Three-dimensional reconstruction method based on digital image acquisition in green reconstruction of existing building

Abstract

The application discloses a kind of three-dimensional reconstruction methods based on digital image acquisition in green reconstruction of existing building, belong to building informatization technical field, including: for the different types of building to be surveyed, the technical framework of field survey is established, the data of three aspects of geometry, color, disease is measured;For the applicability of different algorithms in different types of existing building, the three-dimensional reconstruction technology of existing building based on image is established;With BIM combined with GIS, the survey results are expressed and managed in three levels of building components, geographical environment and time evolution;With carbon emissions, thermal comfort, cost-effectiveness as the target, the green reconstruction technology integration coupled with wind, light and thermal environment is established.Through the application, more accurate data basis and convenience can be provided for green reconstruction of existing building in data collection and management.

Description

Technical Field

[0001] This invention belongs to the field of building information technology, and more specifically, relates to a three-dimensional reconstruction method based on digital image acquisition in the green renovation of existing buildings. Background Technology

[0002] Surveying is fundamental to the green renovation of existing buildings. The applicability of surveying techniques is particularly important when dealing with different types of existing buildings. In recent years, the application of laser scanning technology has brought new technical approaches to the surveying of existing buildings. However, the use of this technology is constrained by the existing buildings and their surrounding environment; for example, laser surveying is very difficult for mountainous buildings and high-rise buildings.

[0003] Image-based 3D reconstruction expands the measurement scope (from individual buildings to their surrounding environment) and increases measurement dimensions (from geometry to color and structural defects), which is crucial for the renovation and post-construction assessment of existing buildings. However, commonly used 2D drawing methods cannot effectively represent and manage these measurement results. Furthermore, most research on image-based 3D reconstruction technology in my country comes from fields such as civil engineering, surveying, and computer science. While these fields may also focus on architecture, their research is not geared towards the information-based approach to green renovation of existing buildings.

[0004] For example, how to ensure the accuracy of this technology meets the needs of surveying existing buildings of different types and sites, how to effectively express and manage the results of 3D reconstruction, and how to establish corresponding technical frameworks based on the characteristics of different types of existing buildings. Summary of the Invention

[0005] In response to the above-mentioned deficiencies or improvement needs of existing technologies, this invention proposes a three-dimensional reconstruction method based on digital image acquisition and led by architecture. This method can provide a more accurate data foundation and convenience for the green renovation of existing buildings in terms of data acquisition and management.

[0006] To achieve the above objectives, this invention provides a three-dimensional reconstruction method based on digital image acquisition for green renovation of existing buildings, comprising:

[0007] S1: Establish a technical framework for field measurement for different types of buildings to be measured, and measure data in three aspects: geometry, color, and defects.

[0008] S2: To assess the applicability of different algorithms in different types of existing buildings, establish an image-based 3D reconstruction technology for existing buildings;

[0009] S3: Use BIM combined with GIS to express and manage measurement results at three levels: building components, geographical environment, and time evolution;

[0010] S4: Establish an integrated green transformation technology that couples wind, solar, and thermal environments, with the goals of reducing carbon emissions, improving thermal comfort, and achieving cost-effectiveness.

[0011] In some alternative implementations, step S1 includes:

[0012] S11: Acquire color images of the building to be surveyed through a camera sensor and thermal infrared images of the building to be surveyed through a thermal infrared imager, so as to identify hidden defects inside the wall through abnormal temperature distribution.

[0013] S12: Determine the shooting distance, image overlap rate, and number of images to be acquired for the building to be surveyed;

[0014] S13: The color image is corrected by using the multi-scale Retinex algorithm with color restoration to obtain the corrected color image.

[0015] In some alternative implementations, step S12 includes: by Determine the shooting distance, where s x p represents the camera sensor size. x f represents pixels, and f represents focal length.

[0016] Depend on Determine the overlap area of ​​adjacent images by Determine the number of images N to be acquired, where B represents the tolerance, R represents the radius of the circumscribed circle of the building surface, and L is the visible length of the building in the world coordinate system.

[0017] More preferably, step S1 can collect image information at different times to observe the building's change trend and obtain the physical performance of the building under test more intuitively.

[0018] In some alternative implementations, step S2 includes:

[0019] The system generates feature recognition and matching for color images, then generates sparse point clouds through measured ground control points, then generates dense point clouds from sparse point clouds, and finally generates grid surfaces from dense point clouds. Combined with thermal infrared images, a texture mapping is formed to obtain the final GIS model. Based on the texture-mapped point cloud data, a BIM model is created. A visual programming platform is used to combine functional nodes to automatically generate a BIM entity model by constructing the outline of three-dimensional point cloud slices.

[0020] In some alternative implementations, step S2 includes:

[0021] S21: Use Bentley ContextCapture to import positioning data, ensure the interoperability of coordinate system data with GIS solutions, use Descartes to process color images, convert raster images into vector engineering drawings, and associate color images with vector information;

[0022] S22: Utilize Bentley Pointools to enhance, segment, and classify dense point clouds, and combine them with engineering models to detect conflicts in dense point clouds. Perform conflict detection between real-world data and proposed designs, and generate GIS models that can be updated to the latest version in real time by synchronizing with the original data source.

[0023] S23: Create a BIM model based on textured point cloud data, use a visual programming platform to combine functional nodes, and automatically generate a BIM building model by constructing dense point cloud slice outlines.

[0024] In some alternative implementations, step S22 includes:

[0025] S221: Image changes cause point cloud changes. Compare dense point clouds in the same region and identify any increases or decreases. Use the difference tool to detect changes and monitor them in real time. Through monitoring, the dense point cloud is synchronized with the original image data source, and the generated model can be updated to the latest version in real time.

[0026] S222: Uses point layer technology to edit large datasets of dense point clouds, manipulating, cleaning, or subdividing point cloud models. It divides point clouds into various layers, streamlining and reducing noise, and enabling isolated point cloud editing and attribute differentiation. Isolated editing allows for individual editing of changed point clouds detected in S221, shortening modeling time. Attribute differentiation yields unique, build-level point clouds, providing a data foundation for subsequent BIM modeling.

[0027] S223: Manipulates real-world meshes and scalable models with hundreds of millions of triangular tiles, imports, refines, and exports meshes in numerous formats, generates accurate georeferenced 3D models in various GIS formats, utilizes PCL for texture mapping, and obtains the final GIS environment model.

[0028] In some alternative implementations, step S23 includes:

[0029] S231: Calculate the bounding box surface area S of the dense point cloud, and then calculate the average surface density α of the dense point cloud based on the number N of the dense point cloud.

[0030] S232: Point cloud slicing. To ensure the accuracy of the model shape, the point cloud density α is set as the slicing threshold, and x is input as the slicing quantity parameter to calculate the point set on the slicing plane.

[0031] S233: The set of points on the tangent plane, project two adjacent sets of points onto the tangent plane to find their intersection, fit the discrete points, and thus obtain x tangent plane boundary contours;

[0032] S234: Loft and merge adjacent contour boundaries to form a segmented solid model;

[0033] S235: By combining segmented solid models and processing data manually, corresponding material, structural, and defect information is assigned to building components to obtain a BIM building model.

[0034] In some alternative implementations, step S3 includes:

[0035] S31: Utilize the convenient BIM import mechanism provided by SuperMap GIS software to achieve interaction between the GIS environmental model and the BIM building model, and obtain a high-precision three-dimensional model with micro-architectural semantic information and macro-geographical environmental information.

[0036] S32: Establish external references for the building to be surveyed, unify the high-precision 3D models at different times into a coordinate system, and obtain the change trend of the existing building within the same time period.

[0037] By observing the changing trends, we can intuitively obtain hidden building defects that are invisible to the naked eye and existing surveying methods, providing a basis for subsequent renovations. At the same time, we can simulate the lifespan and changes of a building without renovations based on the changing patterns, and then compare whether or not to carry out renovations.

[0038] S33: The CFD-DEM coupled simulation method is used for analysis. Based on the traditional finite element simulation analysis, the influence of terrain factors is comprehensively considered. Multiple technology platforms are linked to simulate and analyze the wind, light and heat of the obtained high-precision three-dimensional model to obtain the wind speed and temperature cloud map of the building to be surveyed.

[0039] S34: Import the high-precision 3D model into the energy consumption calculation software to calculate carbon emissions and greenhouse gas emissions.

[0040] In some alternative implementations, step S4 includes:

[0041] S41: Based on the BIM building model, carry out a renovation, and import the renovation results into finite element simulation software to simulate the wind, solar and thermal environment and obtain wind speed and temperature cloud maps of the renovated building.

[0042] S42: Based on the BIM building model, carry out a renovation, and import the renovation results into energy consumption calculation software to obtain carbon emissions and greenhouse gas emissions;

[0043] S43: Compare the wind speed and temperature cloud map of the modified building with the wind speed and temperature cloud map of the building to be measured, and continue to modify the building envelope parameters based on the comparison results;

[0044] S44: Compare the carbon emissions and greenhouse gas emissions obtained in step S42 with the carbon emissions and greenhouse gas emissions of the building to be measured in step S34. Calculate the incremental cost based on the comparison results, and calculate the cost and benefits of green retrofitting.

[0045] S45: Repeat the above steps until the optimal green renovation plan is obtained.

[0046] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects:

[0047] Expanding the dimensions of building surveying to include color and hidden defects, image-based 3D reconstruction allows this information to be quantitatively presented in a 3D model along with geometric information, greatly expanding the types of data in building surveying. It expresses and manages measurement results at three levels: building components, geographical environment, and temporal evolution, enabling applications such as building defect analysis, safety monitoring, and structural simulation, providing more accurate and systematic data support for the green renovation of existing buildings. Attached Figure Description

[0048] Figure 1 This is a flowchart of a three-dimensional reconstruction method based on digital image acquisition provided by an embodiment of the present invention;

[0049] Figure 2 This invention provides an image acquisition technology framework for existing buildings.

[0050] Figure 3 This is a schematic diagram of a modeling process provided in an embodiment of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0052] like Figure 1As shown, this embodiment provides a 3D reconstruction method based on digital image acquisition for the green renovation of existing buildings. The method starts with digital image acquisition, followed by high-precision 3D modeling, and ultimately obtains a database of the buildings, based on different building categories. In this process, semi-automated image acquisition yields digital images of the existing buildings; an automated 3D reconstruction algorithm based on digital images transforms non-quantitative local information (images with shooting tilt and lens distortion) into quantitative global information (3D model); and manual data processing endows the model with semantic information (at the level of building components, information on materials, structure, defects, etc.) and performs georeferencing to meet subsequent applications based on the actual building condition, such as safety monitoring and structural simulation. The specific steps are as follows:

[0053] S1: Establish a technical framework for field measurement for different types of existing buildings to ensure accurate measurement of data in three aspects: geometry, color, and defects.

[0054] In step S1, measurement methods for different types of building characteristics are designed. In many current renovations of existing buildings, in building-oriented surveying, simply obtaining three-dimensional data of the wall surface is far from sufficient; color and hidden defects (such as wall hollowing) are important bases for technical intervention. This invention summarizes the advantages and disadvantages of different types of building characteristics and surveying methods, and outlines commonly used surveying methods applicable to various buildings. In the early stages of surveying, appropriate surveying tools and methods must first be selected based on the building characteristics, such as... Figure 2 As shown, the specific steps include:

[0055] S11: For 3D reconstruction, laser scanners require an external camera. This is because current built-in cameras in laser scanners typically have 8-bit color depth, which can only distinguish 256 colors. However, due to cost and portability considerations, the combination of laser scanning and an external camera is currently difficult to use on a large scale. Passive optical sensors used in digital image acquisition have better sensitivity to light and heat: current SLR cameras have 24-bit color depth and can distinguish more than 16 million colors; thermal infrared imagers can detect hidden defects inside walls through abnormal temperature distribution. (Color depth is the number of bits used to store the color of 1 pixel in computer graphics. If the color depth is n, it can distinguish the superposition of 2n colors.)

[0056] S12: Balancing measurement accuracy and modeling completeness, excessive image acquisition may reduce measurement accuracy; however, insufficient image acquisition may lead to incomplete 3D reconstruction. To ensure modeling accuracy and completeness and avoid wasting resources on a large number of acquired images, it is essential to standardize the measurement parameters of the building to be surveyed before shooting. According to the principles of image modeling, the parameters that need to be determined include shooting distance, image overlap rate, and the number of images acquired. Among these, the image overlap rate and the number of images acquired are calculated based on the visible range.

[0057] The above parameters are obtained based on the relationship between the world coordinate system, pixel coordinate system, and image coordinate system. The specific calculation formula is as follows:

[0058] Shooting distance: Among them, s x p represents the camera sensor size. x f represents pixels, and f represents focal length.

[0059] The formula for calculating the number of images N and the image overlap rate is as follows:

[0060] in, B represents the overlapping area of ​​adjacent images, R represents the tolerance, and L represents the radius of the circumscribed circle of the building surface. L is the visible length of the building in the world coordinate system.

[0061] S13: For existing buildings and greening renovations of neighborhoods, the accuracy of color measurement is crucial. Currently, the Commission Internationale de L'Eclairage (CIE) specifies standard color measurement methods, defines various standard light sources, and recommends standard measurement geometry. However, in actual measurements, it is difficult to achieve standard light sources and geometry. To avoid image color differences, color correction is necessary. The traditional solution is to use a 24-color chart during photography, and then extract color templates from it during post-processing, assigning them to other images under the same lighting conditions. However, this method is labor-intensive. This invention proposes to use a multi-scale Retinex algorithm with color restoration, namely a uniform illumination algorithm based on an illuminance and reflectance model. Uniform illumination is performed first, followed by correction. Specifically, the Retinex algorithm is input using MATLAB programming, a color correction factor is calculated before multi-scale operations, the image to be corrected is imported, and automatic correction is performed. Much research has been conducted on the Retinex algorithm for color correction, and will not be elaborated upon here.

[0062] S2: To determine the applicability of different algorithms in different types of existing buildings, a technical process for image-based 3D reconstruction of existing buildings is established.

[0063] In a specific stage (such as image recognition), the accuracy of related algorithms (such as ASFIT, SIFT, SURF) is compared horizontally (e.g., the number of correctly matched corresponding points or comparison with point cloud models obtained by laser scanning). Other stages involving algorithm evaluation include: image processing, generating sparse point clouds, generating dense point clouds, generating mesh surfaces, texture mapping, and generating BIM models from point clouds.

[0064] In image-based 3D reconstruction, current commercial software suffers from opaque operation processes, preventing users from intervening in the modeling process once image acquisition is complete. However, open-source software in the field of computer vision offers the possibility of optimizing modeling accuracy and completeness at each stage of the modeling process. It can quantitatively evaluate key algorithms in each stage of 3D reconstruction based on the characteristics of different building types in terms of layout, appearance, materials, and texture. Figure 3 As shown, the specific steps include:

[0065] S21: In the image processing stage, Bentley ContextCapture is used to import various types of positioning data, such as GPS markers and anchor points, ensuring the interoperability of coordinate system data with the GIS solution. Descartes is used to process raster images, converting them into vector engineering drawings. Old documents are vectorized using raster and vector editing, cleaning, and processing tools, handling hybrid workflows. Output rulers, scales, and positioning are used to set image size and scale, associating images with vector information for accurate reuse later.

[0066] S22: Utilize Bentley Pointools to enhance, segment, and classify point clouds, and combine this with an engineering model to detect conflicts within the point cloud, performing conflict detection between real-world data and proposed designs. Specific steps include:

[0067] S221 distinguishes point clouds;

[0068] Compare two point clouds within the same region and identify any additions or subtractions in the data. Changes can be detected using the difference tool and monitored in real time. By synchronizing with the original data source, the generated model is updated to the latest version in real time. The value of this is having a global, up-to-date, and comprehensive representation of all data, which can be used to perform analyses using various display modes.

[0069] S222: Edit point cloud;

[0070] Edit large datasets of point clouds using point layer techniques. Move points between layers to isolate areas for detailed editing. Manipulate, clean, or subdivide point cloud models to clean and enrich them, making them more reusable and reducing modeling time.

[0071] S223: Mesh mapping, texture mapping;

[0072] Manipulate realistic meshes and scalable models with hundreds of millions of triangular tiles. Import, refine, and export meshes in numerous formats. Generate accurate georeferenced 3D models in various GIS formats, including true orthophotos and the new Cesium 3D Tiles. Utilize PCL to convert the images from step S1 into texture materials, perform texture mapping, and obtain the final GIS model.

[0073] S23: Create a BIM model based on point cloud data. Use a visual programming platform to combine functional nodes and construct a node chain for automatically generating the BIM entity model from the outline of 3D point cloud slices. Reconstruct the 3D point cloud through step S22, filter and classify the point cloud to obtain component-level point clouds with unique attributes, export coordinate point formats with 3D semantics, and import them into Dynamo. The specific modeling process is as follows:

[0074] S231: Calculate the bounding box surface area of ​​the point cloud, and then calculate the average surface density of the point cloud based on the number of points. The calculation of the feature vectors and feature values ​​of the point cloud can be implemented directly using C++ programming, or using the PCL point cloud library, or using Python's numpy functions.

[0075] S232: Point cloud slicing, input x as the slice quantity parameter to obtain the threshold of the point set on the slicing plane;

[0076] S233: Project two adjacent point sets to the tangent plane to find their intersection, fit the discrete points, and thus obtain x tangent boundary contours;

[0077] S234: Lofting and merging adjacent contour boundaries to form a segmented solid model;

[0078] S235: Segmented entity combination, through manual data processing, endowing the entity model with architectural semantic information (at the level of building components, endowing them with corresponding information such as materials, structure, and defects).

[0079] S3: Using BIM combined with GIS to express and manage measurement results at three levels: building components, geographical environment, and time evolution, and to provide scientific data support for the research and management of post-green building renovation assessment of existing buildings;

[0080] In step S3, based on the BIM / GIS measurement transformation results, finite element software is used to simulate and calculate the wind, light, and thermal environments of the existing building and site. DesignBuilder is used to calculate the energy consumption of the original building and greenhouse gas emissions. Combining BIM and GIS, the following are expressed: 1) Building form, building shape, and site environment; 2) Building color and materials; 3) Hidden building defects. Based on the building results, the existing physical environment, energy consumption, and carbon emissions of the building are analyzed and calculated. Step S3 includes the following steps:

[0081] S31: Utilize the convenient BIM import mechanism provided by SuperMap GIS software to achieve interaction between the GIS environmental model and the BIM building model, and obtain a high-precision 3D model result with micro-level architectural semantic information (materials, structure, defects, etc.) and macro-level geographic environmental information.

[0082] S32: Establish an external reference for the measured building (using a total station or GPS), unify the measurement results from different times into a single coordinate system, and obtain the building's change trend within that time period (such as floor settlement, structural displacement, wall cracking, surface peeling, changes in surrounding vegetation, etc.), providing scientific data for safety monitoring and technical intervention.

[0083] S33: A CFD-DEM coupled simulation method is employed, considering the influence of terrain factors and utilizing multi-technology platforms in a coordinated simulation. The obtained model results are then used to simulate and analyze the real-world wind, light, and heat environments, generating wind speed and temperature cloud maps of existing buildings. This provides comparative parameters for subsequent green renovation and structural modifications of existing buildings. Current commercial finite element software is compatible with various models, can transfer data error-free with subsequent ANSYS meshing software, and can read file formats output by almost all mainstream CAD software.

[0084] S34: Import the model results into energy consumption calculation software to calculate carbon emissions and greenhouse gas emissions.

[0085] S4: Establish an integrated green renovation technology system that couples wind, solar, and thermal environments, targeting carbon emissions, thermal comfort, and cost-effectiveness. Quantify the single-objective sensitivity of various forms, structures, systems, and equipment strategies for typical buildings through multi-technology platform-based simulation. Based on the data results from step S3, redesign and renovate, and then use multi-platform-based simulation again to calculate the differences in various simulation calculation data before and after the renovation. This process is repeated until a reasonable renovation method is found to assist in subsequent renovation decisions. Specifically, this includes the following steps:

[0086] S41: Based on the BIM model obtained in step S2, carry out a renovation, and import the results into finite element simulation software to simulate the wind, solar and thermal environment, and obtain wind speed and temperature cloud maps of the renovated building.

[0087] S42: Based on the BIM model obtained in step S2, carry out a transformation and import the results into energy consumption calculation software to obtain data such as carbon emissions and greenhouse gas emissions;

[0088] S43: Compare the data obtained in step S41 with the existing building wind speed and temperature cloud map in step S3, and continue to modify the building envelope parameters based on the comparison results.

[0089] S44: Compare the data obtained in step S42 with the carbon emissions and greenhouse gas emissions of existing buildings in step S3, calculate the cost increment based on the comparison results, and calculate the cost and benefits of green renovation.

[0090] S45: Repeat the above steps until the optimal green renovation plan is obtained.

[0091] In step S1, digital images of the existing building are obtained through semi-automated image acquisition, accurately reflecting the building's color information, geometric information, damage information, and surrounding environment information.

[0092] In step S2, a precise 3D geographic reference model in GIS format is quickly generated by processing the point cloud data. This model includes point clouds, break lines, raster digital elevation models, and existing triangulated irregular networks. By synchronizing with the original data source, the scalable environment model can be updated to the latest version in real time.

[0093] Furthermore, by filtering, segmenting, slicing, and projecting point cloud data to obtain a uniquely attributed build-level point cloud, and inputting it into Dynamo, a node chain is constructed, enabling automated BIM modeling based on real-world data.

[0094] In step S3, BIM and GIS are combined through the mediation between measurement and application, and the GPS positioning information input in the image processing stage of step S2, to fully express the measurement results at the micro level (building components) and the macro level (geographical environment).

[0095] Based on the measurement transformation results of BIM / GIS, the hidden defects and service life of the building to be measured can be obtained by observing the changing trends of the model.

[0096] Finite element method (FEM) software was used to simulate and calculate the wind, light, and thermal environments of the existing building and site. Energy consumption calculation software was used to calculate the energy consumption of the original building and the greenhouse gas emissions.

[0097] In step S4, the energy consumption calculation results obtained from the modified building model are compared with the data in step S3. The design is continuously improved until the most reasonable modification method is explored to assist in the later modification decision.

[0098] In summary, by utilizing the technical solutions described above, the dimensions of building surveying can be expanded to include color and hidden defects. Through image-based 3D reconstruction, this information, along with geometric information, can be quantitatively presented in a 3D model, greatly expanding the data types available for building surveying. Measurement results can be expressed and managed at three levels: building components, geographical environment, and temporal evolution. This data can be used for building defect analysis, safety monitoring, structural simulation, and other applications, providing more accurate and systematic data support for the green renovation of existing buildings.

[0099] It should be noted that, depending on the implementation needs, the various steps / components described in this application can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0100] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A three-dimensional reconstruction method based on digital image acquisition for green renovation of existing buildings, characterized in that, include: S1: Establish a technical framework for field measurement for different types of buildings to be measured, and measure data in three aspects: geometry, color, and defects. S2: To assess the applicability of different algorithms in different types of existing buildings, establish an image-based 3D reconstruction technology for existing buildings; S3: Use BIM combined with GIS to express and manage measurement results at three levels: building components, geographical environment, and time evolution; S4: Establish an integrated green transformation technology that couples wind, solar, and thermal environments, with the goals of reducing carbon emissions, improving thermal comfort, and achieving cost-effectiveness. Step S3 includes: S31: Utilize the convenient BIM import mechanism provided by SuperMap GIS software to achieve interaction between the GIS environmental model and the BIM building model, and obtain a high-precision three-dimensional model with micro-architectural semantic information and macro-geographical environmental information. S32: Establish external references for the building to be surveyed, unify the high-precision 3D models at different times into a coordinate system, and obtain the building's change trend within the same time period; S33: The CFD-DEM coupled simulation method is used for analysis, taking into account the influence of terrain factors and multi-technology platform linkage simulation. The high-precision three-dimensional model is used to simulate and analyze the wind, light and heat real environment to obtain the wind speed and temperature cloud map of the building to be surveyed. S34: Import the high-precision 3D model into the energy consumption calculation software to calculate carbon emissions and greenhouse gas emissions.

2. The method according to claim 1, characterized in that, Step S1 includes: S11: Acquire color images of the building to be surveyed through a camera sensor and thermal infrared images of the building to be surveyed through a thermal infrared imager, so as to identify hidden defects inside the wall through abnormal temperature distribution. S12: Determine the shooting distance, image overlap rate, and number of images to be acquired for the building to be surveyed; S13: The color image is corrected by using the multi-scale Retinex algorithm with color restoration to obtain the corrected color image.

3. The method according to claim 2, characterized in that, Step S12 includes: by Determine the shooting distance, among which, Represents the camera sensor size. f represents pixels, and f represents focal length. Depend on =arctan Determine the angular range corresponding to the overlapping regions of adjacent images, by Determine the number of image acquisitions N, where, R represents the tolerance, R represents the radius of the circumscribed circle of the building surface, and L is the visible length of the building in the world coordinate system.

4. The method according to claim 1, characterized in that, Step S2 includes: The system generates feature recognition and matching for color images, then generates sparse point clouds from measured ground control points, then generates dense point clouds from sparse point clouds, and finally generates grid surfaces from dense point clouds. Combined with thermal infrared images, a texture mapping is formed to obtain the final GIS model. Based on the texture-mapped point cloud data, a BIM model is created. The Dynamo visualization programming platform is used to combine functional nodes to automatically generate a BIM entity model by constructing the outline of three-dimensional point cloud slices.

5. The method according to claim 4, characterized in that, Step S2 includes: S21: Use Bentley ContextCapture to import location data, ensure the interoperability of coordinate system data with GIS solutions, process color images using Descartes, convert raster images to vector drawings, vectorize legacy documents using raster and vector editing, cleaning and processing tools, handle mixed workflows, set image size and scale using output rulers, scales and positioning, and associate color images with vector information; S22: Use Bentley Pointools to enhance, segment, and classify dense point clouds, and combine them with engineering models to detect conflicts in dense point clouds. Perform conflict detection between real-world data and proposed designs, and obtain a GIS model that can be updated to the latest version in real time by synchronizing with the original data source. S23: Create a BIM model based on the textured point cloud data, and use the Dynamo visual programming platform to combine functional nodes to automatically generate a BIM building model by constructing the outline of point cloud slices.

6. The method according to claim 5, characterized in that, Step S22 includes: S221: Compare dense point clouds in the same area and identify any additions or subtractions. Use the difference tool to detect changes and monitor them in real time to synchronize the dense point cloud with the original image data source. S222: Use point layer technology to edit large datasets of dense point clouds, manipulate, clean or subdivide dense point clouds, divide dense point clouds into various different layers, realize isolated editing and attribute differentiation of dense point clouds. Among them, isolated editing edits the monitored changing point clouds separately, and attribute differentiation obtains a build-level point cloud with unique attributes. S223: Manipulate realistic meshes and scalable models with hundreds of millions of triangular tiles, import, refine and export meshes in numerous formats, generate accurate georeferenced 3D models in various GIS formats, utilize PCL for texture mapping, and obtain the final GIS environment model.

7. The method according to claim 6, characterized in that, Step S23 includes: S231: Calculate the bounding box surface area S of the dense point cloud, and then calculate the average surface density of the dense point cloud based on the number N of dense point clouds. ; S232: Point cloud slicing, setting point cloud density. Given the slice threshold and x as the number of slices parameter, calculate the set of points on the tangent plane. S233: For a set of points on the tangent plane, project two adjacent sets of points onto the tangent plane to find their intersection, fit the discrete points, and thus obtain x boundary contours of the cutting plane; S234: Loft and merge adjacent contour boundaries to form a segmented solid model; S235: By combining segmented solid models and processing data manually, corresponding material, structural, and defect information is assigned to building components to obtain a BIM building model.

8. The method according to claim 7, characterized in that, Step S4 includes: S41: Based on the BIM building model, carry out a renovation, and import the renovation results into finite element simulation software to simulate the wind, solar and thermal environment and obtain wind speed and temperature cloud maps of the renovated building. S42: Based on the BIM building model, carry out a renovation, and import the renovation results into energy consumption calculation software to obtain carbon emissions and greenhouse gas emissions; S43: Compare the wind speed and temperature cloud map of the modified building with the wind speed and temperature cloud map of the building to be measured, and continue to modify the building envelope parameters based on the comparison results; S44: Compare the carbon emissions and greenhouse gas emissions obtained in step S42 with the carbon emissions and greenhouse gas emissions of the building to be measured in step S34. Calculate the incremental cost based on the comparison results, and calculate the cost and benefits of green retrofitting. S45: Repeat the above steps until the optimal green renovation plan is obtained.

Images