Real estate surveying and mapping rapid vectorization method based on deep learning

Abstract

The embodiment of the invention provides a real estate surveying and mapping rapid vectorization method based on deep learning, and belongs to the technical field of image processing, and the method specifically comprises the steps: 1, dividing real estate surveying and mapping data, and obtaining a table region and a main body graph region; 2, extracting coordinate number data of the main body shape from the table region by using a high-precision character recognition technology; step 3, adopting a trained UNET + + deep learning model to carry out automatic vectorization processing on the main body graph area to obtain a vectorization result; and step 4, accurately matching the vectorization result with the house building map by establishing a corresponding relation of the coordinate number data, and performing standardization processing on the vectorization result by applying an affine transformation technology based on actual space size information in the house building map to obtain final vector data with actual geographic coordinates. Through the scheme of the invention, the processing efficiency and accuracy are improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of image processing technology, and in particular, to a fast vectorization method for real estate surveying and mapping drawings based on deep learning. Background Art

[0002] Real estate surveying and mapping drawings are important documents in real estate management, used to record and express key information such as the floor plan layout, spatial dimensions, and property rights of houses. These drawings are not only the basic basis for real estate transactions and property rights registration, but also important reference materials for urban planning and real estate management. The vectorization processing of real estate surveying and mapping drawings is to convert traditional paper or scanned drawings into a vector format that can be recognized and processed by a computer, which is of great significance for establishing a modern real estate information management system and improving real estate management efficiency. Through vectorization processing, digital storage, rapid retrieval, and intelligent analysis of real estate data can be achieved, providing technical support for the standardized development of the real estate market.

[0003] At present, the vectorization work of real estate surveying and mapping drawings faces many technical problems. On the one hand, the quality of the surveying and mapping drawings is uneven, and the clarity of drawings from different eras and sources varies significantly. The drawings have been preserved for a long time, resulting in blurred edges or breaks in some lines. On the other hand, a large number of text information such as dimension markings and room numbers, as well as various seals and other imprints, are often superimposed on the surveying and mapping drawings. These additional information not only covers the original graphic elements, but also has obvious differences in complexity and integrity, bringing huge challenges to automated processing.

[0004] It can be seen that there is an urgent need for a fast vectorization method for real estate surveying and mapping drawings based on deep learning with high processing efficiency and accuracy. Summary of the Invention

[0005] In view of this, the embodiments of the present invention provide a fast vectorization of real estate surveying and mapping drawings based on deep learning, which at least partially solves the problem of poor processing efficiency and accuracy in the prior art.

[0006] The embodiments of the present invention provide a fast vectorization method for real estate surveying and mapping drawings based on deep learning, including:

[0007] Step 1, dividing the real estate surveying and mapping drawing data to obtain a table area and a main graphic area; Step 2, using high-precision text recognition technology to extract the coordinate number data of the main shape from the table area; The specific steps of Step 2 include: Step 2.1, by traversing all the recognized texts in the table area and applying a preset filtering rule, the system preferentially recognizes the text content without Chinese characters, and selects the longest string as the candidate measurement number; Step 2.2, perform format verification on the candidate measurement numbers using a preset regular expression; Step 2.3, determine the number of layers by identifying pure numeric text and screening the largest value not greater than 99, and determine the hierarchical information by identifying integers with a negative sign; Step 2.4, when some information cannot be recognized, the system will automatically mark the corresponding field as N, and comprehensively form the coordinate number data; Step 3, use a trained UNET++ deep learning model to automatically vectorize the main graphic area to obtain a vectorization result; Step 4, by establishing the corresponding relationship of the coordinate number data, precisely match the vectorization result with the building outline map, and based on the actual spatial dimension information in the building outline map, apply the affine transformation technology to standardize the vectorization result to obtain the final vector data with actual geographical coordinates; The specific steps of Step 4 include: Step 4.1, match the measurement numbers of the attribute information in the table area with the measurement numbers in the attribute table of the building outline map to generate a mapping relationship table; Step 4.2, select at least 4 evenly distributed corner points from the mapping relationship table as control points, and for the selected set of control points , construct a transformation matrix solution equation: ; Among them, is the pixel coordinate in the source image, is the coordinate in the target geographical coordinate system, is the affine transformation coefficient to be solved; Step 4.3, use the least squares method to solve the transformation matrix solution equation to obtain the optimal affine transformation matrix T, and apply the affine transformation matrix T to all vertex coordinates of the vectorization result generated by the UNET++ model. The transformed vertex coordinates are converted from the pixel coordinate system to the actual geographical coordinate system to obtain the final vector data with actual geographical coordinates.

[0008] According to a specific implementation manner of the embodiment of the present invention, the specific steps of Step 1 include: Step 1.1, calculate the row-column ratio of the image in the real estate surveying and mapping map data , when is greater than 1, rotate the image 90 degrees clockwise; Step 1.2, apply a Gaussian filter to perform noise reduction processing on the image, and on this basis, use the Canny edge detection operator to extract the contour of the image ; ; ; Among them, is the Gaussian kernel formula, is the pixel coordinate, is the standard deviation of the Gaussian kernel, which controls the smoothness, is the gradient intensity, is the horizontal gradient, is the vertical gradient, is the gradient direction. Along the gradient direction, the local maximum value is retained, the non-maximum points are suppressed, and the edge is refined; Step 1.3: For each detected contour, calculate its area and perimeter, set an area threshold, and retain the contours with an area greater than the area threshold; Step 1.4: Calculate the longitudinal position ratio of the contour, and based on this, select the optimal segmentation threshold through the analysis of the property mapping data , and according to the optimal segmentation threshold divide each contour with an area greater than the area threshold into a table area and a main graphic area.

[0009] According to a specific implementation manner of the embodiment of the present invention, after the step 1.2, the method further includes: Use morphological operations to eliminate small noises in the extracted contours: ; Among them, is the structural element, represents the dilation operation, represents the erosion operation, is the processed binary image, is the extracted contour.

[0010] According to a specific implementation manner of the embodiment of the present invention, the step 1.4 specifically includes: When the height of the contour , this contour is divided into a table area, which contains attribute information; When the height of the contour , this contour is divided into a main graphic area, which contains the geometric shape information of the house, is the total height of the image.

[0011] According to a specific implementation manner of the embodiment of the present invention, the step 4.1 specifically includes: Step 4.1.1: Construct the measurement number information extracted from the table area into a first association table; Step 4.1.2: Extract the measurement number from the house building line map attribute table and construct it into a second association table; Step 4.1.3: Perform a spatial join operation using the measurement ID as the join key to generate a mapping relationship table.

[0012] The fast vectorization solution for real estate surveying and mapping drawings based on deep learning in the embodiments of the present invention includes: Step 1, divide the real estate surveying and mapping drawing data to obtain a table area and a main graphic area; Step 2, use high-precision character recognition technology to extract the coordinate number data of the main shape from the table area; Step 3, use the trained UNET++ deep learning model to perform automatic vectorization processing on the main graphic area to obtain a vectorization result; Step 4, by establishing the corresponding relationship of the coordinate number data, accurately match the vectorization result with the building line drawing, and based on the actual spatial dimension information in the building line drawing, apply the affine transformation technology to standardize the vectorization result to obtain the final vector data with actual geographical coordinates.

[0013] The beneficial effects of the embodiments of the present invention are as follows: Through the solution of the present invention, through the innovative integration of deep learning and character recognition technology, the full-automatic vectorization processing of real estate surveying and mapping drawings is realized. Compared with the semi-automatic method that relies on manual intervention in the prior art, the processing efficiency and accuracy are greatly improved; the affine transformation method based on coordinate number matching is applied to solve the problem of accurate conversion from pixel coordinates to actual geographical coordinates, overcome the defect of relying on manual alignment in the prior art, and realize the high-precision automatic alignment of the vectorization result with the actual building outline. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0015] Figure 1 It is a flowchart of a fast vectorization method for real estate surveying and mapping drawings based on deep learning provided by the embodiments of the present invention; Figure 2 It is a specific implementation flowchart of a fast vectorization method for real estate surveying and mapping drawings based on deep learning provided by the embodiments of the present invention; Figure 3 It is a vectorization result of a real estate surveying and mapping drawing provided by the embodiments of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] The embodiments of the present invention will be described in detail below with reference to the drawings.

[0017] The following describes the embodiments of the present invention through specific examples. Those skilled in the art can easily understand the other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0018] It should be noted that the following describes various aspects of the embodiments within the scope of the appended claims. It should be obvious that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is illustrative only. Based on the present invention, those skilled in the art should understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects described herein can be used to implement the device and / or practice the method. In addition, this device can be implemented and this method can be practiced using other structures and / or functions in addition to one or more of the aspects described herein.

[0019] It should also be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present invention schematically. The diagrams only show the components related to the present invention and are not drawn according to the number, shape and size of the components in actual implementation. The type, quantity and proportion of each component in its actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0020] In addition, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0021] Traditional methods for vectorizing real estate surveying and mapping drawings mainly rely on manual tracing or semi-automated processing. The manual tracing method requires operators to manually trace the outlines and internal structures of houses through digitizers or computer software. This method not only takes a long time but is also easily affected by the subjective judgment of operators, making it difficult to ensure measurement accuracy and consistency. In addition, the efficiency of manual processing is low and it is difficult to meet the needs of large-scale real estate data processing. With the development of image processing technology, various automatic vectorization methods have emerged. Among them, the more common ones include edge detection-based methods and image segmentation-based methods. Edge detection-based methods mainly use operators such as Canny and Sobel to detect edge information in images and extract house outlines through techniques such as morphological processing and line tracking. However, such methods have high requirements for image quality. When processing images with large noise interference or blurred lines, incomplete edge detection results or false edge misidentifications are likely to occur. Image segmentation-based methods attempt to divide images into different semantic regions through techniques such as region growing and watershed algorithms. However, when dealing with real estate surveying and mapping drawings with complex structures, it is often difficult to accurately identify and segment the detailed features of houses. Especially when dealing with areas with overlapping seals or dense annotations, the segmentation effect is not ideal.

[0022] In recent years, deep learning technology has made breakthrough progress in the field of image processing. Especially, methods based on convolutional neural networks have shown powerful performance in tasks such as image classification, object detection, and semantic segmentation. This provides a new technical idea for solving the problem of vectorizing real estate surveying and mapping drawings. However, at present, the research on applying deep learning technology to the vectorization of real estate surveying and mapping drawings is relatively less, and there is still a lack of a fast vectorization method for real estate surveying and mapping drawings based on deep learning.

[0023] It can be seen that: (1) Traditional manual vectorization methods are time-consuming and laborious with low efficiency; existing automatic vectorization methods have high requirements for image quality and poor anti-interference ability. (2) There is a lack of the ability to identify and process the unique attribute table areas in real estate surveying and mapping drawings; the correlation between the vectorization results and the actual spatial coordinate system is poor, making it difficult to perform actual size calibration.

[0024] The embodiments of the present invention provide a fast vectorization method for real estate surveying and mapping drawings based on deep learning, and the method can be applied to the process of surveying and mapping drawing management in the real estate management scenario.

[0025] See Figure 1 , which is a schematic flowchart of a fast vectorization method for real estate surveying and mapping drawings based on deep learning provided by the embodiments of the present invention. As Figure 1 and Figure 2 shown, the method mainly includes the following steps: Step 1, divide the real estate surveying and mapping drawing data to obtain a table area and a main graphic area; In specific implementation, the specific process of dividing the real estate surveying and mapping map can be as follows: 1.1 Image preprocessing and contour extraction First, by analyzing the data characteristics of the real estate surveying and mapping map, it is found that its width is always greater than its height. Therefore, calculate the row-column ratio of the image , when is greater than 1, that is, when the image height is greater than the width, rotate the image 90 degrees clockwise to ensure that the image always maintains a horizontal direction: ; Among them, , are the number of rows and columns of the image respectively.

[0026] Furthermore, apply a Gaussian filter to denoise the image, and on this basis, use the Canny edge detection operator to extract the image edges. The calculation formula is as follows: ; ; ; Among them, is the Gaussian kernel formula, is the pixel coordinate, is the standard deviation of the Gaussian kernel, which controls the smoothness; is the gradient intensity, is the horizontal gradient, is the vertical gradient; is the gradient direction. Along the gradient direction, retain the local maximum value, suppress non-maximum points, and refine the edges.

[0027] To further optimize the detection results, morphological operations are used to eliminate small noises: ; Among them, is the structure element; represents the dilation operation; represents the erosion operation; is the processed binary image.

[0028] For each detected contour, calculate its area and perimeter. Set the area threshold , and only retain the contours with an area greater than the threshold: ; Among them, is the area of the contour.

[0029] 1.2 Region division Based on the geometric characteristics of the rectangle, perform adaptive region division. First, calculate the longitudinal position ratio of the rectangle , by analyzing the real estate surveying and mapping drawings, the optimal segmentation threshold is selected , and this threshold can effectively distinguish the table area and the main graphic area. Specifically, when the ordinate , this area is divided into the table area, which contains attribute information such as measurement numbers; when the ordinate , this area is divided into the main graphic area, which contains the geometric shape information of the house.

[0030] ; Among them, is the ordinate of the center point of the rectangle, is the total height of the image.

[0031] Step 2, use high-precision text recognition technology to extract the coordinate number data of the main shape from the table area; In specific implementation, a high-precision text recognition system based on PaddleOCR can be adopted. By loading the PP-OCRv4 pre-trained model, a complete recognition framework is constructed in combination with the text direction classification model and the Chinese recognition model. The system will automatically correct the text direction during the processing and perform intelligent preprocessing on the input image to ensure that the image always maintains the best recognition posture. In particular, when it is detected that the height of the image is greater than the width, the system will automatically perform a 90-degree clockwise rotation to ensure the accuracy of text recognition.

[0032] During the recognition process, the system adopts a multi-level information extraction strategy. First, by traversing all recognized texts and applying specific filtering rules, the system preferentially recognizes the text content without Chinese characters and selects the longest string as the candidate value of the measurement number. Subsequently, strict regular expressions are used to verify the format of the candidate measurement number to ensure that it conforms to the standard specifications. At the same time, the system can also intelligently extract other key information, such as determining the number of floors by recognizing pure digital texts and screening the largest value not greater than 99, and recognizing integers with negative signs to determine the hierarchical information. When some information cannot be recognized, the system will automatically mark the corresponding field as "N" to ensure the integrity of the data.

[0033] Step 3, use the trained UNET++ deep learning model to perform automatic vectorization processing on the main graphic area to obtain the vectorization result; In specific implementation, the embodiment of the present invention adopts the UNet++ network architecture to realize the automatic vectorization of the main graphic. For any node in the network, its feature map calculation expression is: ; Among them, represents the dense convolutional block operation of the i-th node in the l-th layer.

[0034] The feature map after convolution operation can be expressed as: ; where is the weight of the convolution kernel at position , is the input feature map, is the bias term, is the half-width of the convolution kernel.

[0035] During the network training process, the cross-entropy loss function is adopted: ; where is the true label, is the probability predicted by the model, is the number of samples.

[0036] After training the UNET++ deep learning model, for the input image , its feature extraction process is achieved through multi-scale convolution operations: ; where is the feature map of the th layer, is the convolution kernel weight of the th layer, is the bias term of the th layer, is the non-linear activation function.

[0037] Step 4: By establishing the corresponding relationship of the coordinate number data, the vectorized result is precisely matched with the building line graph of the house. Based on the actual spatial dimension information in the building line graph of the house, the vectorized result is standardized by applying the affine transformation technology to obtain the final vector data with actual geographical coordinates.

[0038] Specifically, the process of measurement number matching and affine transformation can be as follows: 4.1 Measurement number matching: Match the measurement number data of the main shape extracted from the table area in the previous step with the measurement number information in the attribute table of the building line graph of the house. The building line graph of the house is vector data, and there is measurement number information in the data attribute table, which is the outer contour of the real house in the real world, and its size and direction are consistent with the real world. The main steps include: ① Construct the measurement number information extracted from the table area into an association table {measurement number ID, pixel coordinates (x, y)}; ② Extract the measurement number information from the attribute table of the building line graph of the house and construct it into an association table {Measurement number ID, geographical coordinates (lon, lat)}; ③ Using the measurement number ID as the connection key, perform a spatial join operation to generate a mapping relationship table M {measurement number ID, pixel coordinates (x, y), geographical coordinates (lon, lat)}. 4.2 Affine transformation calculation based on control points To achieve an accurate transformation from the pixel space to the actual geographical space, this method uses affine transformation technology. Select at least 4 evenly distributed corner points from the mapping relationship table in step 4.1 as control points; for the selected set of control points , construct an equation for solving the transformation matrix: ; where, are the pixel coordinates in the source image, are the coordinates in the target geographical coordinate system, are the affine transformation coefficients to be solved.

[0039] Use the least squares method to solve the system of equations to obtain the optimal affine transformation matrix T, and apply the affine transformation matrix T to all vertex coordinates of the preliminary vectorization result generated by the UNET++ model. The transformed vertex coordinates are converted from the pixel coordinate system to the actual geographical coordinate system.

[0040] The method for rapid vectorization of real estate surveying and mapping drawings based on deep learning provided in this embodiment realizes the full-automatic vectorization processing of real estate surveying and mapping drawings through the innovative integration of deep learning and character recognition technology. Compared with the semi-automatic method that relies on manual intervention in the prior art, it greatly improves the processing efficiency and accuracy; it applies an affine transformation method based on coordinate number matching to solve the problem of accurate conversion from pixel coordinates to actual geographical coordinates, overcomes the defect of relying on manual alignment in the prior art, and realizes the high-precision automatic alignment of the vectorization result with the actual outer contour of the house.

[0041] The method of the present invention will be further described below in combination with a specific embodiment. The specific implementation process of the present invention will be described using the real estate surveying and mapping drawing data and building line data of Area A: (1) In the embodiment, a real estate surveying and mapping drawing of a certain area in Area A is selected as the input data. First, preprocess the input image, including Gaussian filtering and Canny edge detection, extract the image edges and apply morphological operations to eliminate noise. Perform direction correction by calculating the row-column ratio of the image to ensure that the image always maintains a horizontal direction. Subsequently, perform regional division on the image based on rectangular geometric features and adaptive thresholds to separate the real estate surveying and mapping drawing into a main graphic area and a table area.

[0042] (2) The table area is processed using a high-precision text recognition system based on PaddleOCR. The system loads the PP-OCRv4 pre-trained model and constructs a complete recognition framework by combining the text direction classification model and the Chinese recognition model. Through a multi-level information extraction strategy, the system identifies key information such as measurement numbers and floor numbers. For the measurement number, the system preferentially identifies the text content without Chinese characters and validates the format using regular expressions to ensure compliance with the standard specifications.

[0043] (3) The trained UNet++ deep learning model is used to perform automatic vectorization processing on the main graphic area. The model extracts image features through multi-scale convolutional operations and is optimized using the cross-entropy loss function. During implementation, 180 real estate surveying and mapping maps of Area A are used as the dataset, with 150 for model training and 30 for validation. The model is applied to the preprocessed main graphic area to generate a preliminary vectorization result, which contains the geometric shape information of the house but lacks actual geographical coordinates.

[0044] (4) The measurement number information extracted in step (2) is matched with the measurement number data in the attribute table of the building line drawing of the house. A spatial join operation is performed using the measurement number ID as the connection key to establish the correspondence between the preliminary vectorization result and the actual geographical coordinates. Based on the correspondence, four evenly distributed corner points are selected as control points, and an affine transformation matrix is constructed and solved to convert the vectorization result from the pixel coordinate system to the actual geographical coordinate system, obtaining the final vectorization result as Figure 3 shown.

[0045] It should be understood that each part of the present invention can be implemented using hardware, software, firmware, or a combination thereof.

[0046] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A rapid vectorization method for real estate surveying and mapping drawings based on deep learning, characterized in that, Including: Step 1: Divide the real estate surveying and mapping data to obtain a table area and a main graphic area; Step 2: Use high-precision character recognition technology to extract the coordinate number data of the main shape from the table area; The specific steps of Step 2 include: Step 2.1: By traversing all the recognized texts in the table area and applying a preset filtering rule, the system preferentially recognizes the text content without Chinese characters, and selects the longest string as the candidate measurement number; Step 2.2: Use a preset regular expression to verify the format of the candidate measurement number; Step 2.3: Determine the number of floors by recognizing pure digital texts and screening the largest value not greater than 99, and determine the hierarchical information by recognizing integers with negative signs; Step 2.4: When some information cannot be recognized, the system will automatically mark the corresponding field as N, and comprehensively form the coordinate number data; Step 3: Use a trained UNET++ deep learning model to perform automatic vectorization processing on the main graphic area to obtain a vectorization result; Step 4: By establishing the corresponding relationship of the coordinate number data, accurately match the vectorization result with the building line graph of the house. Based on the actual spatial dimension information in the building line graph of the house, apply the affine transformation technology to standardize the vectorization result to obtain the final vector data with actual geographical coordinates; The specific steps of Step 4 include: Step 4.1: Match the measurement number of the attribute information in the table area with the measurement number in the attribute table of the building line graph of the house to generate a mapping relationship table; Step 4.2, select at least 4 evenly distributed corner points from the mapping relation table as control points, and for the selected set of control points , construct a transformation matrix solution equation: ； Among them, is the pixel coordinate in the source image, is the coordinate in the target geographic coordinate system, is the affine transformation coefficient to be solved; Step 4.3: Use the least squares method to solve the transformation matrix solution equation to obtain the optimal affine transformation matrix T, and apply the affine transformation matrix T to all vertex coordinates of the vectorization result generated by the UNET++ model. The transformed vertex coordinates are converted from the pixel coordinate system to the actual geographical coordinate system to obtain the final vector data with actual geographical coordinates.

2. The method according to claim 1, wherein The specific steps of Step 1 include: Step 1.1, calculate the row-column ratio of the images in the real estate surveying and mapping map data , when is greater than 1, rotate the image 90 degrees clockwise; Step 1.2: Apply a Gaussian filter to denoise the image, and then use the Canny edge detection operator to extract the contour of the image ； ； ； Among them, is the Gaussian kernel formula, is the pixel coordinate, is the standard deviation of the Gaussian kernel, controlling the smoothness degree, is the gradient intensity, is the horizontal gradient, is the vertical gradient, is the gradient direction. Along the gradient direction, local maxima are retained, non-maximum points are suppressed, and the edges are refined; Step 1.3: For each detected contour, calculate its area and perimeter, set an area threshold, and retain the contours with an area greater than the area threshold; Step 1.4, calculate the longitudinal position ratio of the contour, and based on this, select the optimal segmentation threshold by analyzing the data of the real estate surveying and mapping map , and based on the optimal segmentation threshold divide each contour with an area larger than the area threshold into a table area and a main graphic area.

3. The method according to claim 2, characterized in that, After Step 1.2, the method further includes: Using morphological operations to eliminate small noises in the extracted contours: ； Among them, is a structural element, represents dilation operation, represents erosion operation, is the processed binary image, is the extracted contour.

4. The method according to claim 3, wherein The specific steps of Step 1.4 include: When the height of the contour is reached, the contour is divided into table areas containing attribute information; When the height of the contour the contour is divided into a main graphic area, which contains the geometric shape information of the house, is the total height of the image.

5. The method according to claim 4, wherein The specific steps of Step 4.1 include: Step 4.1.1: Construct the measurement number information extracted from the table area into a first association table; Step 4.1.2: Extract the measurement number from the attribute table of the building line graph of the house and construct it into a second association table; Step 4.1.3: Based on the measurement number ID as the connection key, perform a spatial join operation to generate a mapping relationship table.

Images