigger identifier, it is determined that the interpretation intent is basic remote sensing interpretation, the user's operation result on the special interpretation mode option is detected, and the current interpretation type is determined based on the result. The analysis result is input into the matching professional domain interpretation resource library for feature matching, and accurate remote sensing interpretation results are output.

[0047] This solution analyzes the user's interpretation needs and intentions through a domain-specific adaptation interpretation model, automatically allocating corresponding interpretation resources: if the intention is to call a tool, it matches a high-precision interpretation tool specific to urban buildings and structures; if the intention is to perform basic interpretation, it analyzes the user's accuracy requirements through mode options, calls a domain-specific interpretation resource library containing building and structure planning standards and material spectral characteristics for feature matching, and achieves accurate interpretation in complex scenarios based on professional data. In this way, this solution simultaneously realizes domain-specific tool interpretation and high-precision resource library-adapted interpretation for complex urban buildings and structures, solving the core problems of insufficient adaptability and low interpretation accuracy of general remote sensing interpretation technology in scenarios such as dense urban villages and irregularly shaped commercial buildings; at the same time, through automatic intent recognition and automatic allocation of domain-specific resources / tools, it reduces the cost of manual processing of remote sensing data and improves the efficiency of remote sensing interpretation in scenarios such as urban planning verification and land surveys.

[0048] To further facilitate understanding of the solution, specific examples are used to describe and illustrate the remote sensing interpretation method for complex urban structures based on human-machine collaboration provided in this application.

[0049] This solution utilizes training data specific to the urban complex building domain to enhance the remote sensing interpretation model, while simultaneously using a domain-specific interpretation resource library to accurately match and enhance the interpretation results. The enhancement processing for the domain-specific adaptation interpretation model includes: enhancing the original remote sensing interpretation model using training data specific to the urban complex building domain (including 0.5-meter resolution remote sensing samples and material spectral annotation data) to improve the model's interpretation capabilities for scenarios such as dense urban villages and irregularly shaped commercial buildings; the collected domain-specific training data undergoes preprocessing such as spatial registration, noise removal, and feature annotation to transform it into a format suitable for model training; and during use, the trained model continuously adjusts the interpretation feature weights based on user feedback on interpretation accuracy, optimizing the accuracy of identifying complex buildings.

[0050] For example, even models trained in specialized domains may confuse the spectral characteristics of glass curtain wall buildings with those of reflective pavements, leading to errors in feature type identification and affecting the accuracy of interpretation results. By providing user feedback on these biases, the model can increase the weighting of the spectral features specific to glass curtain walls, reducing confusion errors in similar scenarios later. This domain-specific interpretation model employs a U-Net+Transformer hybrid architecture, allowing for incremental learning and continuous parameter updates based on user feedback data to improve interpretation accuracy in complex scenarios. The process of introducing a domain-specific interpretation resource library for complex urban buildings to enhance accuracy includes: constructing a domain-specific interpretation resource library to store structured data such as building outline features and material spectral characteristics.

[0051] The incremental learning mechanism employs two trigger thresholds: a feedback magnitude threshold and a precision deviation threshold. The first threshold triggers incremental learning when ≥10 valid precision deviation feedbacks are received for the same type of building (e.g., glass curtain wall commercial buildings). Valid feedback must include annotations of interpretation deviations and comparative appendices of measured data. The second threshold triggers incremental learning when the interpretation precision compliance rate (outline deviation ≤0.5 meters) for a certain type of building is below 90%, regardless of the number of feedbacks.

[0052] The system compiles interpretation accuracy data for various buildings and structures every quarter. If the accuracy compliance rate of a certain type of building is ≥95% for three consecutive quarters, the feedback threshold is raised to 20 to reduce invalid updates; if the accuracy compliance rate is below 85%, the feedback threshold is lowered to 5 to speed up model optimization.

[0053] After triggering incremental learning, the system automatically extracts feature bias information from the feedback data, adjusts the weight values ​​of the corresponding features based on the attention mechanism of the U-Net+Transformer hybrid architecture, fine-tunes the model using the adjusted weight values, and verifies the model through the test set. If the accuracy improvement is ≥3%, the model is updated synchronously to the online model.

[0054] The resource library is equipped with a spatial index retrieval engine, which can quickly locate and match data based on the interpreted regional coordinates and land cover types. For the basic data of the professional field that are entered into the library, it will first be parsed into structured features of "contour parameters + spectral curves + planning attributes". Each type of building corresponds to a set of exclusive features. When a user's interpretation request enters the resource library, the related data will be filtered according to the interpretation object and regional scope in the request and the matching degree with the structured features in the library. For example, in the spatial feature resource pool of this solution, the structured feature data of three typical urban buildings are pre-stored. Each feature corresponds to an exclusive feature matching weight value (the higher the value, the higher the feature representativeness and matching priority): A (high-rise residential buildings) contains contour parameter A (15), spectral feature B (20), and planning attribute C (5); B (irregular commercial buildings) contains contour parameter D (8), spectral feature B (12), and planning attribute E (10); C (self-built houses in urban villages) contains contour parameter F (25), spectral feature G (18), and planning attribute C (8). When a user's interpretation request contains a single feature condition, the system will perform accurate matching and result sorting based on the feature matching weight value: if the request contains "contour parameter A", only A matches, and the complete feature data of A is returned directly; if the request contains "spectral feature B", both A and B match, and A and B are returned in descending order of weight value; if the request contains "planning attribute C", both A and C match, and C and A are returned in descending order of weight value.

[0055] The specific method for determining the feature matching weights mentioned above is as follows: Step 1: Collect sample data covering typical urban buildings, including remote sensing images and measured data of 500 sets of high-rise residential buildings, 300 sets of irregularly shaped commercial buildings, and 800 sets of self-built houses in urban villages. Step 2: Based on the random forest algorithm, perform contribution quantification analysis on three types of features: contour parameters, spectral features, and planning attributes, and calculate the weight of the influence of a single feature on the interpretation accuracy. Step 3: In accordance with the industry standard "Technical Specifications for Urban Remote Sensing Interpretation," calibrate the weight values ​​calculated by the algorithm to ensure that the weight settings meet the needs of industry applications. Step 4: Through 10 rounds of cross-validation, adjust the weight values ​​to achieve the optimal interpretation accuracy, and finally determine the matching weights for each type of feature.

[0056] like Figure 2As shown in the figure, this application provides a deployment architecture diagram of a remote sensing interpretation system for complex urban buildings based on human-machine collaboration. The system adopts a three-level security isolation domain architecture, which is divided into a user interaction domain, a task scheduling domain, and a computing power support domain. All services are deployed on a classified cloud platform. The user interaction domain is deployed on a cloud-native lightweight container cluster, the task scheduling domain is deployed on a private network cluster, and the computing power support domain is deployed on a GPU heterogeneous computing power cluster.

[0057] The three-tiered isolation architecture enables the isolation of classified data and dynamic scheduling of computing power for remote sensing interpretation tasks. It adapts to the interpretation needs of sensitive areas such as urban core areas and military-managed zones, while also supporting elastic scaling to improve response efficiency and data security in complex scenarios. The interpretation interaction portal serves as the sole external entry point to the user interaction domain, providing users with an interface for submitting remote sensing interpretation requests, configuring interpretation parameters, and visualizing results vectors. It also incorporates surveying and mapping qualification verification logic, ensuring that only users with compliant qualifications can submit interpretation requests for sensitive areas.

[0058] As the core unit of the mission scheduling domain, the spatial mission hub undertakes the functions of interpreting intent, resource routing, and permission verification. It routes user requests to a dedicated interpretation computing power pool or third-party remote sensing services, and achieves rapid matching of interpretation features through a spatial data storage pool. The dedicated interpretation computing power pool deploys interpretation models trained in the field of urban buildings and structures, providing core interpretation services such as contour extraction and material recognition. All data interactions are completed through a classified dedicated line.

[0059] like Figure 3 As shown in the figure, this application provides a functional architecture diagram of a remote sensing interpretation system for complex urban buildings based on human-machine collaboration. The system is divided into a front-end interaction layer and a core processing layer. The front-end interaction layer provides users with compliance verification and task collaboration capabilities, while the core processing layer supports the entire chain processing of interpretation tasks.

[0060] The main functions of the user qualification verification module include: Surveying and mapping qualification verification: Connect to the national surveying and mapping qualification database to verify the user's remote sensing interpretation qualification. Only compliant users can submit interpretation requests for sensitive areas. Dynamic permission adaptation: Based on user qualifications and the sensitivity of the interpretation area, dynamic allocation of interpretation function permissions and remote sensing data access permissions is performed.

[0061] The main functions of the interpretation task collaboration module include: Task progress tracking: Real-time display of computing power allocation, feature matching, and result generation progress of the interpretation task, and supports users to adjust parameters midway; Multi-scenario demand iteration: Supports users to refine and interpret their needs through multiple rounds of interaction for complex scenarios such as dense urban villages and irregularly shaped commercial buildings.

[0062] The core processing layer includes a professional domain interpretation engine module, a spatial feature resource pool module, and a professional toolchain module, which serve as the core capability support for the system and enable accurate interpretation of complex buildings and structures.

[0063] The professional domain interpretation engine module mainly includes: Feature-guided training: Based on professional domain features such as the 3D contours and material spectra of urban buildings, the interpretation model is fine-tuned in a guided manner to improve the recognition accuracy of complex scenes; Dynamic computing power scheduling: GPU computing power resources are dynamically allocated according to the complexity of the interpretation task to ensure the efficient execution of large-area interpretation tasks. The inter-feature resource pool module mainly includes: 3D feature indexing: Constructing spatial indexes of features such as the 3D contours and material spectra of urban buildings to achieve millisecond-level feature matching; Spatial data registration: Performing spatial registration and noise removal on multi-source remote sensing data to ensure the consistency and accuracy of interpretation features. The professional toolchain module mainly includes: Image preprocessing toolset: Providing general remote sensing preprocessing functions such as cloud removal and radiometric correction to reduce the impact of noise in the original image on interpretation; Building interpretation toolset: Providing professional tools such as sub-pixel contour extraction and material spectrum recognition to adapt to the interpretation needs of complex urban buildings.

[0064] This embodiment also discloses an apparatus for applying the above-mentioned remote sensing interpretation method for complex urban buildings based on human-machine collaboration, the apparatus comprising: The system comprises three modules: a request receiving module, which receives remote sensing interpretation request information input by the user; an intent parsing module, which parses the remote sensing interpretation request information based on a pre-built professional-domain-adaptive interpretation model for complex urban buildings, obtaining a parsing result that represents the user's remote sensing interpretation intent; and a tool invocation module, which, if the parsing result contains a preset interpretation tool trigger identifier, determines that the remote sensing interpretation intent is to invoke a specific dedicated interpretation tool, invokes the target dedicated interpretation tool matching the interpretation tool trigger identifier to perform targeted interpretation processing on the input remote sensing data, and outputs a targeted interpretation processing result. This includes fine extraction of building outlines, material spectral identification, height dimension inversion, change area detection, and / or multi-source data fusion interpretation. An interpretation matching module is used to determine if the remote sensing interpretation intent is basic remote sensing interpretation if the analysis results do not contain a preset interpretation tool trigger identifier. It also detects the user's selection of specific interpretation mode options when inputting remote sensing interpretation requirements, determines the current remote sensing interpretation type based on the selection results, inputs the analysis results into a professional interpretation resource library for urban complex buildings that matches the current remote sensing interpretation type, performs feature matching, and outputs accurate remote sensing interpretation results associated with the analysis results.

[0065] This embodiment also provides a computer device applicable to the remote sensing interpretation method for complex urban buildings based on human-machine collaboration, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the remote sensing interpretation method for complex urban buildings based on human-machine collaboration as proposed in the above embodiment.

[0066] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0067] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the remote sensing interpretation method for complex urban buildings based on human-machine collaboration, as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0068] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A remote sensing interpretation method for complex urban buildings based on human-machine collaboration, characterized in that, include: The system receives remote sensing interpretation request information input by the user, and performs interpretation intent analysis on the remote sensing interpretation request information based on a pre-built professional domain-adaptive interpretation model for complex urban buildings, and obtains the analysis result, which represents the user's remote sensing interpretation intent. If the analysis result contains a preset interpretation tool trigger identifier, then the remote sensing interpretation intention is to call a special dedicated interpretation tool, call the target dedicated interpretation tool that matches the interpretation tool trigger identifier to perform targeted interpretation processing on the input remote sensing data, and output targeted interpretation processing results. The targeted interpretation processing includes fine extraction of building outlines, material spectrum identification, height dimension inversion, change area detection and / or multi-source data fusion interpretation. If the analysis result does not contain a preset interpretation tool trigger identifier, then the remote sensing interpretation intention is determined to be basic remote sensing interpretation. The user selects the special interpretation mode option when inputting the remote sensing interpretation requirement information, and determines the current remote sensing interpretation type based on the selection result. The analysis result is then input into the urban complex building professional field interpretation resource library that is compatible with the current remote sensing interpretation type for feature matching, and the accurate remote sensing interpretation result associated with the analysis result is output.

2. The remote sensing interpretation method for complex urban buildings based on human-machine collaboration as described in claim 1, characterized in that: The detection process involves the user selecting a specific interpretation mode option when inputting remote sensing interpretation requirements, determining the current remote sensing interpretation type based on the selection result, and inputting the analysis result into a specialized interpretation resource database for urban complex buildings that matches the current remote sensing interpretation type for feature matching. This includes: The interface for selecting interpretation modes is displayed. The interface includes specialized interpretation mode options, which at least include high-precision professional interpretation options and fast conventional interpretation options. The system detects whether to check or confirm the specialized interpretation mode options when the user inputs the remote sensing interpretation requirement information. If so, based on the operation result, the current remote sensing interpretation type is determined to be the professional domain interpretation of the corresponding mode, and the analysis result is input into the professional domain interpretation resource library of urban complex buildings that is compatible with the special interpretation mode option for feature matching; The method further includes: If not, based on the operation result, the current remote sensing interpretation type is determined to be general basic interpretation. The parsing result is then input into the preset general remote sensing interpretation model for rapid interpretation and response, and the basic remote sensing interpretation result is output.

3. The remote sensing interpretation method for complex urban buildings based on human-machine collaboration as described in claim 2, characterized in that: The step of detecting whether to perform a check or confirmation operation on the special interpretation mode option when the user inputs the remote sensing interpretation requirement information includes: When the user inputs the remote sensing interpretation request information, whether the user selects or confirms the resource database matching interpretation option and / or the real-time remote sensing retrieval interpretation option.

4. The remote sensing interpretation method for complex urban buildings based on human-machine collaboration as described in claim 3, characterized in that: Before receiving the remote sensing interpretation request information input by the user, the method further includes: Acquire basic data in the professional field of complex urban buildings and structures, extract interpretation feature information from the basic data in the professional field, and store the interpretation feature information in the professional field interpretation resource library of complex urban buildings and structures; When the specialized interpretation mode option is the resource library matching interpretation option, the step of inputting the analysis result into the urban complex building structure professional field interpretation resource library adapted to the specialized interpretation mode option for feature matching, and outputting accurate remote sensing interpretation results associated with the analysis result, includes: The analysis results are input into the professional domain interpretation resource library of complex urban buildings and structures, and compared and matched with the interpretation feature information stored in the library. Based on the matching degree threshold between the analysis results and the interpretation feature information, the professional domain basic data that meet the correlation requirements are selected. The relevant professional domain basic data are integrated with the interpretation analysis results to form accurate remote sensing interpretation results, which are then returned to the user.

5. The remote sensing interpretation method for complex urban buildings based on human-machine collaboration according to claim 3, characterized in that: When the specialized interpretation mode option is the real-time remote sensing retrieval interpretation option, the step of inputting the analysis result into the urban complex building structure professional field interpretation resource library adapted to the specialized interpretation mode option for feature matching, and outputting accurate remote sensing interpretation results associated with the analysis result, includes: The analysis results are input into a preset third-party remote sensing data engine, and real-time remote sensing data retrieval is performed based on the core information such as the interpretation object and area range in the analysis results. Based on the relevance and timeliness of the search results, remote sensing data sources that meet the preset thresholds are selected. The selected remote sensing data sources are correlated with the interpretation and analysis results, and the data source's acquisition time, resolution, and key information of the data provider are labeled to form accurate remote sensing interpretation results, which are then returned to the user.

6. The remote sensing interpretation method for complex urban buildings based on human-machine collaboration according to claim 1, characterized in that: Before receiving the remote sensing interpretation request information input by the user, the method further includes: Acquire training data for the professional domain of complex urban buildings, and enhance the original remote sensing interpretation model based on the training data to obtain the pre-constructed professional domain-adapted interpretation model for complex urban buildings. After outputting targeted interpretation results or outputting accurate remote sensing interpretation results associated with the analysis results, the method further includes: Receive user feedback on the interpretation results evaluation information; based on the interpretation results evaluation information, perform incremental learning on the professional domain-adapted interpretation model for the complex urban buildings, and update the model's interpretation feature weights and professional domain adaptation parameters.

7. The remote sensing interpretation method for complex urban buildings based on human-machine collaboration according to claim 1, characterized in that: After receiving the remote sensing interpretation request information input by the user, the method further includes: Obtain the user's interpretation function permissions and remote sensing data permissions; From the set of specialized interpretation tools for complex urban buildings, obtain specialized interpretation tools that match the interpretation function permissions to obtain the target specialized interpretation tool; from the basic data resources of the professional field of complex urban buildings, obtain remote sensing data that matches the remote sensing data permissions, and construct the professional field interpretation resource library of complex urban buildings that is adapted to the current remote sensing interpretation type using the matched remote sensing data.

8. A remote sensing interpretation device for complex urban buildings based on human-machine collaboration, applied to the remote sensing interpretation method for complex urban buildings based on human-machine collaboration as described in any one of claims 1-7, characterized in that: The device includes: The demand receiving module is used to receive remote sensing interpretation demand information input by the user; The intent parsing module is used to perform interpretation intent parsing on the remote sensing interpretation requirement information based on a pre-built professional domain-adaptive interpretation model for complex urban buildings, and to obtain the parsing result, which represents the user's remote sensing interpretation intent. The tool invocation module is used to determine that the remote sensing interpretation intention is to invoke a special dedicated interpretation tool if the parsing result contains a preset interpretation tool trigger identifier. The module then invokes a target dedicated interpretation tool that matches the interpretation tool trigger identifier to perform targeted interpretation processing on the input remote sensing data and outputs targeted interpretation processing results. The targeted interpretation processing includes fine extraction of building outlines, material spectrum identification, height dimension inversion, change area detection, and / or multi-source data fusion interpretation. The interpretation matching module is used to determine that the remote sensing interpretation intention is basic remote sensing interpretation if the parsing result does not contain a preset interpretation tool trigger identifier, detect the user's selection operation result of the special interpretation mode option when inputting the remote sensing interpretation requirement information, determine the current remote sensing interpretation type based on the selection operation result, input the parsing result into the urban complex building professional field interpretation resource library that is compatible with the current remote sensing interpretation type for feature matching, and output accurate remote sensing interpretation results associated with the parsing result.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the remote sensing interpretation method for complex urban buildings based on human-machine collaboration as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the remote sensing interpretation method for complex urban buildings based on human-machine collaboration as described in any one of claims 1 to 7.

Images