Surveying and mapping result data processing method and device based on intelligent identification and related equipment

Abstract

The invention relates to the technical field of intelligent surveying and mapping and data management, discloses a surveying and mapping result data processing method and device based on intelligent identification, equipment and a medium, and aims to solve the problem that home-entry surveying and mapping result data cannot be subjected to automatic intelligent verification and consistency conflict resolution in the prior art. Normativity and integrity of collected data are ensured from the source by introducing a dynamic surveying and mapping list; discrete data is converted into a computable structured model through automatic modeling, and a foundation is laid for subsequent processing; an independent and traceable verification source is provided for actually measured data through computer visual identification and estimation; based on a conflict resolution strategy of a precision weight, data contradictions can be objectively and automatically corrected, artificial experience judgment is replaced, the improvement points act synergistically, automatic and intelligent counting and summarizing of home-entry surveying and mapping multi-source result data are jointly achieved, and the overall quality, efficiency and reliability of surveying and mapping result data processing are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent surveying and data management technology, and in particular to a method, apparatus, equipment and medium for processing surveying and mapping results data based on intelligent recognition. Background Technology

[0002] Surveying and mapping data processing refers to the process of organizing data obtained from surveying and mapping real estate in fields such as real estate registration, urban renewal, and engineering management, and ultimately forming standardized reports. It is a key link in the digitization of real estate registration, property appraisal, and other businesses. Taking the inventory and summarization of on-site surveying and mapping results as an example, the quality of its processing directly affects the fairness and efficiency of key businesses such as property rights confirmation and asset appraisal.

[0003] Traditional techniques primarily employ a manual workflow for data processing: field personnel collect data using various tools, while office staff rely on personal experience to manually correlate, compare, and integrate scattered measurement data, attribute records, and photographs using computer-aided design or office software, ultimately compiling a report. However, the inventors discovered that because this process relies entirely on manual judgment, it cannot automatically and systematically cross-validate and logically verify multi-source data (e.g., dimensions measured by professional tools and visual information reflected in media evidence). This not only makes it difficult to effectively identify inherent conflicts between different data sources but also lacks an objective and effective decision-making mechanism to resolve conflicts when they are discovered. Ultimately, this results in low data processing efficiency, poor controllability of the quality of the compiled results, and a high degree of reliance on personal experience for accuracy.

[0004] Therefore, how to achieve automated and intelligent verification and consistency conflict resolution of multi-source survey data has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] This invention provides a method, device, computer equipment, and medium for processing surveying and mapping results data based on intelligent recognition. By guiding structured data collection through a dynamic surveying and mapping list and utilizing computer vision and a conflict resolution mechanism based on precision weights, it achieves automated verification and intelligent correction of multi-source surveying and mapping data, thereby ensuring the accuracy and efficiency of results inventory and summarization.

[0006] Firstly, a method for processing surveying and mapping results data based on intelligent recognition is provided, comprising: generating and sending a dynamic surveying and mapping list of a target property to a mobile terminal; collecting target data through the mobile terminal based on the dynamic surveying and mapping list, and attaching corresponding spatial location tags and target component attributes to the target data through the mobile terminal, wherein the target data includes measured dimensions and media evidence data containing calibration references; automatically associating the target data with corresponding nodes in a preset property data model based on the spatial location tags and target component attributes to construct a preliminary model; performing computer vision recognition on the media evidence data associated with the preliminary model to identify the calibration references and the property components to be measured, and estimating the target key dimensions of the property components based on the identified calibration references; comparing the target key dimensions with the corresponding measured dimensions of the corresponding components recorded in the preliminary model based on a preset accuracy weight associated with the data source, and determining whether there is a conflict in the comparison results; if the above determination is yes, initiating a corresponding conflict resolution strategy to correct the preliminary model according to the accuracy weight; and automatically generating a surveying and mapping results summary report based on the corrected property data model.

[0007] Secondly, a data processing device for surveying and mapping results based on intelligent recognition is provided, comprising: a first distribution module, used to generate and distribute a dynamic surveying and mapping list of target properties to a mobile terminal; a first acquisition module, used to acquire target data through the mobile terminal based on the dynamic surveying and mapping list, and to attach corresponding spatial location tags and target component attributes to the target data through the mobile terminal, wherein the target data includes measured dimensions and media evidence data containing calibration references; a first construction module, used to automatically associate the target data with corresponding nodes in a pre-set property data model based on the spatial location tags and the target component attributes to construct a preliminary model; and a first estimation module, used to... The system performs computer vision recognition on the media evidence data associated with the preliminary model to identify the calibration reference and the property component to be measured, and estimates the target key dimensions of the property component based on the identified calibration reference. A first comparison module compares the target key dimensions with the corresponding measured dimensions of the corresponding component recorded in the preliminary model based on a preset accuracy weight associated with the data source, and determines whether there is a conflict in the comparison results. A first correction module, if the above determination is correct, initiates a corresponding conflict resolution strategy to correct the preliminary model based on the accuracy weight. A first summary module automatically generates a surveying results summary report based on the corrected property data model.

[0008] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0009] Fourthly, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the steps of the above-described method.

[0010] The aforementioned solution, implemented using intelligent recognition-based surveying and mapping data processing methods, devices, computer equipment, and storage media, constructs a closed-loop intelligent data processing workflow through key improvements such as introducing dynamic surveying and mapping lists to guide structured data acquisition, automatic modeling based on spatial semantics, automated cross-validation of computer vision and measured data, and intelligent conflict resolution based on precision weights. Specifically, the dynamic surveying and mapping list ensures the standardization and integrity of the collected data from the source; automatic modeling transforms discrete data into a computable structured model, laying the foundation for subsequent processing; computer vision recognition and estimation provide an independent and traceable source of verification for the measured data; and finally, the conflict resolution strategy based on precision weights can objectively and automatically correct data contradictions, replacing manual judgment. These improvements work synergistically to achieve automated and intelligent inventory and summarization of multi-source surveying and mapping data, fundamentally solving the core technical problems of low efficiency, difficulty in consistency verification, and poor controllability of accuracy caused by reliance on manual labor in traditional methods. This significantly improves the overall quality, efficiency, and reliability of surveying and mapping data processing. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention 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.

[0012] Figure 1 A flowchart illustrating the surveying and mapping results data processing method based on intelligent recognition provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the first sub-process of the surveying and mapping result data processing method based on intelligent recognition provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the second sub-process of the surveying and mapping result data processing method based on intelligent recognition provided in an embodiment of the present invention; Figure 4 A schematic block diagram of a surveying and mapping result data processing device based on intelligent recognition provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention; Figure 6 This is another structural schematic diagram of a computer device according to one embodiment of the present invention. Detailed Implementation

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0015] This invention provides a method for processing surveying and mapping data based on intelligent recognition. The method can be applied to computer devices including but not limited to smartphones, tablets, desktop computers, servers, and cloud computing, and can be used for air quality monitoring in scenarios including but not limited to operating rooms, sterile preparation workshops, biological laboratories, and ICU wards.

[0016] The following detailed description of some embodiments of the present invention is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0017] Please see Figure 1 , Figure 1 This is a flowchart illustrating the surveying and mapping data processing method based on intelligent recognition provided in an embodiment of the present invention. Figure 1 As shown, the method includes, but is not limited to, the following steps S11-S18: S11. Generate and send a dynamic survey list of the target property to the mobile terminal.

[0018] Interpretatively, a target property refers to a single physical property unit that is identified and requires data collection during real estate registration, surveying, or valuation operations.

[0019] A dynamic surveying checklist represents a set of structured data collection instructions that can be dynamically generated based on task requirements. This checklist digitally defines the data items to be collected, their data types, relationships, and collection constraints. This means the checklist content can adaptively adjust to the type of target property (e.g., residential, commercial), the business scenario (e.g., real estate registration, demolition assessment), and known basic property information (e.g., floor plans). For example, for residential properties, the checklist might focus on collecting room dimensions and door / window attributes; while for commercial properties, the checklist might additionally require collecting storefront dimensions and the location of power supply facilities.

[0020] Based on the above description and concept, a dynamic survey list of the target property is generated and sent to the mobile terminal, mainly including the following two core steps: 1) Inventory Generation: After receiving the surveying task, the system first retrieves a matching basic template from a pre-set inventory template library based on the task metadata (such as business type and target property ID). Subsequently, the system can instantiate and dynamically adjust this basic template based on known information about the target property (such as the floor plan obtained from the real estate database). For example, it automatically pre-populates the inventory with known room numbers and component lists, and configures the data items to be collected for each room node (such as configuring "length," "width," "east wall photo," and "window attributes" under the "master bedroom" node). Finally, it generates an executable dynamic surveying inventory that precisely matches the current target task. 2) Inventory Distribution: The generated dynamic surveying inventory is pushed via wireless network (such as 4G / 5G) to a dedicated application on the mobile terminal (such as a smartphone, tablet, or dedicated surveying terminal) held by field personnel. This ensures that field personnel can immediately access standardized work instructions, thereby standardizing data collection behavior from the source and laying the foundation for subsequent automated data processing.

[0021] S12. Based on the dynamic mapping list, target data is collected through the mobile terminal, and corresponding spatial location tags and target component attributes are attached to the target data through the mobile terminal. The target data includes measured dimensions and media evidence data containing calibration references.

[0022] Interpretatively, target data refers to the core set of data collected during the on-site surveying process, in accordance with the specifications of the dynamic surveying list, which describes the spatial and attribute information of the target property. It is the original data source for the subsequent construction of the property digital model.

[0023] Spatial location tags are digital markers used to uniquely identify data collection points in space. These tags are generated by fusing data from position and attitude sensors integrated into the mobile terminal (such as GPS, BeiDou, Wi-Fi positioning, Bluetooth beacons, and inertial measurement units, IMU). Their content can be a set of coordinates (such as latitude and longitude), a relative position relative to a reference point (such as "5 meters from the entrance, east wall"), or a room number on a pre-set floor plan (such as "living room"). Their function is to give each piece of data spatial semantics.

[0024] A target component refers to a physical part of the target property that has a specific function and structure, such as a wall, a window, a door, a beam, or a switch or socket. Target component attributes represent structured parameters used to describe the characteristics of the target component, such as {Component Type: "Window", Material: "Aluminum Alloy", Opening Method: "Sliding", Room: "Master Bedroom"}. These attributes are typically recorded by selecting from a predefined list or by manually entering them.

[0025] Dimensional measurement refers to physical dimensional data about a target property or its components that are directly obtained through professional surveying tools (such as laser rangefinders or Bluetooth rangefinders). For example, the length, width, and height of a room, or the width and height of doors and windows. This data is transmitted to a mobile terminal via Bluetooth or manual input.

[0026] A calibration reference is an object intentionally placed in the frame during the capture of media evidence data, whose physical size is known. Its core function is to provide a scale benchmark for subsequent computer vision size estimation. For example, a calibration reference can be, but is not limited to, the mobile terminal itself, a standard-sized mapping calibration board, or a coin with a known diameter.

[0027] Media evidence data refers to digital files that record the visual state of a target property and its components in the form of images or videos, captured by a mobile terminal camera. The core requirement is that each frame must contain a reference object and the target component to be recorded, so as to ensure that the media data is not only used for archiving, but also as a measurable data source.

[0028] Based on the above concept and setup, the core of this step in executing the on-site data collection loop for the dynamic mapping inventory lies in achieving synchronous collection, structuring, and association of multi-source data through mobile terminals. Specifically, this includes, but is not limited to, the following: 1) Data collection: Field personnel operate mobile terminals to complete the data collection item by item according to the inventory guide. For measurement dimensions, this is obtained by calling or connecting to professional surveying tools; for media evidence data, this is obtained by capturing images with a camera, ensuring that the image includes the calibration reference object and the target component. 2) Data attachment (binding): During or after the collection process, the mobile terminal application performs the following operations: a) Automatically calls the sensors to generate the current spatial location label; b) Provides an interactive interface to guide the user to select or input target component attributes for the currently operated target (such as the window being photographed). Furthermore, within the system, the synchronously collected measurement dimensions, media evidence data, automatically generated spatial location labels, and manually selected target component attributes are associated and bound to form a structured data package. Therefore, by integrating the three actions of data collection, positioning, and description through mobile terminals, the "immediate automatic structuring" of on-site data collection is achieved, obtaining multi-source data packets with inherent correlation and rich semantics, fundamentally eliminating data isolation and correlation errors.

[0029] S13. Based on the spatial location label and the target component attribute, the target data is automatically associated with the corresponding node in the preset real estate data model to construct a preliminary model.

[0030] Explained, a pre-defined property data model refers to a standardized, hierarchical, tree-structured digital framework defined in the backend system before data processing begins. This framework logically represents the physical composition of a property. Essentially, the model is a node tree with parent-child relationships. The top-level node is the "Property" root node, which can sequentially contain child nodes such as "Floor," "Room," and "Building Components" (e.g., doors, windows, beams, columns). Each node is a container that can hold its corresponding attribute data (e.g., dimensions, materials) and media data (e.g., photos). Those skilled in the art can construct this model as follows: based on industry standards (e.g., the "Property Measurement Standard") or business requirements, define a general data pattern and implement this fixed hierarchical logic in the system using a database (e.g., a parent-child table structure in a relational database) or data structure (e.g., JSON, XML). For example, an empty model template can be represented as: Property -> Floor 1 -> [Room 1, Room 2...] -> [Walls, Doors, Windows...].

[0031] The preliminary model represents the first complete digital instance of real estate, containing specific measurement values ​​and media evidence, formed by automatically filling the target data collected from the front end, which has been labeled with spatial location tags and target component attributes, into the corresponding nodes of the aforementioned pre-set real estate data model according to the semantics indicated by its tags and attributes.

[0032] Based on the above concept and setup, after receiving the structured data packet uploaded by the front end, the system performs the following steps: 1) Parsing location information: The system parses the spatial location labels (e.g., master bedroom - east wall) and target component attributes (e.g., {component type: "window"}) in the data packet. 2) Node path matching: The system performs step-by-step matching and navigation in the pre-set property data model based on the parsed information. For example, based on the master bedroom, the system locates the "master bedroom" room node, and then locates or dynamically creates an "east wall - window" component sub-node under that room node based on the east wall and window. 3) Data mounting: The system mounts the target data (i.e., measurement dimensions and media evidence data) in the data packet as attribute data to the finally located component sub-node. 4) Model instantiation: After all data packets have been processed in this way, the originally empty property data template is filled with specific data, thereby constructing a preliminary model reflecting the current true state of the target property. Thus, through a semantic-based automatic association mechanism, the data processing flow is fully automated, resulting in an order-of-magnitude improvement in efficiency. At the same time, it completely avoids errors such as mismatch and omission that may be caused by manual association, providing a high-quality and timely data foundation for the digital management of real estate data.

[0033] S14. Perform computer vision recognition on the media evidence data associated with the preliminary model to identify the calibration reference and the property component to be measured, and estimate the target key dimensions of the property component based on the identified calibration reference.

[0034] Interpretatively, a property component to be tested refers to a corresponding architectural element that has been recorded in the preliminary model and whose corresponding media evidence data includes a clear image of the component available for visual analysis. For example, if a photograph is attached to the "Master Bedroom - East Wall" node in the preliminary model, then the "window" appearing in that photograph is the property component to be tested.

[0035] The target critical dimension refers to the physical dimensions that characterize the core geometric features of the building component under test, derived from media evidence data through calculation. These dimensions are indirectly estimated through image analysis and are mainly used for data cross-validation. For example, for a window, the target critical dimensions are usually the width and height of the window frame.

[0036] Computer vision recognition refers to the technology of automatically processing, analyzing, and understanding digital images or videos using computer algorithms to identify specific targets within them. Those skilled in the art can achieve the recognition function required in this step as follows: Employing a deep learning-based object detection model (such as YOLO, SSD, or Faster R-CNN), a large amount of image data containing various calibration references (such as mobile phones, calibration boards) and real estate components (such as windows, doors) is collected and labeled as a training set. The model is trained to accurately select (i.e., identify) image regions of these target objects from any input media evidence data. The trained model is then deployed on a server or in the cloud. When recognition processing is required, the system automatically calls the model to infer the media evidence data associated with the initial model, outputting the bounding box coordinates and category labels of all identified objects.

[0037] Based on the above concept and description, this step provides a specific and feasible computer vision recognition and size estimation scheme, and clarifies its accuracy assurance measures to ensure that its output can be used for subsequent automated conflict resolution. The specific implementation includes, but is not limited to, the following steps: 1) Target recognition using a pre-trained object detection model: a) Specific implementation of model training: To achieve high-precision recognition, the object detection model adopts the YOLOv5 architecture. Its training dataset is specifically built for real estate surveying scenarios, containing over 50,000 labeled images. The dataset covers various calibration references (such as different models of smartphones, standard A4 paper, credit cards, and dedicated surveying calibration boards) and real estate components (such as various doors, windows, beams, columns, switches, and sockets). Images were collected under various lighting conditions, shooting angles, and partially occluded scenarios to enhance model robustness. b) Labeling and performance requirements: Training labels use pixel-level instance segmentation (e.g., using variants of architectures such as Mask R-CNN or YOLACT) to obtain accurate target contours, rather than coarse bounding boxes. The performance requirement for this model on the reserved test set is an average accuracy (mAP@0.5) of no less than 95%. This accuracy metric is the primary technical guarantee for the accuracy of subsequent size estimation.

[0038] 2) Pixel size calculation and scale conversion based on accurate contours: a) The system uses the accurate contours (masks) output by the model to calculate the "pixel size of the largest bounding rectangle" of the calibration reference object and the building component to be measured in the image. This avoids errors introduced by the bounding box including the background or inaccurate positioning. b) The scale conversion relationship is calculated by "Scale conversion relationship (mm / pixel) = Known physical size of reference object (mm) / Pixel size of reference object (pixels)".

[0039] 3) Target Key Dimension Estimation and System Error Control: a) Based on the scale conversion relationship and the pixel size of the target property component, its physical size is calculated as the estimated target key dimension. b) Accuracy Guarantee Mechanism: To control estimation errors, this embodiment of the invention introduces "shooting specifications" and "system error calibration." "Shooting Specifications": When collecting media evidence data, the mobile terminal application will prompt the user through built-in sensors to keep the camera's optical axis perpendicular to the surface of the measured component as much as possible to minimize errors caused by perspective distortion. "System Error Calibration": Before system deployment, a large number of tests are conducted using standard-sized objects at different distances and angles. Statistical analysis shows that the average relative error of this visual estimation scheme can be controlled within ±3%. This system error can serve as an important reference for setting subsequent "preset tolerance thresholds."

[0040] Therefore, it endows the media evidence data in the preliminary model, which was originally only used for qualitative archiving, with the ability to conduct quantitative verification. Through automated CV recognition and size estimation, it generates traceable verification data that is independent of traditional measurement tools, thus providing an objective basis for discovering potential contradictions between "measured dimensions" and "visual evidence" and realizing automated and intelligent quality inspection of surveying and mapping results.

[0041] S15. Based on a preset accuracy weight associated with the data source, the target key dimension is compared with the corresponding measured dimension of the corresponding component recorded in the preliminary model, and it is determined whether there is a conflict in the comparison result.

[0042] Interpretationally, the pre-set precision weights associated with the data source represent a set of parameter values ​​pre-set to quantify the reliability of different types of data acquisition methods, based on their inherent measurement precision and reliability, before the system performs the comparison.

[0043] Data source refers to the technical means by which the dimensional data was generated. For example, professional surveying tools (such as laser rangefinders) are considered high-precision sources due to their direct working principle and precise calibration. Computer vision estimation, which indirectly infers from images, is considered a relatively low-precision source because its accuracy is affected by factors such as shooting angle, lens distortion, and model recognition errors. Methods for presetting this weight by those skilled in the art include, but are not limited to, the following: assigning a calculable weight value to different sources based on equipment technical specifications, extensive experimental statistical data, and domain consensus. For example, in the system configuration, {"Laser Rangefinder": 0.9, "Computer Vision Estimation": 0.7} can be set, and this weight will serve as an objective basis for decision-making.

[0044] Whether the comparison results conflict means that, through an objective and quantitative judgment process, it is determined whether two dimensional data from different sources are consistent within an acceptable error range. Here, "conflict" does not mean that the two data are simply unequal, but that their difference exceeds the tolerance threshold preset according to business rules, so that they cannot be regarded as reliable measurements of the same physical quantity, thus triggering an abnormal state that requires system intervention.

[0045] Based on the above concept and setup, this step is an automated quality check of multi-source data. Its core is a quantitative comparison algorithm based on weights and thresholds. The system executes steps including but not limited to: 1) Data Acquisition and Weight Matching: The system acquires dimensional data from different sources from the specified component nodes of the preliminary model. This includes the target key dimension estimated by CV (S_cv) and the corresponding measured dimension measured by professional tools (S_measure). Simultaneously, based on the data source, it matches preset accuracy weights (e.g., weight W_m corresponds to S_measure, and weight W_cv corresponds to S_cv). 2) Absolute Difference Calculation: The absolute difference between the two dimensional data is calculated as Δ = |S_measure - S_cv|. 3) Conflict Detection: The absolute difference Δ is compared with a preset tolerance threshold T. This threshold T can be set according to measurement specifications or business requirements (e.g., 10 mm). If Δ > T, a conflict is determined in the comparison results; otherwise, the data is considered consistent. The setting of the "preset tolerance threshold" comprehensively considers measurement specifications, business requirements, and the visual estimation system error measured in S14 above. For example, for a 1-meter-wide window, the threshold can be set to ±30mm (i.e., ±3%). This is a quantitative technical standard that allows for reasonable error while effectively capturing real conflicts. Thus, it transforms the traditional data rationality judgment, which relies on human experience, into an objective, quantitative, and reproducible automated rule check. By introducing the concepts of precision weight and tolerance threshold, the system can automatically and efficiently identify significant anomalies between different data sources, providing clear decision-making trigger signals for subsequent intelligent conflict resolution, thereby greatly improving the efficiency and reliability of data quality inspection.

[0046] S16. If there is a conflict in the comparison results, the corresponding conflict resolution strategy shall be initiated according to the accuracy weight to correct the preliminary model.

[0047] Explained, a conflict resolution strategy refers to a pre-defined automated decision-making rule followed when the system determines that measurement data from different sources contain unacceptable conflicts. Its purpose is to select the most reliable data from the conflicting data and use it to correct the data model, restoring the model's inherent consistency. Methods for setting this strategy by those skilled in the art include, but are not limited to, the following: Based on the objective indicator of the accuracy weight of the data sources, the strategy is formulated as a "weight-first principle," that is, unconditionally adopting the data provided by the data source with the higher accuracy weight among the conflicting data, because a higher weight represents higher prior confidence and measurement accuracy. This is the most direct, objective, and easily implemented automated strategy. For example, those skilled in the art can set various strategies based on business logic, as shown in the following examples: 1) Weight-first strategy: This is the core strategy, directly comparing the accuracy weights of the conflicting data, unconditionally adopting the value provided by the data source with the higher weight, and using it to cover the data with the lower weight. For example, the rule can be defined as: "If laser rangefinder data (weight = 0.9) conflicts with CV estimation data (weight = 0.7), then adopt the laser data and cover the CV data." 2) Marking and review strategy: When the weights of conflicting data are the same or differ by a very small range, the system does not automatically overwrite them. Instead, it packages the conflicting entries, related data and their context (such as associated photos) to generate a "to be reviewed" task, which is then assigned to the designated reviewer or the original field personnel for manual confirmation and processing.

[0048] Based on the above concept and setup, this step is the decision-making and execution stage of the entire intelligent verification process, and its core lies in realizing the self-correction of the model. After determining a conflict, the system automatically runs the following steps, including but not limited to: 1) Strategy matching and activation: Once a conflict is determined, the system matches the conflict resolution strategy applicable to the current conflict scenario according to the preset rule base. In this embodiment of the invention, the "weight priority" conflict resolution strategy is activated by default. The system first identifies the data sources corresponding to the conflicting parties (i.e., S_measure and S_cv). 2) Decision and execution: The system compares the precision weight values ​​of the two data sources (e.g., S_measure and S_cv). Subsequently, it automatically selects the data provided by the data source with the higher precision weight value as the reliable data. 3) Model correction: The system uses this reliable data to update (e.g., overwrite update) the unreliable size data recorded under the corresponding component node in the initial model. At the same time, the system records the version change log of this model correction, including the correction time, the data being corrected, the adopted data and its reasons (i.e., the weight comparison result), thereby ensuring the traceability of the data evolution process.

[0049] Furthermore, conflict resolution strategies can be enhanced to make them more intelligent. This can be achieved through the following steps: 1) Weight-based arbitration: The system compares the accuracy weights of conflicting data (such as laser measurement data and CV estimation data), prioritizing the adoption of higher-weighted data (such as laser data) to correct the model. This is the basic strategy. 2) To improve the system's intelligence, this embodiment also provides enhancement strategies: a) Continuous conflict learning: The system records each conflict event and arbitration result. If the same component or the same CV model experiences multiple conflicts within a short period and is consistently judged as having CV data errors, the system generates a "model performance degradation warning," prompting the administrator to review or retrain the CV model; b) Multi-channel review: If the weights of the conflicting parties are the same or very close, the system does not automatically overwrite the conflict data packet (including photos, measurements, and location information) but automatically pushes it to a second review channel (such as another senior operator or expert group) for final adjudication, recording the adjudication result for optimization of weight allocation or model training.

[0050] Thus, it achieves closed-loop automated processing after the system discovers data quality problems. By executing preset conflict resolution strategies, the system can mimic the decision-making logic of senior experts and intelligently, objectively and efficiently correct erroneous data in the model, thereby ensuring the accuracy and reliability of the preliminary model and even the final result report. It minimizes human intervention throughout the process and achieves truly intelligent data governance.

[0051] S17. Based on the revised real estate data model, automatically generate a summary report of surveying and mapping results.

[0052] Explained, this step is the final output of the entire data processing workflow. Its core function is to automatically transform the intelligently validated and corrected, high-quality, standardized data into compliant business documents. The system can achieve this process through, but is not limited to, a pre-built report generation engine. 1) Data Extraction and Population: The engine traverses every node of the corrected real estate data model (from property to floor, room, and even specific components), extracting the final, validated attribute data according to the structure and order specified in the predefined report template. This includes, but is not limited to, room name, area, component type, key dimensions (such as door and window width and height), and associated media evidence data indexes. 2) Template Rendering and Composition: The engine automatically populates the extracted data into the corresponding positions in the report template. Simultaneously, the system calls the media processor embedded in the template to insert the specified media evidence data (such as images) into the report according to preset sizes and formats. For numerical data (such as area), the engine automatically performs the summary calculations defined in the template (such as usable floor area, building area, etc.). 3) Report Generation and Output: Finally, the engine combines the populated templates into a complete, richly illustrated, standardized electronic document (such as a PDF-formatted "House Survey Report"), storing it in a designated system location or directly pushing it to the relevant business system.

[0053] S18. If there is no conflict in the comparison results, a summary report of surveying and mapping results will be automatically generated based on the preliminary model.

[0054] Explained, this step implements a direct output path when the automated verification process does not detect data conflicts. It shares the same report generation engine as S17, but differs in the data source. Specifically, after performing the comparison judgment in S15, if the system determines that the differences between all key dimensions and measured dimensions are within the preset tolerance threshold, i.e., "there are no conflicts in the comparison results," then the system will assume that the data quality in the current preliminary model meets the requirements and does not need to initiate a conflict resolution process to correct it. It will then trigger the report generation engine, which will directly use this uncorrected, original preliminary model as the sole data source and perform the same data extraction, template filling, and document synthesis operations as in S17 to automatically generate a summary report of surveying and mapping results.

[0055] This invention, through key improvements such as introducing a dynamic surveying inventory to guide structured data acquisition, automatic modeling based on spatial semantics, automated cross-validation of computer vision and measured data, and intelligent conflict resolution based on precision weights, constructs a closed-loop intelligent data processing workflow. Specifically, the dynamic surveying inventory ensures the standardization and completeness of the collected data from the source; automatic modeling transforms discrete data into a computable structured model, laying the foundation for subsequent processing; computer vision recognition and estimation provide an independent and traceable source of verification for the measured data; and finally, the conflict resolution strategy based on precision weights can objectively and automatically correct data inconsistencies, replacing manual judgment. These improvements work synergistically to achieve automated and intelligent inventory and summarization of multi-source surveying data, fundamentally solving the core technical problems of low efficiency, difficulty in consistency verification, and poor controllability of accuracy caused by reliance on manual labor in traditional methods. This significantly improves the overall quality, efficiency, and reliability of surveying data processing.

[0056] In one embodiment, based on a preset accuracy weight associated with the data source, the target key dimension is compared with the corresponding measured dimensions of the corresponding component recorded in the preliminary model, including: Based on the accuracy weight, preset accuracy weight values ​​are obtained for different data sources. The measured dimensions collected by professional surveying tools are assigned a first weight value, and the target key dimensions estimated by computer vision recognition are assigned a second weight value. The first weight value is higher than the second weight value. Calculate the absolute difference between the target critical dimension and the corresponding measured dimension; The absolute difference is compared with a preset tolerance threshold. When the absolute difference is greater than or equal to the preset tolerance threshold, it is determined that there is a conflict in the comparison result.

[0057] Explained as above, precision weights represent a strategy for assigning varying degrees of credibility to different data sources. The precision weight value is a concrete quantitative manifestation of this strategy, a specific numerical parameter assigned to each data source for objective mathematical comparison during conflict resolution. The relationship between the two is that of strategy and execution parameter. For example, the strategy (precision weight) stipulates that "professional tool data takes precedence over CV estimation data," and the execution parameter (precision weight value) of this strategy is specifically quantified as: {"laser rangefinder": 0.9, "computer vision estimation": 0.7}. Here, the first weight value (0.9) corresponds to professional surveying tools, the second weight value (0.7) corresponds to computer vision recognition, and 0.9 > 0.7 reflects the rule that "the first weight value is higher than the second weight value."

[0058] A preset tolerance threshold represents a pre-defined maximum permissible error value used to determine whether two measurement results can be considered consistent. It is an objective, quantifiable standard, and its value is determined based on measurement specifications, business requirements, or engineering experience. For example, in real estate surveying, this threshold can be set to 10 millimeters, meaning that when the difference between two dimensions is less than 10 millimeters, it is considered a measurement error or minor deformation, which is acceptable; when the difference reaches or exceeds 10 millimeters, it is considered a "conflict" that needs to be addressed.

[0059] Based on the above concept and setup, this embodiment of the invention specifies the core steps of "comparison" and "conflict judgment," realizing an automated data quality inspection checkpoint based on threshold judgment. The system executes the following steps: 1) Parameter acquisition: The system acquires the specific precision weight values ​​(i.e., the first weight value and the second weight value) corresponding to the conflicting data according to a preset precision weight strategy. This step prepares a decision basis for subsequent potential conflict resolution. 2) Difference quantification: The system calculates the absolute difference Δ between the target key dimension and the corresponding measured dimension. This operation transforms the difference between the two data into a comparable scalar. 3) Threshold comparison: The system mathematically compares the calculated absolute difference Δ with a preset tolerance threshold T. This is a definite and unambiguous logical judgment: Δ>= T. 4) Conflict judgment: Based on the comparison result, a judgment conclusion is output: If Δ>= T is true, the comparison result is judged to have a conflict; otherwise, it is judged to have no conflict.

[0060] This invention, through its embodiment, concretizes the abstract "comparison" operation into a parameter-driven, threshold-based automated rule engine, achieving objectivity and proceduralization of data consistency verification. Its core improvement lies in constructing a repeatable decision-making logic that requires no manual intervention, using two key parameters: precision weight values ​​and preset tolerance thresholds. Specifically, firstly, by pre-setting quantified precision weight values ​​for different data sources, an objective and unified decision-making basis is provided for the system to intelligently arbitrate when conflicts are detected, eliminating the subjectivity of manual processing. Secondly, by introducing a preset tolerance threshold and comparing differences with it, "conflict" is transformed from a vague concept into a clear, quantifiable state, ensuring the consistency and accuracy of conflict judgment standards. In summary, through the above-mentioned parameterized configuration and automated judgment techniques, the problems of inconsistent data quality inspection standards, low efficiency, and unreliable reliability caused by reliance on subjective experience in traditional methods are effectively solved, providing crucial technical support for the intelligent and highly reliable processing of the entire surveying and mapping results data workflow.

[0061] In one embodiment, a corresponding conflict resolution strategy is initiated based on the accuracy weight to correct the initial model, including: After determining that there is a conflict in the comparison results, identify the data source corresponding to the conflicting data; Compare the precision weight values ​​corresponding to each of the conflicting data; Select data provided by the data source with the highest precision weight value; Based on the selected data, the size information of the corresponding components recorded in the preliminary model is updated to generate a revised real estate data model, and the model version and its change log are recorded.

[0062] Interpretationally, conflicting data refers to two or more sets of data that are determined to have unacceptable differences during the comparison step, namely, the target critical dimensions estimated by computer vision identification and the corresponding measured dimensions collected and recorded in the preliminary model by professional surveying tools.

[0063] The data source indicates the technical approach or equipment type that generated the conflicting data. For example, the data source for the target's critical dimensions is "computer vision recognition and estimation," while the data source for the corresponding measured dimensions is "professional surveying tools (such as laser rangefinders)." The purpose of identifying the data source is to accurately match its corresponding accuracy weight value in the preset rules.

[0064] Based on the above concept and description, this embodiment of the invention specifies the core decision-making and execution steps of "conflict resolution," which implements a rule-based, automated model self-correction mechanism. After determining a conflict, the system executes the following steps: 1) Conflict identification and source tracing: The system first identifies the specific data pairs that caused the conflict and clearly identifies their respective data sources. 2) Weight comparison and decision: The system accesses the preset precision weight configuration, obtains the precision weight values ​​corresponding to the data sources of both conflicting parties, and performs numerical comparison. 3) Trusted data selection: According to the principle of "selecting the highest value," the system automatically selects the data provided by the data source with the highest precision weight value as trusted data, thereby replacing subjective judgment with objective parameters. 4) Model update and log recording: The system uses the trusted data selected in the previous step to overwrite the size information of the corresponding components recorded in the preliminary model. Specifically, the system locates the node where the component is located in the preliminary model and replaces its size attribute value with the value of the trusted data. At the same time, the system will generate a new version of the revised real estate data model and record the revision in detail in a change log, including: the revised components, the original values, the adopted values, the reasons for adoption (i.e., the weight comparison results), and the timestamp, thereby ensuring the complete traceability of the data evolution process.

[0065] This invention, by concretizing "conflict resolution" into an automated decision-making and execution process based on weighted comparison, achieves a leap from "error identification" to "error autonomy" in data processing systems. Specifically, by automatically identifying the source, comparing weights, selecting high-weight data, and updating the model, the system not only instantly completes decisions that would otherwise require repeated manual verification, but also completely eliminates the uncertainty and inefficiency of manual intervention. Furthermore, version and log management ensures the transparency and auditability of each correction, thereby endowing the system with the ability to self-purify and self-evolve. This greatly improves the data quality and reliability of the initial model and even the final results, providing a decisive guarantee for producing highly credible surveying and mapping reports.

[0066] In one embodiment, please refer to Figure 2 , Figure 2 This is a schematic diagram of the first sub-process of the surveying and mapping result data processing method based on intelligent recognition provided in an embodiment of the present invention. Figure 2 As shown, computer vision recognition is performed on media evidence data associated with the preliminary model to identify the calibrated reference objects and the property components they depict, and based on the identified calibrated reference objects, target key dimensions of the property components are estimated, including: S21. Using a pre-trained object detection model, identify the image regions of the calibrated reference object and the target property component from the media evidence data; S22. Calculate the pixel dimensions of the calibration reference object and the target real estate component in the image respectively; S23. Based on the known physical dimensions and pixel dimensions of the calibration reference object, calculate the scale conversion relationship of the current image region; S24. Based on the scale conversion relationship and the pixel size of the target property component, calculate the physical size of the target property component as the estimated target key size.

[0067] Explained, a pre-trained object detection model refers to a computer vision model (such as YOLO, Faster R-CNN, etc.) that is based on deep learning, trained using a large amount of labeled image data, and capable of automatically identifying the location and category of specific objects in an image. Those skilled in the art can construct such a model as follows: First, collect and label an image dataset containing various "calibration references" (such as surveying boards, mobile phones) and "real estate components" (such as doors, windows, beams, and columns), where the annotation information consists of the bounding boxes of objects in the images and their category labels; then, select a general object detection network architecture, and train it using the dataset through backpropagation to optimize the network parameters, enabling it to learn to extract features from pixels and accurately locate and classify target objects; finally, solidify and save the trained model parameters for inference.

[0068] An image region represents a local set of pixels identified by an object detection model in media evidence data that corresponds to a specific target (calibrated reference object or property component). It is usually defined by a rectangular bounding box, and this region is the object of subsequent pixel calculations.

[0069] Pixel size represents the number of pixels (width or height) occupied by a target object in its digital image; it is a unitless number based on image resolution. Physical size, on the other hand, represents the actual length, width, etc., of the target object in the real world, measured in metric units (such as millimeters). Within a single image, pixel size and physical size are related through a scale conversion relationship; only after determining this conversion relationship can the pixel size be converted into a physically meaningful size.

[0070] The scale conversion relationship represents the conversion factor used to convert the pixel size in an image to the physical size in the real world. It represents "the actual physical length represented by one pixel unit in the current image". This relationship is calculated by dividing the known physical size of the reference object by its pixel size in the image, that is: Scale conversion relationship (mm / pixel) = Known physical size of reference object (mm) / Pixel size of reference object (pixel).

[0071] Based on the above concept and setup, the system executes the following steps: 1) Target recognition and region extraction: The system calls the pre-trained object detection model to infer the associated media evidence data (such as images). The model outputs the bounding box coordinates of the calibration reference object and the target property component, thereby accurately defining the image regions of both. 2) Pixel size calculation: Based on the identified bounding box coordinates, the system calculates the pixel size (e.g., the number of pixels in the width direction) of the calibration reference object and the target property component in the image. 3) Scale conversion calculation: Based on the known physical size of the calibration reference object (e.g., the width of a mobile phone is 76.7 mm) and its pixel size in the image, the system calculates the scale conversion relationship of the current image region using the above formula. 4) Physical size estimation: The system multiplies the pixel size of the target property component by the calculated scale conversion relationship to obtain the physical size of the component. This value is recorded as the key target size for comparison.

[0072] This invention, through the construction of a computable pipeline from pixels to the physical world, transforms ordinary field photographs into quantifiable verification data sources, thus creatively applying computer vision technology to the field of surveying and mapping quality inspection. This provides an independent and traceable cross-verification method for traditional single-point measurements. Specifically, by automatically identifying reference objects and components in images and estimating dimensions based on rigorous geometric proportions, the system can efficiently and in batches generate verification dimensions parallel to the tool measurement data. This not only greatly expands the coverage and automation level of data quality inspection, but more importantly, it establishes a quantitative logical bridge between "image evidence" and "numerical records" within the data processing flow for the first time. This provides a solid technical foundation for discovering hidden data contradictions, thereby upgrading the quality control of the output data from a "probabilistic guarantee" relying on sampling to a "deterministic guarantee" that can be comprehensively and intelligently verified.

[0073] In one embodiment, identifying the image regions of the calibrated reference object and the target property component from the media evidence data using a pre-trained object detection model includes: The object detection model is configured to recognize multiple types of calibration references, including at least one of the following: a mapping calibration board with standard dimensions, the mobile terminal itself, and everyday objects of known dimensions. The object detection model is configured to recognize multiple types of property components, including at least one of the following: walls, floors, ceilings, doors and windows, beams and columns, pipes, and switches and sockets. The identified image region is subjected to contour optimization processing to accurately define the pixel boundary between the calibration reference and the target property component.

[0074] Explained, contour optimization refers to an image post-processing algorithm performed after an object detection model has initially identified the approximate bounding box of a target, in order to further improve the localization accuracy of its pixel boundaries. Its purpose is to eliminate problems such as localization deviations, jagged edges, or the inclusion of redundant background pixels in the initial bounding box, thereby obtaining a precise pixel contour that highly matches the true shape of the target object. Those skilled in the art can implement this processing using existing mature computer vision algorithms, such as: 1) instance segmentation algorithms (e.g., Mask R-CNN), which can directly output a pixel-level precise mask of the target while detecting objects; 2) optimization based on an active contour model, which defines an energy function to drive the initial contour curve to evolve towards the true boundary of the target, ultimately fitting precisely to the edge.

[0075] Everyday items of known size, specifically those not designed or carried for surveying purposes, but whose shape is regular, size is standardized, common in daily life and easily accessible, so that their precise physical dimensions can be pre-recorded and identified by the system, including but not limited to aluminum cans, common mineral water / beverage bottles, smartphones, A4 paper, credit cards / debit cards.

[0076] Based on the above concept and description, the embodiments of the present invention have made key refinements and enhancements to the core step of "identifying image regions". By expanding the model recognition capabilities and introducing post-processing refinement, the accuracy of the size estimation source data is ensured.

[0077] The pre-trained object detection model is specifically trained to recognize a broad library of object categories, specifically including but not limited to the following: 1) In the calibration reference category, the model can recognize standard-sized mapping calibration boards, the mobile terminal itself (such as a specific model of smartphone or tablet), and everyday objects of known size (such as standard-sized beverage cans or credit cards). This diversity ensures that field personnel can still complete calibration using readily available items when specialized tools are lacking. 2) In the property component category, the model can recognize walls, floors, ceilings, doors and windows, beams and columns, pipes, switches and sockets, etc., ensuring that the system can cover the main measurement objects in in-home surveying. In another example, the recognition of calibration references and target property components can also be performed using corresponding pre-trained object detection models, for example, calibration references can be recognized using corresponding pre-trained object detection model A, and target property components can be recognized using corresponding pre-trained object detection model B. This embodiment of the invention is not limited to this.

[0078] Contour optimization precisely defines the boundaries. For each target's rectangular image region initially identified by the model, the system further calls the contour optimization processing algorithm to calculate the target's precise pixel-level contour. This directly determines the accuracy of subsequent pixel size calculations and avoids measurement errors caused by the rectangle containing the background or failing to fit the edge.

[0079] Furthermore, in the implementation of using a mobile terminal as a calibration reference, when the calibration reference is the mobile terminal itself, the specific operation of capturing it into the media evidence data can include, but is not limited to, the following example: Field personnel use another mobile terminal (such as a colleague's mobile phone or tablet) as an auxiliary device to capture the target component, while placing their own mobile terminal (the device running the acquisition APP) next to the component as a calibration reference and capturing it together in the picture. The acquisition APP can automatically associate the device model and known size information of the mobile terminal used as the reference with the captured media evidence data through Bluetooth or network communication.

[0080] This invention provides robust and high-quality data input assurance for the entire computer vision measurement process by expanding the versatility of the model and enhancing the accuracy of recognition. Its core improvements are: first, by supporting multiple types of calibration references and real estate components, it greatly enhances the system's adaptability and practicality in complex on-site environments, ensuring that measurement tasks can be performed under various conditions; second, by introducing contour optimization processing, the definition of image regions is improved from a coarse "box selection" to a precise "outline," fundamentally improving the accuracy of pixel size calculation and significantly reducing subsequent size estimation errors; in particular, designing the mobile terminal itself as a calibration reference is a highly ingenious system design that cleverly solves the industry pain point of field personnel potentially forgetting to carry dedicated calibration objects, achieving a seamless "tool as ruler" experience, further improving the reliability and efficiency of the method. These improvements work together to make the vision-based automated verification process more robust and reliable.

[0081] In one embodiment, please refer to Figure 3 , Figure 3 This is a schematic diagram of the second sub-process of the surveying and mapping result data processing method based on intelligent recognition provided in an embodiment of the present invention. For example... Figure 3 As shown, the mobile terminal is used to attach corresponding spatial location tags and target component attributes to the target data, including: S31. Parse the dynamic mapping list and obtain at least one list item therein; S32. In the display interface of the mobile terminal, a structured data entry form corresponding to the list item is presented; S33. In response to the user operation that triggers the collection of the media evidence data, the positioning module and inertial measurement unit of the mobile terminal are invoked to obtain the current spatial position and attitude data, and the spatial position tag is generated. S34. Provide a component attribute selection interface for the user to select or input the target component attribute from a predefined attribute list. The target component attribute includes at least one of component type, room to which it belongs, and material information. S35. Associate and bind the target data, the spatial location label, and the target component attribute to attach corresponding spatial location labels and target component attributes to the target data.

[0082] Explained, a list item represents the basic unit that constitutes the dynamic mapping list; it is a specific instruction within the list for a particular data collection target. For example, a list item could be "collect the dimensions, material, and photograph of the window on the east wall of the master bedroom."

[0083] Structured data entry forms are standardized data collection interfaces that are dynamically generated on mobile terminals based on the content of list items and contain specific input fields (such as numeric input boxes, drop-down selection menus, and file upload controls). They are used to guide users to complete data entry in a standardized and comprehensive manner.

[0084] A positioning module (such as a GPS or BeiDou chip) is used to acquire the device's absolute or relative spatial location data, such as latitude and longitude coordinates, or a rough indoor location determined via Bluetooth / Wi-Fi signals. An inertial measurement unit (IMU), containing accelerometers and gyroscopes, is used to measure the device's attitude data, i.e., the device's orientation, tilt angle, etc., in space.

[0085] Spatial location and attitude data together constitute the data source for spatial location labels. Spatial location describes "where" while attitude data helps determine "which direction". When combined, they can generate a more accurate semantic location description, such as "master bedroom, east wall, 1.2 meters above the ground".

[0086] Component type and material information are core components of the target component's attributes. Component type indicates the functional classification of the component, such as "window", "door", "wall", and "pipe". Material information indicates the material that makes up the component. For example, for a window, the material may be "aluminum alloy", "PVC", or "wood".

[0087] Based on the above concept and description, this embodiment of the invention specifies the entire process of front-end data acquisition and structuring. Through task guidance, sensor invocation, and the collaboration of the interactive interface, it achieves synchronous acquisition and real-time correlation of multi-source data. The system executes the following steps on the mobile terminal, including but not limited to: 1) Task parsing: The application parses the dynamic mapping list, obtains at least one list item, and clarifies the specific acquisition task to be performed. 2) Interface guidance: Based on the list item content, a structured data entry form corresponding to the list item is presented on the screen. This form pre-sets the data fields to be collected. 3) Spatial tag generation: When the user performs a user operation that triggers the acquisition of media evidence data (such as clicking the photo button), the system synchronously invokes the positioning module and inertial measurement unit of the mobile terminal to obtain the current spatial position and attitude data, and generates the spatial position tag accordingly. 4) Attribute entry: The system provides a component attribute selection interface. This interface contains a predefined attribute list for the user to select or input the target component attribute. 5) Data binding: Finally, the system logically associates and binds the three types of elements—the collected target data (measured dimensions, media evidence), the generated spatial location tags, and the entered target component attributes—to form a structured data package, thus completing the operation of attaching corresponding spatial location tags and target component attributes to the target data.

[0088] This invention addresses the fundamental problems of messy and error-prone outdoor surveying data by constructing a highly structured, sensor-enhanced front-end acquisition process. Its core improvement lies in merging the previously separate actions of data acquisition, spatial positioning, and attribute description into a single atomic operation, achieving "structured acquisition." Specifically, through list parsing and form presentation, standardization and unambiguous guidance of acquisition tasks are achieved; by automatically invoking sensors to generate spatial location tags, the real-time nature and accuracy of spatial information are ensured, avoiding errors from subsequent manual association; and by selecting and inputting from a predefined attribute list, the standardization and consistency of attribute data are guaranteed. Finally, through association and binding, a data package with unified internal logic and rich semantic information is generated. This provides high-quality, cleansing-free data input for automated modeling and intelligent verification in the back-end system, forming the cornerstone for the efficient and reliable operation of the entire technical solution.

[0089] In one embodiment, based on the spatial location label and the target component attribute, the target data is automatically associated with the corresponding node in a pre-set real estate data model, including: The spatial location tag is parsed to determine the target property identifier it points to, the target property identifier including the floor number and the room number; Parse the target component's attributes to determine its described component type and component identifier; In the preset property data model, the target room node is located based on the floor number and the room number; Under the target room node, locate or create the corresponding component child node according to the component type and the component identifier; The target data is attached as attribute data to the component child node to complete the construction of the preliminary model.

[0090] Explained, a target property identifier represents coded information used to uniquely identify a spatial location within a pre-defined property data model. It is typically extracted from spatial location labels, including but not limited to floor numbers (e.g., "F1", "2nd floor") and room numbers (e.g., "101", "Master Bedroom"). The floor number and room number together constitute a hierarchical location from macro to micro levels. For example, after parsing the spatial location label "Master Bedroom - East Wall," the target property identifier would be {Floor Number: "1", Room Number: "Master Bedroom"}.

[0091] Based on the above concept and description, this embodiment of the invention specifies the core steps of automatic data association and modeling, implementing an intelligent data location algorithm based on semantic parsing. The system executes, but is not limited to, the following steps: 1) Spatial semantic parsing: The system parses the spatial location label and determines the target property identifier (i.e., floor and room information) it points to from its content. 2) Component semantic parsing: The system parses the target component attributes and determines the component type (e.g., "window") and component identifier (e.g., "east wall window," which can be obtained from the attributes or generated by combining the location label and type). 3) Room node positioning: In the pre-set property data model, the system performs step-by-step matching based on the floor number and room number to locate the unique target room node. 4) Component node management: Under the target room node, the system queries based on the component type and component identifier. If the node already exists, it is located directly; if it does not exist, a corresponding component child node is created according to this type and identifier. 5) Data mounting and model construction: Finally, the system mounts the target data (measured dimensions and media evidence) as attribute data to the component child node. Once all data packets have completed this operation, the initial model construction is complete.

[0092] This invention, through the construction of an automatic node location and creation mechanism based on dual semantic parsing (space + component), achieves precise and lossless conversion from unstructured data packets to structured data models. Its core improvement lies in transforming the traditional CAD drawing or data table association work, which relies on manual understanding and operation, into a programmable and reproducible automated algorithm. Specifically, by parsing spatial location tags and component attributes, the system can accurately understand the semantic attribution of each data packet; through intelligent node location or creation, it ensures a strict correspondence between the model structure and the real property; and finally, through data mounting, it forms a complete digital asset rich in semantic relationships. This automated process completely eliminates the subjectivity and operational errors of manual association, not only improving modeling efficiency by several orders of magnitude but also providing a unique, accurate, and structured data foundation for subsequent intelligent verification, making it the key hub for achieving full-process automation in the entire technical solution.

[0093] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0094] In one embodiment, a surveying and mapping result data processing device based on intelligent recognition is provided, which corresponds one-to-one with the surveying and mapping result data processing method based on intelligent recognition described in the above embodiments. Please refer to [link / reference]. Figure 4 , Figure 4This is a schematic block diagram of a surveying and mapping result data processing device based on intelligent recognition, provided in an embodiment of the present invention. Figure 4 As shown, the intelligent recognition-based surveying and mapping data processing device 40 includes a first distribution module 41, a first acquisition module 42, a first construction module 43, a first estimation module 44, a first comparison module 45, a first correction module 46, and a first summary module 47. The detailed descriptions of each of these functional modules are as follows: The first distribution module 41 is used to generate and distribute a dynamic survey list of the target property to the mobile terminal; the first acquisition module 42 is used to acquire target data through the mobile terminal based on the dynamic survey list, and to attach corresponding spatial location tags and target component attributes to the target data through the mobile terminal, wherein the target data includes measured dimensions and media evidence data containing calibration references; the first construction module 43 is used to automatically associate the target data with corresponding nodes in a preset property data model based on the spatial location tags and target component attributes to construct a preliminary model; the first estimation module 44 is used to estimate the media evidence associated with the preliminary model. The data undergoes computer vision recognition to identify the calibration reference and the property component to be measured, and estimates the target key dimensions of the property component based on the identified calibration reference. A first comparison module 45 is used to compare the target key dimensions with the corresponding measured dimensions of the corresponding component recorded in the preliminary model based on a preset accuracy weight associated with the data source, and to determine whether there is a conflict in the comparison results. A first correction module 46 is used to initiate a corresponding conflict resolution strategy to correct the preliminary model based on the accuracy weight if the above determination is correct. A first summary module 47 is used to automatically generate a surveying results summary report based on the corrected property data model.

[0095] In one embodiment, the first comparison module 45 includes: a first acquisition submodule, configured to acquire, according to the accuracy weight, preset accuracy weight values ​​for different data sources, wherein the measured dimensions acquired by professional surveying tools are assigned a first weight value, the target key dimensions estimated by computer vision recognition are assigned a second weight value, and the first weight value is higher than the second weight value; a first calculation submodule, configured to calculate the absolute difference between the target key dimensions and the corresponding measured dimensions; a first comparison submodule, configured to compare the absolute difference with a preset tolerance threshold; and a first determination submodule, configured to determine that the comparison result conflicts when the absolute difference is greater than or equal to the preset tolerance threshold.

[0096] In one embodiment, the first correction module 46 includes: a first identification submodule, used to identify the data source corresponding to the conflicting data after determining that there is a conflict in the comparison results; a second comparison submodule, used to compare the precision weight values ​​corresponding to each of the conflicting data; a first selection submodule, used to select the data provided by the data source with the higher precision weight value; and a first update submodule, used to update the size information of the corresponding components recorded in the preliminary model according to the selected data, generate a corrected real estate data model, and record the model version and its change log.

[0097] In one embodiment, the first estimation module 44 includes: a second identification submodule, used to identify the image regions of the calibration reference object and the target property component from the media evidence data using a pre-trained object detection model; a second calculation submodule, used to calculate the pixel dimensions of the calibration reference object and the target property component in the image respectively; a third calculation submodule, used to calculate the scale conversion relationship of the current image region based on the known physical dimensions and pixel dimensions of the calibration reference object; and a fourth calculation submodule, used to calculate the physical dimensions of the target property component according to the scale conversion relationship and the pixel dimensions of the target property component, as the estimated target key dimensions.

[0098] In one embodiment, the second identification submodule includes: a first configuration submodule, configured to enable the object detection model to identify multiple types of calibration reference objects, the calibration reference objects including at least one of: a surveying calibration board with standard dimensions, the mobile terminal itself, and everyday objects of known dimensions; a second configuration submodule, configured to enable the object detection model to identify multiple types of real estate components, the real estate components including at least one of: walls, floors, ceilings, doors and windows, beams and columns, pipes, and switches and sockets; and a first delimitation submodule, configured to perform contour optimization processing on the identified image region to accurately delimit the pixel boundaries between the calibration reference object and the target real estate component.

[0099] In one embodiment, the first acquisition module 42 includes: a first parsing submodule, used to parse the dynamic mapping list and obtain at least one list item therein; a first presentation submodule, used to present a structured data entry form corresponding to the list item in the display interface of the mobile terminal; a first generation submodule, used to respond to a user operation that triggers the acquisition of media evidence data, call the positioning module and inertial measurement unit of the mobile terminal to obtain the current spatial position and attitude data, and generate the spatial position label; a first providing submodule, used to provide a component attribute selection interface for the user to select or input the target component attribute from a predefined attribute list, the target component attribute including at least one of component type, room to which it belongs and material information; and a first association submodule, used to associate and bind the target data, the spatial position label and the target component attribute to attach corresponding spatial position labels and target component attributes to the target data.

[0100] In one embodiment, the first construction module 43 includes: a second parsing submodule, used to parse the spatial location label and determine the target property identifier it points to, the target property identifier including a floor number and a room number; a third parsing submodule, used to parse the target component attributes and determine the component type and component identifier it describes; a first positioning submodule, used to locate the target room node in the preset property data model according to the floor number and the room number; a first determining submodule, used to locate or create a corresponding component subnode under the target room node according to the component type and the component identifier; and a first mounting submodule, used to mount the target data as attribute data to the component subnode to complete the construction of the preliminary model.

[0101] Specific limitations regarding the intelligent recognition-based surveying and mapping data processing device can be found in the above-mentioned limitations on the intelligent recognition-based surveying and mapping data processing method, and will not be repeated here. Each module in the aforementioned intelligent recognition-based surveying and mapping data processing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0102] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used for communication with external clients via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a server-side method for processing surveying and mapping data based on intelligent recognition.

[0103] In one embodiment, a computer device is provided, which may be a client, and its internal structure diagram may be as follows: Figure 6 As shown, the computer device includes a processor, memory, network 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 in the non-volatile storage media. The network interface is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it implements client-side functions or steps of a surveying and mapping data processing method based on intelligent recognition.

[0104] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the intelligent recognition-based surveying and mapping result data processing method described in the above embodiments.

[0105] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the intelligent recognition-based surveying and mapping result data processing method described in the above embodiments.

[0106] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0107] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0108] The software tools or components not belonging to our company that appear in the embodiments of this invention are merely illustrative examples and do not represent actual use.

[0109] The data collection in this embodiment of the invention complies with the requirements of relevant laws and regulations, such as China's Personal Information Protection Law, GDPR (General Data Protection Regulation of the European Union), or information security standards of other countries and regions.

[0110] The above-described 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 the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for processing surveying and mapping result data based on intelligent identification, characterized in that, The method comprises: generating and issuing a dynamic survey list for a target property to a mobile terminal; collecting target data through the mobile terminal based on the dynamic survey list, and attaching corresponding spatial position tags and target component attributes to the target data through the mobile terminal, wherein the target data comprises measured dimensions and media evidence data containing calibration reference objects; automatically associating the target data to corresponding nodes in a preset property data model based on the spatial position tags and the target component attributes to construct a preliminary model; performing computer vision recognition on the media evidence data associated with the preliminary model to identify the calibration reference objects and the property components to be measured therefrom, and estimating target key dimensions of the property components based on the identified calibration reference objects; comparing the target key dimensions with corresponding measured dimensions of the corresponding components recorded in the preliminary model based on preset precision weights associated with the data sources, and determining whether there is a conflict in the comparison result; if the determination is yes, starting a corresponding conflict resolution strategy according to the precision weights to correct the preliminary model; generating a survey result summary report automatically based on the corrected property data model. 2.The smart recognition-based surveying and mapping result data processing method of claim 1, wherein, comparing the target key dimensions with corresponding measured dimensions of the corresponding components recorded in the preliminary model based on preset precision weights associated with the data sources, and determining whether there is a conflict in the comparison result, comprising: obtaining preset precision weight values for different data sources according to the precision weights, wherein the measured dimensions collected by professional survey tools are assigned a first weight value, the target key dimensions estimated by computer vision recognition are assigned a second weight value, and the first weight value is higher than the second weight value; calculating the absolute difference between the target key dimensions and the corresponding measured dimensions; comparing the absolute difference with a preset tolerance threshold; when the absolute difference is greater than or equal to the preset tolerance threshold, determining that there is a conflict in the comparison result. 3.The smart recognition-based surveying and mapping result data processing method of claim 2, wherein, starting a corresponding conflict resolution strategy according to the precision weights to correct the preliminary model, comprising: after determining that there is a conflict in the comparison result, identifying the data sources corresponding to the conflict data; comparing the precision weight values corresponding to the conflict data respectively; selecting data provided by the data source with a high precision weight value; updating the dimension information of the corresponding components recorded in the preliminary model according to the selected data, generating a corrected property data model, and recording the model version and its change log. 4.The smart recognition-based surveying and mapping result data processing method of claim 1, wherein, performing computer vision recognition on the media evidence data associated with the preliminary model to identify the calibration reference objects and the property components depicted therefrom, and estimating target key dimensions of the property components based on the identified calibration reference objects, comprising: identifying image regions of the calibration reference objects and target property components from the media evidence data through a pre-trained object detection model; respectively calculating the pixel dimensions of the calibration reference objects and the target property components in the image; calculating a scale conversion relationship of the current image region based on the known physical size of the calibration reference object and its pixel size; calculating the physical size of the target house property component based on the scale conversion relationship and the pixel size of the target house property component, as the estimated target key size. 5.The smart recognition-based surveying and mapping result data processing method of claim 4, wherein, identifying image regions of the calibration reference object and the target house property component from the media evidence data through a pre-trained object detection model, including: the object detection model is configured to identify multiple types of calibration reference objects, including at least one of a surveying calibration board with a standard size, the mobile terminal itself, and a daily item with a known size; the object detection model is configured to identify multiple types of house property components, including at least one of a wall, a floor, a ceiling, a door and window, a beam and column, a pipeline, and a switch and socket; performing contour optimization processing on the identified image regions to accurately define the pixel boundaries of the calibration reference object and the target house property component. 6.The smart recognition-based surveying and mapping result data processing method of claim 1, wherein, attaching corresponding spatial location labels and target component attributes to the target data through the mobile terminal, including: parsing the dynamic surveying list to obtain at least one list item; presenting a structured data entry form corresponding to the list item in the display interface of the mobile terminal; in response to a user operation triggering the media evidence data collection, calling the positioning module and inertial measurement unit of the mobile terminal to obtain the current spatial location and attitude data, and generating the spatial location label; providing a component attribute selection interface for the user to select or input the target component attribute from a pre-defined attribute list, the target component attribute including at least one of component type, belonging room, and material information; associating and binding the target data, the spatial location label, and the target component attribute to attach corresponding spatial location labels and target component attributes to the target data. 7.The smart recognition-based surveying and mapping result data processing method of claim 1, wherein, based on the spatial location label and the target component attribute, automatically associating the target data to the corresponding node in the pre-set house property data model, including: parsing the spatial location label to determine the target house property identifier it points to, the target house property identifier including floor number and room number; parsing the target component attribute to determine the component type and component identifier it describes; locating the target room node in the pre-set house property data model according to the floor number and the room number; locating or creating the corresponding component sub-node under the target room node according to the component type and the component identifier; mounting the target data as attribute data under the component sub-node to complete the construction of the preliminary model.

8. An intelligent recognition-based surveying and mapping result data processing device, characterized in that, including: a first issuing module for generating and issuing a dynamic surveying list for a target house property to a mobile terminal; A first collection module is configured to collect target data based on the dynamic survey list via the mobile terminal and attach corresponding spatial position tags and target component attributes to the target data via the mobile terminal, wherein the target data includes measured dimensions and media evidence data containing calibration reference objects; A first construction module is configured to automatically associate the target data to corresponding nodes in a preset housing data model based on the spatial position tags and the target component attributes to construct a preliminary model; A first estimation module is configured to perform computer vision recognition on the media evidence data associated to the preliminary model to identify the calibration reference objects and housing components to be measured therefrom and estimate target key dimensions of the housing components based on the identified calibration reference objects; A first comparison module is configured to compare the target key dimensions with corresponding measured dimensions of components recorded in the preliminary model based on preset precision weights associated with data sources and determine whether there is a conflict in the comparison result; A first correction module is configured to initiate a corresponding conflict resolution strategy to correct the preliminary model according to the precision weights if the determination is positive. A first summary module is configured to automatically generate a survey result summary report based on the corrected housing data model.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The computer program is executed by the processor to implement the steps of the intelligent recognition-based survey result data processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the steps of the intelligent recognition-based survey result data processing method according to any one of claims 1 to 7.

Images