Disaster area evaluation method and system for realizing BIM (Building Information Modeling) rapid modeling based on visual large model

Abstract

The invention discloses a disaster area evaluation method and system for realizing BIM (Building Information Modeling) rapid modeling based on a visual large model. The method comprises the following steps of collecting pictures of different angles of a local area of a disaster area; extracting a feature vector of each picture, and performing feature fusion to obtain a comprehensive feature vector of the local area of the disaster area; performing three-dimensional reconstruction on the comprehensive feature vector by using a trained ZeroRF model to obtain three-dimensional coordinate data of a local area of the disaster area; according to the three-dimensional coordinate data, constructing a disaster area BIM model of the local area of the disaster area; and performing difference operation by using the disaster area BIM model and the complete BIM model of the target area to obtain a disaster area evaluation result. According to the method, the pictures are collected in an automatic mode and are processed by using the visual large model, so that the time period from image collection to BIM modeling and evaluation is greatly shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of disaster assessment, and more particularly to a disaster surface assessment method and system for realizing BIM rapid modeling based on a visual large model. Background Art

[0002] With the frequent occurrence of natural disasters today, it is crucial to quickly and accurately assess the damage and impact range of the disaster-affected areas for emergency response and post-disaster reconstruction. Traditional disaster area assessment methods mainly rely on manual on-site surveys and manual measurements, which are not only time-consuming and inefficient, but also easily affected by subjective factors, making it difficult to provide comprehensive and accurate assessment results.

[0003] In recent years, with the rapid development of computer vision and deep learning technology, image recognition and 3D reconstruction technology based on visual big models have been widely used. Visual big models can extract rich feature information from images through deep learning algorithms and generate 3D models with a certain degree of accuracy. Therefore, how to combine building information modeling (BIM) technology to achieve 3D reconstruction and assessment of local disaster-stricken areas in a short time is a technical problem that needs to be solved urgently. Summary of the invention

[0004] In view of this, the present invention provides a disaster-affected surface assessment method and system for realizing BIM rapid modeling based on a visual large model.

[0005] In order to achieve the above object, the present invention adopts the following technical solution:

[0006] On the one hand, a disaster area assessment method for realizing BIM rapid modeling based on a visual large model is disclosed, comprising the following steps:

[0007] Collect pictures of the local area of ​​the disaster area from different angles;

[0008] Extract the feature vector of each image and perform feature fusion to obtain the comprehensive feature vector of the local area of ​​the disaster-stricken surface;

[0009] The trained ZeroRF model is used to perform three-dimensional reconstruction on the comprehensive feature vector to obtain three-dimensional coordinate data of the local area of ​​the disaster-affected surface;

[0010] Construct a disaster-affected surface BIM model of the local area of ​​the disaster-affected surface based on the three-dimensional coordinate data;

[0011] The disaster-affected surface BIM model is used to perform difference calculation with the complete BIM model of the target area to obtain the disaster-affected surface assessment results.

[0012] Furthermore, pictures of the local area of ​​the disaster area from different angles are collected, including:

[0013] Horizontal rotation acquisition: Fix the drone with a camera at a certain height, rotate horizontally around the target area to obtain 360-degree continuous images of the local area of ​​the disaster area;

[0014] Vertical rotation acquisition: Fix the drone equipped with a camera in a vertical position and shoot up and down to obtain pictures of the local area of ​​the disaster area from different height perspectives.

[0015] Furthermore, the feature vector of each image is extracted and feature fusion is performed to obtain the comprehensive feature vector of the local area of ​​the disaster-stricken surface; specifically, the following steps are included:

[0016] Pre-process each collected picture of the local area of ​​the disaster-stricken surface;

[0017] Use convolutional neural networks to extract features from the pre-processed local area images of the disaster-stricken surface and obtain feature vectors for each pre-processed image;

[0018] The feature vectors of the images of the local area of ​​the disaster-stricken surface at different angles are fused to obtain a comprehensive feature vector of the local area of ​​the disaster-stricken surface.

[0019] Furthermore, the trained ZeroRF model is used to perform three-dimensional reconstruction on the comprehensive feature vector to obtain three-dimensional coordinate data of the local area of ​​the disaster-affected surface; specifically, the method includes:

[0020] The 3D reconstruction module in the trained ZeroRF model is used to convert the extracted comprehensive feature vectors of the local area of ​​the disaster-stricken surface into 3D coordinates.

[0021] Furthermore, a disaster-affected surface BIM model of the local area of ​​the disaster-affected surface is constructed based on the three-dimensional coordinate data, specifically including:

[0022] According to the data import requirements and formats of BIM software, the format of the 3D coordinate data of the local area of ​​the disaster area is converted;

[0023] The format-converted three-dimensional coordinate data is imported into the BIM software to generate a disaster-affected surface BIM model of the local area of ​​the disaster-affected surface.

[0024] Furthermore, the BIM model of the affected area is used to perform a difference operation with the complete BIM model of the target area to obtain the affected area assessment result, which specifically includes the following steps:

[0025] The BIM model of the affected area and the complete BIM model of the target area are simultaneously input into the BIM software. The difference operation function provided by the BIM software is used to compare and analyze the two models element by element to obtain the analysis difference operation results.

[0026] If the analysis difference calculation result exceeds the preset threshold, it is considered that the component may be damaged.

[0027] Furthermore, the difference calculation function provided by the BIM software is used to compare and analyze the two models element by element to obtain the analysis difference calculation results; specifically, the following are included:

[0028] Assume that the coordinates of a point in the component of the complete BIM model are: original ,Y original ,Z original , the component coordinates of the BIM model of the disaster-affected surface are X new ,Y new ,Z new , then use the following difference operation formula to get the analytical difference operation result:

[0029] ΔX=X original -X new

[0030] ΔY=Y original -Y new

[0031] ΔZ=Z original -Z new .

[0032] Another aspect of the present invention discloses a disaster surface assessment system for realizing BIM rapid modeling based on a visual large model, comprising:

[0033] Data collection module: used to collect pictures of the local area of ​​the disaster area from different angles;

[0034] Feature extraction module: used to extract the feature vector of each image and perform feature fusion to obtain the comprehensive feature vector of the local area of ​​the disaster area;

[0035] A module for acquiring the three-dimensional coordinates of a local area of ​​the disaster-affected surface: used to perform three-dimensional reconstruction of the comprehensive feature vector using the trained ZeroRF model to obtain the three-dimensional coordinate data of the local area of ​​the disaster-affected surface;

[0036] Disaster surface BIM model construction module: used to construct the disaster surface BIM model of the local area of ​​the disaster surface based on the three-dimensional coordinate data;

[0037] Difference calculation module: used to perform difference calculation using the disaster-affected surface BIM model and the complete BIM model of the target area to obtain the disaster-affected surface assessment results.

[0038] Preferably, in the above-mentioned system, in the disaster-affected surface BIM model construction module, a disaster-affected surface BIM model of a local area of ​​the disaster-affected surface is constructed according to the three-dimensional coordinate data, specifically including:

[0039] According to the data import requirements and formats of BIM software, the format of the 3D coordinate data of the local area of ​​the disaster area is converted;

[0040] The format-converted three-dimensional coordinate data is imported into the BIM software to generate a disaster-affected surface BIM model of the local area of ​​the disaster-affected surface.

[0041] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a disaster surface assessment method and system for realizing BIM rapid modeling based on a visual large model, which has the following beneficial effects:

[0042] Efficiency: The present invention collects images in an automated manner and processes them using a large visual model, which greatly shortens the time period from image acquisition to BIM modeling and evaluation.

[0043] Rich information: The present invention utilizes the existing BIM information and the generated XYZ coordinates to construct a BIM model containing detailed structural and attribute information, providing comprehensive and accurate data support for the assessment of the disaster-stricken area in the local area.

[0044] High degree of automation: The entire process of the present invention is largely automated, which reduces manual intervention and subjective judgment and reduces the probability of error.

[0045] Strong flexibility and adaptability: The present invention can be applied to disaster-stricken areas of different types and sizes, and has strong versatility and scalability. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0047] Figure 1 The present invention provides a schematic diagram of the overall method flow chart. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] The embodiment of the present invention discloses a disaster area assessment method based on a visual large model to realize BIM rapid modeling, such as Figure 1 As shown, the following steps are included:

[0050] Collect pictures of the local area of ​​the disaster area from different angles;

[0051] Extract the feature vector of each image and perform feature fusion to obtain the comprehensive feature vector of the local area of ​​the disaster-stricken surface;

[0052] The trained ZeroRF model is used to perform three-dimensional reconstruction on the comprehensive feature vector to obtain three-dimensional coordinate data of the local area of ​​the disaster-affected surface;

[0053] Construct a disaster-affected surface BIM model of the local area of ​​the disaster-affected surface based on the three-dimensional coordinate data;

[0054] The disaster-affected surface BIM model is used to perform difference calculation with the complete BIM model of the target area to obtain the disaster-affected surface assessment results.

[0055] The technical details of the present invention are further described below in conjunction with the specific operation process:

[0056] First, collect multi-angle pictures of the local area of ​​the disaster area. After a natural disaster occurs, in order to quickly and accurately assess the damage and impact range of the disaster area, it is necessary to collect multi-angle pictures of the local area of ​​the disaster area. These pictures will be used for subsequent image processing. The collection process is described as follows:

[0057] Develop a detailed picture collection plan based on the geographical features and damage of the disaster-stricken area. Determine the area, angle, and height to be collected to ensure that all key areas are covered. Then select the appropriate collection equipment, such as high-resolution drones, satellite cameras, or other professional photography equipment, to ensure that the collected pictures have sufficient clarity and details.

[0058] Before formal collection, it is usually necessary to debug and calibrate the collection equipment to ensure that the device's shooting parameters (such as exposure time, aperture size, etc.) are set correctly to avoid poor image quality due to equipment problems.

[0059] After completing the above preparations, collect pictures of the local area of ​​the disaster area from multiple angles according to the collection plan. The specific steps include the following:

[0060] Horizontal rotation acquisition: Fix the device, such as a drone carrying a camera, at a certain height, and rotate it horizontally around the target area to obtain 360-degree continuous images.

[0061] Vertical rotation acquisition: Fix the device carrying the camera in a certain position, rotate it up and down to take pictures at different heights to capture details in the vertical direction.

[0062] Multi-location acquisition: Alternatively, you can set up the equipment at different locations to capture the same target area, ensuring comprehensive image information is obtained from different perspectives.

[0063] In order to ensure the accuracy of image data, during the acquisition process, pay attention to the influence of environmental conditions, such as light changes, weather conditions, etc. Adjust equipment parameters when necessary to ensure stable image quality.

[0064] In specific operations, it is necessary to store the collected image data in real time, and perform preliminary classification and management. Detailed collection information (such as shooting time, location, angle, etc.) is recorded for each image for subsequent processing and analysis.

[0065] In order to further obtain high-quality images, it is necessary to conduct quality checks on the collected images and select images with high clarity and complete information. For images with poor quality or missing information, retake them in time to ensure the integrity and reliability of the data.

[0066] After obtaining pictures of the local area of ​​the disaster area at different angles, it is necessary to use a large visual model to obtain the XYZ coordinates of the target area; specifically, in the embodiment of the present invention, the ZeroRF model is used to obtain the xyz coordinates of the local area. ZeroRF is a deep learning model that combines image processing and 3D reconstruction technology. Similar models equivalent to this can obtain similar output results.

[0067] The steps are as follows:

[0068] Step 1: Image upload: The user uploads multiple drone aerial images of the same target area.

[0069] Step 2: Image preprocessing: Preprocess the uploaded images, including denoising, scaling, and normalization, to ensure that the images meet the input requirements of the ZeroRF model.

[0070] Step 3: ZeroRF model processing: First, the preprocessed image is input into the input layer of the ZeroRF model.

[0071] The feature extraction layer (convolutional neural network CNN) in the ZeroRF model is used to extract the feature vector in the image. This process can be expressed as: F = CNN (I), where F represents the extracted feature vector and I represents the preprocessed image.

[0072] Then, the 3D reconstruction module of the ZeroRF model is used to convert the extracted feature vector into 3D coordinates. This process can be expressed as: C = R(F), where C represents the predicted 3D coordinate vector [X, Y, Z], R represents the 3D reconstruction function, and F represents the extracted feature vector.

[0073] The output layer of the ZeroRF model outputs the predicted 3D coordinates [X,Y,Z].

[0074] Detailed calculation process example:

[0075] Assume that a user uploads multiple drone aerial images of the same target area I 1 ,I 2 ,…,I n , the following are the detailed calculation steps:

[0076] Image preprocessing: For each image I i Perform denoising, scaling and normalization to obtain the preprocessed image I i ′.

[0077] Feature extraction: Use CNN to extract each preprocessed image I i′ The features in the equation are obtained by taking the feature vector F i : Feature extraction can be expressed by the following formula: F i =CNN(I i ).

[0078] 3D reconstruction: Use the 3D reconstruction module of the ZeroRF model to convert the feature vector Fi into 3D coordinates. The expression is as follows: C i =R(F i ), where R is a complex nonlinear function, and its specific implementation depends on the design and training of the ZeroRF model.

[0079] Multi-view fusion: The feature vectors F of multiple images are 1 ,F 2 ,…,F n Fusion is performed to obtain the comprehensive feature vector F total , the expression is as follows: F total =Fuse(F 1 ,F 2 ,…,F n ), where Fuse represents the multi-view fusion function.

[0080] Output result: The comprehensive feature vector F total Input to the output layer of the ZeroRF model to obtain the final three-dimensional coordinates [X, Y, Z]. This process is expressed as: final =R(F total ).

[0081] Model principle description: Assume that the feature vector F extracted by CNN i for [f i1 ,f i2 ,…,f in ], the 3D reconstruction function R of the ZeroRF model can be expressed as:

[0082] C i =R(F i )=[Xi ,Y i ,Z i ];

[0083] The specific form of R depends on the training and design of the model. Assume that R is a multi-layer perceptron (MLP), and its output layer can be expressed as:

[0084] X i =wX 1 f i1 +wX 2 f i2 +…+wX n f in +b X ;

[0085] Y i =wY 1 f i1 +wY 2 f i2 +…+wY n f in +b Y ;

[0086] Z i =wZ 1 f i1 +wZ 2 f i2 +…+wZ n f in +b Z ;

[0087] Among them, w ij is the weight parameter, b X ,b Y, b Z is the bias parameter.

[0088] The multi-view fusion function Fuse can be expressed as: total =Fuse(F 1 ,F 2 ,…,F n ); wherein, the multi-view fusion function Fuse can be a simple average fusion, weighted fusion or other complex fusion methods, and the specific implementation depends on the design and training of the model.

[0089] Use the known length of the local area to compare with XYZ to obtain the conversion ratio. In this step, in order to ensure the accuracy and reliability of the three-dimensional coordinates (XYZ) generated by the visual large model, it is often necessary to calibrate and convert by comparing with the actual known length. This process can correct the errors in the model prediction and improve the accuracy of the final result.

[0090] The specific calibration and conversion steps are as follows:

[0091] Determine the known length of the local area: In the disaster area, select a local area with a clear known length as a reference. For example, a specific marked line segment on a building, a survey line on a road, etc. Make sure the length of the selected area is accurately known and clearly visible in the image.

[0092] Mark the known length in the image: In the acquired image, accurately mark the known length of the selected local area. This can be achieved by drawing lines or adding annotations.

[0093] Extract the XYZ coordinates of the local area; use the three-dimensional coordinate data previously generated by the visual large model to extract the XYZ coordinates of the key points in the selected local area.

[0094] Calculate the distance between coordinates; Based on the extracted XYZ coordinates, calculate the distance between corresponding points in the local area.

[0095] Compare the known length with the calculated distance: Compare the calculated distance with the known length. If L known is a known length, L calculated For the calculated distance, the following comparison formula is used: Ratio = L calculated / L known , calculate the conversion ratio Ratio through the above comparison formula, if L calculated Greater than L known , then the conversion ratio Ratio is less than 1, indicating that the size predicted by the model is too large; on the contrary, if L calculated Less than L known , then the conversion ratio Ratio is greater than 1, indicating that the size predicted by the model is too small.

[0096] Apply the obtained conversion ratio Ratio to the entire 3D coordinate data, and make corresponding scaling adjustments to all XYZ coordinates. The specific adjustment steps are as follows:

[0097] Assuming the original XYZ coordinates are (X, Y, Z), the conversion ratio is R, and the adjusted coordinates are (X′, Y′, Z′), then:

[0098] X′=X×R;

[0099] Y′=Y×R;

[0100] Z′=Z×R;

[0101] Wherein, R represents the conversion ratio Ratio.

[0102] In actual operation, it is usually necessary to verify the adjusted coordinates to check whether they are more consistent with the actual situation. If there is still a deviation, it may be necessary to reselect the local area, optimize the comparison method, or further adjust the model parameters.

[0103] Use the known BIM information of the local area and the generated XYZ to perform rapid BIM modeling

[0104] Some BIM (Building Information Model) information may already exist in local areas of the disaster area. Combined with the XYZ coordinate data generated from the pictures, this information can be used to quickly build and improve the BIM model. The detailed steps are as follows:

[0105] Collect known BIM information in the local area; organize and collect existing BIM information in the local area of ​​the disaster area, including the structural information, component properties, material information, etc. of the building, which may be stored in previous project files, databases or special BIM management systems. Open BIM software (such as Revit), create a new project file or open an existing related project file. Import the generated XYZ coordinate data: Import the XYZ coordinate data of the local area generated by the visual large model into the BIM software. Before importing, the data needs to be converted and processed as necessary according to the data import requirements and format of the BIM software. Finally, combine the known BIM information and XYZ coordinate data: in the BIM software, associate and match the known BIM information with the imported XYZ coordinate data. For known BIM components, place and adjust them at the corresponding XYZ coordinate position according to their properties and location information; for newly generated XYZ coordinate points, determine the component type and properties to which they belong according to their position and the existing BIM component information around them. Finally, use the modeling tools of the BIM software to quickly build a BIM model of the local area based on the known information and generated coordinate data.

[0106] In order to ensure the accuracy of the BIM model of the constructed local area, it is necessary to verify the constructed BIM model, check its consistency with the actual situation and the accuracy of the data, and optimize and adjust the model according to the verification results to correct possible errors and deviations.

[0107] In this process, the BIM model of the local area can be integrated with the BIM model of the overall project to ensure the consistency and integrity of the model. In addition, the model can be expanded and refined as needed to add more details and information.

[0108] Finally, the present invention performs a difference operation on the BIM model established based on XYZ rapidization and the original BIM model to evaluate the affected area. In the affected area, by combining the XYZ coordinate data generated by the visual large model with the known partial BIM information, we have quickly established a preliminary BIM model of the affected area. In order to evaluate the disaster situation, we need to compare and analyze this newly established BIM model with the BIM model of the original completion drawing. Through the difference operation, we can identify the degree and scope of damage in the affected area.

[0109] The detailed steps are as follows: Before performing the difference operation, ensure that the BIM model of the original as-built drawing and the BIM model established based on XYZ rapidization are under the same measurement scale system and the units are consistent. Perform necessary calibration and alignment on the two models to ensure that they are completely matched in spatial position. Open the BIM software (such as Revit) and load the BIM model of the original as-built drawing and the BIM model established based on XYZ rapidization. Then perform the difference operation: In the BIM software, use the difference operation function it provides to compare and analyze the two models element by element.

[0110] Specifically, the software will compare the components at the same position in the two models one by one and calculate the difference in their XYZ coordinates. Assume that the component coordinates of the original model are (X original ,Y original ,Z original ), the component coordinates of the new model are (X new ,Y new ,Z new ), then the difference operation formula is:

[0111] ΔX=X original -X new ;

[0112] ΔY=Y original -Y new ;

[0113] ΔZ=Z original -Z new .

[0114] Based on the difference calculation results, the damage of the affected area is identified. If the ΔX, ΔY or ΔZ of a component exceeds the preset threshold, it is considered that the component may be damaged. For each component, its difference calculation results are recorded and a corresponding damage report is generated.

[0115] Generate a BIM model of the affected area. Based on the results of the difference calculation, generate a new BIM model that highlights the damage in the affected area. In BIM software, you can visualize the affected area by changing the color, transparency or other properties of the components.

[0116] Verify and optimize the assessment results, in order to verify the generated disaster-affected surface BIM model and ensure its accuracy and reliability. The assessment results can be optimized and adjusted according to the actual situation on site and expert opinions.

[0117] Finally, the BIM model of the disaster area is exported and a detailed assessment report is generated. The report should include information such as the damage in the disaster area, the scope of impact, and the severity. The analysis function of the BIM software is used to further analyze the specific situation of the disaster area and put forward corresponding rescue and reconstruction suggestions.

[0118] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0119] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A disaster-affected surface assessment method based on a visual large model to achieve BIM rapid modeling, characterized in that: The following steps are involved: Collect pictures of the local area of ​​the disaster area from different angles; Extract the feature vector of each image and perform feature fusion to obtain the comprehensive feature vector of the local area of ​​the disaster-stricken surface; The trained ZeroRF model is used to perform three-dimensional reconstruction on the comprehensive feature vector to obtain three-dimensional coordinate data of the local area of ​​the disaster-affected surface; Construct a disaster-affected surface BIM model of the local area of ​​the disaster-affected surface based on the three-dimensional coordinate data; The disaster-affected surface BIM model is used to perform difference calculation with the complete BIM model of the target area to obtain the disaster-affected surface assessment results.

2. According to claim 1, a disaster area assessment method based on a visual large model to achieve BIM rapid modeling is characterized by: Collect pictures of the affected area from different angles, including: Horizontal rotation acquisition: Fix the drone with a camera at a certain height, rotate horizontally around the target area to obtain 360-degree continuous images of the local area of ​​the disaster area; Vertical rotation acquisition: Fix the drone equipped with a camera in a vertical position and shoot up and down to obtain pictures of the local area of ​​the disaster area from different height perspectives.

3. According to claim 1, a disaster area assessment method based on a visual large model to achieve BIM rapid modeling is characterized in that: Extract the feature vector of each image and perform feature fusion to obtain the comprehensive feature vector of the local area of ​​the disaster area; specifically include: Pre-process each collected picture of the local area of ​​the disaster-stricken surface; Use convolutional neural networks to extract features from the pre-processed local area images of the disaster-stricken surface and obtain feature vectors for each pre-processed image; The feature vectors of the images of the local area of ​​the disaster-stricken surface at different angles are fused to obtain a comprehensive feature vector of the local area of ​​the disaster-stricken surface.

4. The disaster-affected surface assessment method based on a visual large model to realize BIM rapid modeling according to claim 1 is characterized in that: The trained ZeroRF model is used to perform three-dimensional reconstruction on the comprehensive feature vector to obtain the three-dimensional coordinate data of the local area of ​​the disaster area; specifically, the reconstruction includes: The 3D reconstruction module in the trained ZeroRF model is used to convert the extracted comprehensive feature vectors of the local area of ​​the disaster-stricken surface into 3D coordinates.

5. The disaster-affected surface assessment method based on a visual large model to realize BIM rapid modeling according to claim 1 is characterized in that: The disaster-affected surface BIM model of the local area of ​​the disaster-affected surface is constructed based on the three-dimensional coordinate data, including: According to the data import requirements and formats of BIM software, the format of the 3D coordinate data of the local area of ​​the disaster area is converted; The format-converted three-dimensional coordinate data is imported into the BIM software to generate a disaster-affected surface BIM model of the local area of ​​the disaster-affected surface.

6. The disaster-affected surface assessment method based on a visual large model to realize BIM rapid modeling according to claim 1 is characterized in that: The BIM model of the affected area is used to perform a difference operation with the complete BIM model of the target area to obtain the affected area assessment result, which specifically includes the following steps: The BIM model of the affected area and the complete BIM model of the target area are simultaneously input into the BIM software. The difference operation function provided by the BIM software is used to compare and analyze the two models element by element to obtain the analysis difference operation results. If the analysis difference calculation result exceeds the preset threshold, it is considered that the component may be damaged.

7. According to claim 6, a disaster area assessment method for realizing BIM rapid modeling based on a visual large model, using the difference operation function provided by the BIM software, performing element-by-element comparative analysis on the two models to obtain the analysis difference operation result; specifically comprising: Assume that the coordinates of a point in the component of the complete BIM model are: original ,Y original ,Z original , the component coordinates of the BIM model of the disaster-affected surface are X new ,Y new ,Z new , then use the following difference operation formula to get the analytical difference operation result: ΔX=X original -X new ΔY=Y original -Y new ΔZ=Z original -Z new 。 8. A disaster area assessment system based on a visual large model to achieve BIM rapid modeling, characterized by: include: Data collection module: used to collect pictures of the local area of ​​the disaster area from different angles; Feature extraction module: used to extract the feature vector of each image and perform feature fusion to obtain the comprehensive feature vector of the local area of ​​the disaster area; A module for acquiring the three-dimensional coordinates of a local area of ​​the disaster-affected surface: used to perform three-dimensional reconstruction of the comprehensive feature vector using the trained ZeroRF model to acquire the three-dimensional coordinate data of the local area of ​​the disaster-affected surface; Disaster surface BIM model construction module: used to construct the disaster surface BIM model of the local area of ​​the disaster surface based on the three-dimensional coordinate data; Difference calculation module: used to perform difference calculation using the disaster-affected surface BIM model and the complete BIM model of the target area to obtain the disaster-affected surface assessment results.

9. The disaster-affected surface assessment system for realizing BIM rapid modeling based on a visual large model according to claim 8 is characterized in that: In the disaster-affected surface BIM model construction module, the disaster-affected surface BIM model of the local area of ​​the disaster-affected surface is constructed according to the three-dimensional coordinate data, specifically including: According to the data import requirements and formats of BIM software, the format of the 3D coordinate data of the local area of ​​the disaster area is converted; The format-converted three-dimensional coordinate data is imported into the BIM software to generate a disaster-affected surface BIM model of the local area of ​​the disaster-affected surface.

Images