An automated design method, device, computer equipment, and medium for corner bracing and anti-bracing structures within a foundation pit.
Through a three-stage design process and generative adversarial networks, the automated design of corner bracing and counterbracing structures within the foundation pit was achieved, solving the problems of low design efficiency and unstable results, and providing a fast and accurate design solution.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA CONSTR THIRD ENG BUREAU GRP CO LTD
- Filing Date
- 2026-01-26
- Publication Date
- 2026-05-26
AI Technical Summary
Existing design methods for corner bracing and cross bracing structures within foundation pits are inefficient, produce inconsistent design results, lack theoretical basis, rely on engineering experience, and are difficult to automate quickly and accurately.
A three-stage design process and generative adversarial network are adopted. By selecting images from the design drawings of the support structure in the foundation pit, performing semantic processing and image generation, a generative adversarial network model is built and trained to realize the automated design of the corner bracing and anti-bracing structure in the foundation pit.
It significantly improves design efficiency, enabling the rapid and accurate completion of preliminary designs for corner bracing and counterbracing structures within the foundation pit, reducing data requirements and human resource consumption, and ensuring the rationality and economy of the design results.
Smart Images
Figure CN122087909A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of smart structures and intelligent structural design technology, specifically to an automated design method, device, computer equipment, and medium for corner bracing and anti-bracing structures in foundation pits based on three-stage design and generative adversarial networks. Background Technology
[0002] The internal support structure of a foundation pit is one of the important measures in the foundation pit support system to ensure the overall stability of the foundation pit structure under the action of horizontal and vertical forces and to enhance the safety of structural construction. If the internal support structure of the foundation pit is damaged, the failure of the support system will lead to the interruption or delay of construction. Severe damage will lead to the instability and collapse of the foundation pit, and secondary disasters will occur in the surrounding buildings due to the impact of construction within the foundation pit. In recent years, construction accidents caused by foundation pit collapses have occurred frequently, resulting in huge economic losses and casualties.
[0003] Corner bracing structures are suitable for local reinforcement of irregular or polygonal foundation pits. By connecting multiple points, they enhance the rigidity of the support in weak areas and reduce deformation and damage in these areas. Counterbracing structures are suitable for large-span support of long or regular foundation pits. They effectively constrain the lateral displacement of the foundation pit retaining structure and avoid impacting surrounding buildings and roads. The combination of the two forms corner bracing and counterbracing structures, which are widely used because they can fully adapt to complex terrain and engineering needs and improve the flexibility and economy of support design.
[0004] Currently, the industry's design for corner bracing and anti-bracing structures within foundation pits includes: designers create an initial layout design for the corner bracing and anti-bracing structures based on the outer contour of the foundation pit and the location of the superstructure; then, based on the design results, structural modeling and finite element calculations are performed; and the structural layout and component dimensions are optimized and adjusted according to the calculation results. Through iterative optimization between finite element calculations and structural design, an economical and structurally reasonable design for corner bracing and anti-bracing structures within the foundation pit is obtained.
[0005] However, current design methods suffer from low efficiency, inconsistent quality of design results, and a lack of theoretical basis and physical significance in designs based on engineers' experience. To overcome the limitations of traditional design methods, there is an urgent need for a fast, efficient, and highly visualized automated design method for corner bracing and support structures within foundation pits, utilizing generative artificial intelligence. Summary of the Invention
[0006] This invention provides an automated design method, device, computer equipment, and medium for corner bracing and anti-bracing structures within a foundation pit.
[0007] In a first aspect, the present invention provides an automated design method for corner bracing and anti-bracing structures within a foundation pit, comprising: Select the design images of the corner bracing and antibracing structures inside the foundation pit from the design drawings of the internal support structure; Based on the design images of the corner bracing and counter-bracing structures within the foundation pit, images of each stage in the three-stage design process of the corner bracing and counter-bracing structures are constructed. Semantic processing is performed on the key structural information in the images at each stage to obtain the semantic images corresponding to each stage; The semantic images corresponding to each stage are decomposed into multiple single-support structure design images, and then the multiple single-support structure design images are combined and stitched together to form a multi-support structure design image. A generative adversarial network model is constructed. The generative adversarial network model is trained based on the design images of each single-support structure and each multi-support structure. The optimal hyperparameters are selected to obtain an automated design model for the corner bracing and counter-bracing structures inside the foundation pit. The automated design model for the corner bracing and counter-bracing structures inside the foundation pit is then used to perform automated design of the corner bracing and counter-bracing structures inside the foundation pit.
[0008] In some instances, the construction of images for each stage of the three-stage design process of the corner brace and anti-brace structure based on the design images of the corner brace and anti-brace structure within the foundation pit includes: Construct an initial stage image, the structural information of which includes the outer contour of the foundation pit and auxiliary lines used to divide the support area; Construct an intermediate stage image, the structural information of which includes the outer contour of the foundation pit, auxiliary lines for dividing the support area, and the void area inside the support. Construct a final stage image, which includes structural information such as the outer contour of the foundation pit, the center lines of the corner braces and anti-braces structural components, and the hollow areas inside the supports.
[0009] In some instances, the semantic processing of key structural information in images at each stage to obtain semantically encoded images corresponding to each stage includes: Semantic processing of key structural information in images at each stage is performed based on the Euclidean distance between different pixels in the three-dimensional pixel space to obtain semantic images corresponding to each stage.
[0010] In some instances, the generative adversarial network model consists of a generator and a discriminator. The generator consists of a global generator and several local detail enhancers, and the discriminator adopts a multi-pixel scale discriminator architecture.
[0011] In some instances, the process of training a generative adversarial network model based on the design images of each single-support structure and each multi-support structure, and selecting the optimal hyperparameters to obtain an automated design model for the corner bracing and anti-bracing structures within the foundation pit, includes: The single-support structure design images and the multi-support structure design images are mixed to obtain a mixed dataset, and the multi-support structure design images are used as the multi-support structure dataset. After pre-training the generative adversarial network model using the hybrid dataset, the generative adversarial network model is further enhanced using the multi-support structure dataset. The generative adversarial network model's generation capability is evaluated based on image generation quality assessment criteria, and the optimal model hyperparameters are selected based on the evaluation results.
[0012] In some instances, after automating the design of the corner bracing and bracing structures within the foundation pit using the automated design model for these structures, the method further includes: Gaussian blurring is performed on the generated image obtained from the automated design model of the corner brace and counterbracing structure in the foundation pit. The image after Gaussian blurring is converted into a grayscale image and binarized. The outline of the key structural information of the binarized image is drawn to obtain the outline of the corner brace and counterbracing structure components in the foundation pit. The local twisted lines in the outline of the corner brace and counterbracing structure components in the foundation pit are straightened.
[0013] In some instances, before semantically processing the key structural information in the images at each stage to obtain the semantically represented images for each stage, the method further includes: Remove non-critical structural information from the images at each stage.
[0014] Secondly, the present invention provides an automated design device for corner bracing and anti-bracing structures within a foundation pit, comprising: The image filtering module is used to filter out the design images of corner braces and anti-braces in the foundation pit from the design drawings of the support structure in the foundation pit. The design process construction module is used to construct images of each stage in the three-stage design process of the corner brace and counter-brace structure based on the design images of the corner brace and counter-brace structure in the foundation pit. The semanticization module is used to semantically process the key structural information in the images at each stage to obtain the semantic images corresponding to each stage. The training set construction module is used to decompose the semantic images corresponding to each stage into multiple single-support structure design images, and then combine and stitch the multiple single-support structure design images into a multi-support structure design image. The training module is used to build a generative adversarial network model. Based on the design images of each single-support structure and each multi-support structure, the generative adversarial network model is trained. The optimal hyperparameters are selected to obtain an automated design model for the corner bracing and counter-bracing structures inside the foundation pit. The automated design model for the corner bracing and counter-bracing structures inside the foundation pit is then used for the automated design of the corner bracing and counter-bracing structures inside the foundation pit.
[0015] Thirdly, the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.
[0016] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the methods described above.
[0017] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects: This invention decouples the complex design of corner bracing and counterbracing structures within foundation pits into a three-stage design method, reducing the difficulty of model training and the amount of data required for training, as well as the workload of data collection and processing. Leveraging the image generation capabilities of generative adversarial networks, it significantly improves design efficiency, enabling the preliminary design of corner bracing and counterbracing structures within foundation pits to be completed in a short time, thus achieving fast, accurate, and convenient automatic design of corner bracing and counterbracing structures within foundation pits. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0019] Figure 1 This is a schematic diagram of the method flow provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating the automated design method for corner bracing and counterbracing structures in foundation pits based on three-stage design and generative adversarial networks provided in this embodiment of the invention. Figure 3 This is a schematic diagram of the design of the corner brace and counterbrace structure in the foundation pit provided in the embodiment of the present invention, broken down into a single support structure design; Figure 4 This is a schematic diagram of the three-stage intelligent structural design provided in an embodiment of the present invention; Figure 5 This is a schematic diagram illustrating the semantic processing of key structural information in the design of corner braces and counterbraces within the foundation pit provided in this embodiment of the invention. Figure 6 This is a schematic diagram of the semantic single-support structure design provided in the embodiments of the present invention, which is disassembled and spliced into a multi-support structure design; Figure 7 This is a schematic diagram of image data rotation enhancement provided in an embodiment of the present invention; Figure 8 This is a schematic diagram of the model generator architecture provided in an embodiment of the present invention; Figure 9 This is a schematic diagram of the model discriminator architecture provided in an embodiment of the present invention; Figure 10 This is a schematic diagram of the model post-processing flow provided in an embodiment of the present invention; Figure 11 This is a schematic diagram of the device provided in an embodiment of the present invention; Figure 12 This is a schematic diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation
[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0021] In the following description, specific embodiments of the invention will be illustrated with reference to steps and symbols performed by one or more computers, unless otherwise stated. Therefore, these steps and operations will be referred to several times as being performed by a computer, and computer execution as referred to herein includes operations by a computer processing unit representing electronic signals of data in a structured format. This operation transforms the data or maintains it at a location in the computer's memory system, which can be reconfigured or otherwise alter the operation of the computer in a manner well known to those skilled in the art. The data structure maintained by the data is the physical location of the memory, which has specific characteristics defined by the data format. However, the principles of the invention described above are not intended to be limiting, and those skilled in the art will understand that many of the following steps and operations can also be implemented in hardware.
[0022] The terms "module" or "unit" as used herein can be considered as software objects executing on the computing system. Different components, modules, engines, and services described herein can be considered as implementations on the computing system. The apparatus and methods described herein are preferably implemented in software, but can also be implemented in hardware, both of which are within the scope of this invention.
[0023] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.
[0024] In this embodiment of the invention, an automated design method for corner bracing and counterbracing structures within a foundation pit is provided. This method, based on a three-stage design and a generative adversarial network (GAN), requires low data volume. The three-stage design is a step-by-step design method determined by the characteristics of the corner bracing and counterbracing structures within the foundation pit, thereby reducing the amount of data required to train the GAN. The GAN is a deep learning-based generative model framework consisting of a generator and a discriminator, which generates data through a game-theoretic process. Figure 1 As shown, the method includes the following steps: S101: Select the design images of corner braces and antibraces in the foundation pit from the design drawings of the supporting structure in the foundation pit; S102: Constructing images of each stage in the three-stage design process of corner bracing and bracing structures based on the design images of corner bracing and bracing structures within the foundation pit; S103: Semantize the key structural information in the images at each stage to obtain the semantic images corresponding to each stage; S104: Decompose the semantic images corresponding to each stage into multiple single-support structure design images, and then combine and stitch the multiple single-support structure design images into a multi-support structure design image. S105: Build a generative adversarial network model. Train the generative adversarial network model based on the design images of each single support structure and each multi-support structure. Select the best hyperparameters to obtain an automated design model for the corner bracing and counter-bracing structures inside the foundation pit. Then, use the automated design model of the corner bracing and counter-bracing structures inside the foundation pit to perform automated design of the corner bracing and counter-bracing structures inside the foundation pit.
[0025] The embodiments of the present invention, through the above-described technical solutions, can assist structural designers in completing the preliminary design of corner bracing and anti-bracing structures, and quickly obtain the initial draft of the layout of corner bracing and anti-bracing structural components. Compared with traditional design methods, it has the characteristics of low cost, fast design speed, and reasonable design results, while significantly reducing the amount of original dataset required, which is more in line with actual engineering needs.
[0026] In this embodiment of the invention, the three-stage design process decomposes the traditional design process of corner bracing and anti-bracing structures within a foundation pit into three stages. By decoupling the complex design of corner bracing and anti-bracing structures, the difficulty of model training and the amount of data required for training are reduced, enabling fast, accurate, and convenient automatic design of corner bracing and anti-bracing structures within a foundation pit. Specifically, the first stage refers to the initial design stage, where the semantic image contains structural information including the outer contour of the foundation pit and manually added auxiliary lines for dividing the support area; the second stage refers to the intermediate design stage, where the semantic image contains structural information including the outer contour of the foundation pit, auxiliary lines for dividing the support area, and the hollow area within the support; the third stage refers to the final design stage, where the semantic image contains structural information including the outer contour of the foundation pit, the center lines of the corner bracing and anti-bracing structural components, and the hollow area within the support.
[0027] In this embodiment of the invention, by collecting design images of the internal support structure of the foundation pit applied to actual projects, DWG format structural design drawings of corner braces and anti-braces are selected, and an original DWG format structural design drawing dataset is constructed. The internal support structure forms included in the design images of the internal support structure of the foundation pit are: corner brace and anti-braces structure, horizontal corner brace structure, horizontal anti-braces structure, orthogonal rod system support structure, and truss side brace structure.
[0028] In this embodiment of the invention, based on the design images of the corner bracing and counterbracing structures in the foundation pit, the design process and structural design characteristics of the corner bracing and counterbracing structures in the foundation pit are summarized, and an intelligent three-stage design process for the corner bracing and counterbracing structures is constructed.
[0029] In this embodiment of the invention, the non-structural information in the design image of the corner bracing and counterbracing structure inside the foundation pit includes the layout information of the pile structure components, the layout information of the slope line, the structural annotation information, and the outer contour information of the corner bracing and counterbracing components; the key structural information in the design image of the corner bracing and counterbracing structure inside the foundation pit includes the layout information of the center lines of the corner bracing and counterbracing structure components, the layout information of the void area inside the support, and the outer contour information of the foundation pit.
[0030] In this embodiment of the invention, the image generation quality evaluation criteria are the rationality of the arrangement of the center lines of the corner braces and counter-braces in the foundation pit, the rationality of the arrangement of the void areas in the supports, and the degree of image line distortion.
[0031] In this embodiment of the invention, after obtaining the automated design model of the corner bracing and anti-bracing structure within the foundation pit, the method further includes a verification step: collecting more design images of the corner bracing and anti-bracing structure within the foundation pit, extracting the outer contour of the foundation pit and adding appropriate auxiliary lines, semantically processing the data, constructing a test set as input data for the automated design model of the corner bracing and anti-bracing structure within the foundation pit, generating the void area within the support in the intermediate design stage, and generating the center lines of the corner bracing and anti-bracing structure components in the final design stage based on the outer contour of the foundation pit, a small number of auxiliary lines, and the void area within the support in the intermediate design stage. Subsequently, based on FID, Structural economic indicators and structural mechanical performance indicators are used to quantitatively evaluate the effectiveness of the model and the rationality of the design results generated by the model.
[0032] In this embodiment of the invention, the generative adversarial network consists of a generator and a discriminator, which compete against each other and improve the quality of the generated data through iterative training. The optimization process of the model can be expressed by the following formula:
[0033] Where D represents the discriminator of the model, G represents the generator of the model, x represents the target data randomly sampled from the target dataset, y represents the input data sampled from the input dataset, and z represents the random noise data sampled from the random noise distribution. During the adversarial training of the model, the generator aims to make D(G(z|y)) approach 1, confusing the discriminator's classification judgment of the target data and the generated data; the discriminator aims to make D(x|y) approach 1 and D(G(z|y)) approach 0, so as to make accurate classification judgment of real data and generated data.
[0034] The model generator consists of a global generator and three local detail enhancers. The global generator consists of six convolutional blocks, nine residual blocks, and five deconvolutional blocks. The local detail enhancers consist of three convolutional blocks, three residual blocks, and two deconvolutional blocks. When the model generates an image, the global generator outputs the overall architecture of the generated image based on random noise z and input data y. This overall architecture is used as input, and the three local detail enhancers perform detail processing on the generated image.
[0035] The model discriminator adopts a multi-pixel scale discriminator architecture, which classifies and distinguishes the target image data from the generated image data at three different pixel scales: the original pixel scale - 1024×1024 pixels, the 1 / 2 pixel scale - 512×512 pixels, and the 1 / 4 pixel scale - 256×256 pixels. The discrimination results of the three different pixel scales are summarized and the final discrimination result is output.
[0036] In this embodiment of the invention, the post-processing step includes: Gaussian blurring is performed on the generated image obtained from the automated design model of the corner bracing and counterbracing structure in the foundation pit. The image after Gaussian blurring is converted into a grayscale image and binarized. The outline of the key structural information of the binarized image is delineated to obtain the outline of the corner bracing and counterbracing structure components in the foundation pit. The local twisted lines in the outline of the corner bracing and counterbracing structure components in the foundation pit are straightened.
[0037] In another embodiment of the present invention, an automated design method for corner bracing and anti-bracing structures within a foundation pit is provided, according to as follows: Figure 2 The illustrated process automates the design of corner bracing and anti-bracing structures within a foundation pit based on a three-stage design and a generative adversarial network. The design is rapid and accurate, significantly reducing the workload of structural designers and effectively avoiding design errors caused by hardware, resource consumption, and human factors. The process includes the following steps: Step 1: Collect DWG format structural design drawings of corner bracing and bracing structures, horizontal corner bracing structures, horizontal bracing structures, orthogonal rod support structures, and truss side bracing structures used in actual projects. Based on the flexibility of the arrangement of corner bracing and bracing structures and the economy of structural cost, select the design drawings of corner bracing and bracing structures and construct the original DWG format structural design drawing dataset.
[0038] Step 2: Based on the design process and characteristics of the corner bracing and counter-bracing structures within the foundation pit, the structural design characteristics are determined to be that the design consists of a support component arrangement area and a support cavity area. The support component arrangement area surrounds the support cavity area, and the design of the corner bracing and counter-bracing structures within the foundation pit can be decomposed into a single-bracing structure design consisting of multiple support cavity areas and support component arrangement areas surrounding the cavity areas. The three-stage intelligent structural design process is determined: using a generative adversarial network, based on the outer contour of the foundation pit and a few auxiliary lines in the initial design stage, the support cavity area in the intermediate design stage is generated; using a generative adversarial network, based on the outer contour of the foundation pit, a few auxiliary lines, and the support cavity area in the intermediate design stage, the center lines of the corner bracing and counter-bracing structural components in the final design stage are generated. Figure 3 This is a schematic diagram showing the breakdown of the corner bracing and counterbracing structure design within the foundation pit into a single-support structure design. Figure 4 This is a schematic diagram of the three-stage design for intelligent structure.
[0039] Step 3: Delete the layout information of the pile structure components, slope line layout information, structural annotation information, and outer contour information of the corner braces and braces in the foundation pit design to reduce data complexity and model learning difficulty; based on the Euclidean distance between different pixels in the three-dimensional pixel space, select an appropriate pixel distribution to perform semantic processing on the key structural information in the corner braces and braces design of the foundation pit: the layout information of the center lines of the corner braces and braces, the layout information of the void areas in the supports, and the outer contour information of the foundation pit. Figure 5 This is a schematic diagram illustrating the semantic processing of key structural information in the design of corner braces and counterbraces within the foundation pit.
[0040] The formula for calculating Euclidean distance is: ,in,( ) represents the coordinates of pixel 1 in the three-dimensional pixel space. ) represents the coordinates of pixel 2 in the three-dimensional pixel space.
[0041] Step 4: Based on the design characteristics of the corner bracing and counter-bracing structures within the foundation pit, the semantically defined corner bracing and counter-bracing structures are decomposed into multiple single-support structure designs. These single-support structure designs are then combined and spliced into a multi-support structure design. For the single-support structure design, a 100,000mm × 100,000mm square frame is created for printing, with a print pixel size of 1024 × 1024, meaning each pixel represents a 97.7mm × 97.7mm square. The print image format is PNG. For the multi-support structure design, a 200,000mm × 200,000mm square frame is created for printing, with a print pixel size of 1024 × 1024, meaning each pixel represents a 195.3mm × 195.3mm square. The print image format is PNG. The two types of image data are rotated by 90°, 180°, and 270° respectively for data augmentation. Figure 6 This is a schematic diagram illustrating the disassembly and assembly of a semantically simplified single-support structure into a multi-support structure design. Figure 7 This is a schematic diagram illustrating rotational data augmentation of image data.
[0042] Mix multi-support structure design data and single-support structure design data to create a hybrid dataset; collect multi-support structure design data to create a multi-support structure dataset, and collect single-support structure design data to create a single-support structure dataset.
[0043] Step 5: Construct a generative adversarial network (GAN), consisting of a generator and a discriminator. The generator comprises a global generator and three local detail enhancers. The global generator consists of six convolutional blocks, nine residual blocks, and five deconvolutional blocks, while the local detail enhancers consist of three convolutional blocks, three residual blocks, and two deconvolutional blocks. The discriminator employs a multi-pixel scale discriminator architecture, classifying and distinguishing between target image data and generated image data at three different pixel scales: the original image pixel scale (1024×1024), the half-pixel scale (512×512), and the quarter-pixel scale (256×256).
[0044] The training strategy employs a two-stage training method. In the first stage, the model is pre-trained using a mixed dataset, and in the second stage, the model is reinforced using a multi-support structure dataset. The model training parameters are shown in Table 1 below.
[0045] Table 1
[0046] Based on the image generation quality evaluation criteria, the model generation capability is judged by the rationality of the centerline arrangement of the corner bracing and counter-bracing structural components in the foundation pit, the rationality of the arrangement of the void area in the support, and the degree of image line distortion. Based on the evaluation results, the best model hyperparameters are selected for model training. Figure 8 This is a schematic diagram of the model generator architecture. Figure 9 This is a schematic diagram of the model discriminator architecture.
[0047] Step 6: Using OpenCV, Gaussian blur is applied to the image to reduce noise. To highlight the center lines of the corner braces and anti-braces in the image, the image is converted to grayscale and binarized. The Canny edge detection algorithm is used to accurately delineate the contours of key structural information, resulting in the contour images of the corner braces and anti-braces within the foundation pit. Combining this with the image generation module in StableDiffusion, positive prompts are added: "Repair the image by reconnecting broken lines based on the overall structure, focusing on the tail ends of unconnected lines. Remove any short lines to enhance visual coherence and clarity." The iteration step is set to 100 steps, and the redrawing amplitude is 0.5. Lines with local distortion are straightened to improve the visual expressiveness of the generated image data. Figure 10 This is a schematic diagram of the model post-processing workflow.
[0048] Step 7: Collect more design drawings of the corner bracing and counter-bracing structures within the foundation pit. Extract the semantically encoded outer contour of the foundation pit and add appropriate auxiliary lines as model input data to generate semantically encoded design drawings of the corner bracing and counter-bracing structures within the foundation pit. Summarize the generated image data and target image data, based on FID (Frechet Inception Distance). Image similarity is evaluated using the (Intersection over Union) evaluation metric. Semantically represented corner braces and parallel braces are extracted from the design drawings using OpenCV. The Canny algorithm is used to extract the component edge contours, and the pixel lengths of the corner braces and parallel braces are obtained based on these contours. These pixel lengths indirectly reflect the economic indicators of the structural design. Based on the design results of the corner braces and parallel braces within the foundation pit in the generated and target images, structural modeling is performed using the "Tongji Qimingxing" structural calculation software. The same geological environment and loads are set, and the mechanical performance and design rationality of the generated model are evaluated by comparing the internal forces of the components in both models. This is achieved by summarizing FID (Functional Identification Number) and other relevant metrics. The effectiveness of the automated design model for the corner bracing and counter-bracing structures within the foundation pit was verified by economic indicators and structural mechanical performance.
[0049] Among them, the FID (Frechet Inception Distance) evaluation index is an objective evaluation index widely used in the quality assessment of generated images. Through the pre-trained Inception network, high-dimensional features of generated image data and target image data are extracted. By Gaussian fitting of the high-dimensional features, the mean vector and covariance matrix of generated image data and target image data are obtained. Based on this, the similarity of the high-dimensional features of the images is evaluated. The (Intersection over Union) evaluation metric assesses the model's generative capability by calculating the degree of overlap between the generated and target image data; image similarity evaluation uses FID (Frame Identifier), ... The formula for calculating the evaluation indicators is:
[0050]
[0051] in, , Let be the mean vector of the target image data and the generated image data. , Generate the covariance matrix of the target image data. , This refers to the image area occupied by internal void regions and support regions in the target image data. , The area occupied by internal void regions and support regions in the image data generated for the model.
[0052] In step 7, the visualization of the design results is improved by using post-processing techniques that combine OpenCV and Stable Diffusion.
[0053] Step 8: Load the automated design model of the corner bracing and counterbracing structure inside the foundation pit into the computer equipment, and carry out the automated design of the corner bracing and counterbracing structure inside the foundation pit in actual engineering design practice.
[0054] The embodiments of the present invention, through the above technical solutions, have the advantages of being fast, accurate, convenient, and stable. Based on the small amount of data required for training, it can generate the initial design draft of the corner bracing and counter-bracing structure in the foundation pit in a short time. While ensuring the rationality of the design, it also reduces the investment cost of hardware equipment and the reliance on human resources, making the design work more economical and efficient.
[0055] In another embodiment of the present invention, to facilitate better implementation of the method provided in the embodiments of the present invention, an apparatus based on the above method is also provided. The meanings of the terms used are the same as in the above method, and specific implementation details can be found in the description of the method embodiments.
[0056] Please see Figure 11 , Figure 11 This is a schematic diagram of the structure of the device provided in an embodiment of the present invention, wherein the device may include an image filtering module 1101, a design flow construction module 1102, a semantic processing module 1103, a training set construction module 1104, and a training module 1105, wherein: Image filtering module 1101 is used to filter out the design images of corner braces and anti-braces in the foundation pit from the design drawings of the support structure in the foundation pit. Design process construction module 1102 is used to construct images of each stage in the three-stage design process of corner bracing and bracing structure based on the design images of corner bracing and bracing structure in the foundation pit. Semanticization module 1103 is used to semanticize the key structural information in the images at each stage to obtain the semantic images corresponding to each stage; The training set construction module 1104 is used to decompose the semantic images corresponding to each stage into multiple single-support structure design images, and then combine and stitch the multiple single-support structure design images into a multi-support structure design image. Training module 1105 is used to build a generative adversarial network model. The generative adversarial network model is trained based on the design images of each single support structure and each multi-support structure. The optimal hyperparameters are selected to obtain an automated design model for the corner bracing and counter-bracing structures inside the foundation pit. The automated design model of the corner bracing and counter-bracing structures inside the foundation pit is used to perform automated design of the corner bracing and counter-bracing structures inside the foundation pit.
[0057] The specific implementation methods of each module can be referred to the description of the above method embodiments, and the embodiments of the present invention will not be repeated.
[0058] In another embodiment of the present invention, a computer device is also provided, such as... Figure 12As shown, it illustrates a structural schematic diagram of a computer device involved in an embodiment of the present invention, specifically: The computer device may include components such as a processor 1201 with one or more processing cores, a memory 1202 with one or more computer-readable storage media, a power supply 1203, and an input unit 1204. Those skilled in the art will understand that... Figure 12 The computer device structure shown does not constitute a limitation on the computer device and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein: The processor 1201 is the control center of the computer device. It connects various parts of the computer device via various interfaces and lines. By running or executing software programs and / or modules stored in the memory 1202, and by calling data stored in the memory 1202, it performs various functions of the computer device and processes data, thereby providing overall monitoring of the computer device. Optionally, the processor 1201 may include one or more processing cores; preferably, the processor 1201 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operation of the storage medium, user interface, and application programs, while the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 1201.
[0059] The memory 1202 can be used to store software programs and modules. The processor 1201 executes various functional applications and data processing by running the software programs and modules stored in the memory 1202. The memory 1202 may mainly include a program storage area and a data storage area. The program storage area may store application programs required for operating the storage medium and at least one function (such as sound playback function, image playback function, etc.); the data storage area may store data created according to the use of the computer device. In addition, the memory 1202 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 1202 may also include a controller to provide the processor 1201 with access to the memory 1202.
[0060] The computer device also includes a power supply 1203 that supplies power to various components. Preferably, the power supply 1203 can be logically connected to the processor 1201 through a power management storage medium, thereby enabling functions such as charging, discharging, and power consumption management through the power management storage medium. The power supply 1203 may also include one or more DC or AC power supplies, recharge storage media, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.
[0061] The computer device may also include an input unit 1204, which can be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.
[0062] Although not shown, the computer device may also include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 1201 in the computer device loads the executable files corresponding to the processes of one or more applications into the memory 1202 according to the following instructions, and the processor 1201 runs the applications stored in the memory 1202, thereby implementing the steps in the above method embodiment.
[0063] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.
[0064] Therefore, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, the computer program being loaded by a processor to execute the steps of any method provided in the embodiments of the present invention.
[0065] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.
[0066] The computer-readable storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.
[0067] Since the computer program stored in the computer-readable storage medium can execute the steps of any of the methods provided in the embodiments of the present invention, the beneficial effects that any of the methods provided in the embodiments of the present invention can achieve can be realized, as detailed in the preceding embodiments, and will not be repeated here.
[0068] The above provides a detailed description of the automated design method, device, computer equipment, and medium for corner bracing and counterbracing structures within a foundation pit provided by embodiments of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.
Claims
1. An automated design method for corner bracing and anti-bracing structures within a foundation pit, characterized in that, include: Select the design images of the corner bracing and antibracing structures inside the foundation pit from the design drawings of the internal support structure; Based on the design images of the corner bracing and counter-bracing structures within the foundation pit, images of each stage in the three-stage design process of the corner bracing and counter-bracing structures are constructed. Semantic processing is performed on the key structural information in the images at each stage to obtain the semantic images corresponding to each stage; The semantic images corresponding to each stage are decomposed into multiple single-support structure design images, and then the multiple single-support structure design images are combined and stitched together to form a multi-support structure design image. A generative adversarial network model is constructed. The generative adversarial network model is trained based on the design images of each single-support structure and each multi-support structure. The optimal hyperparameters are selected to obtain an automated design model for the corner bracing and counter-bracing structures inside the foundation pit. The automated design model for the corner bracing and counter-bracing structures inside the foundation pit is then used to perform automated design of the corner bracing and counter-bracing structures inside the foundation pit.
2. The method according to claim 1, characterized in that, The construction of images for each stage of the three-stage design process of the corner brace and counter-brace structure based on the design images of the corner brace and counter-brace structure within the foundation pit includes: Construct an initial stage image, the structural information of which includes the outer contour of the foundation pit and auxiliary lines used to divide the support area; Construct an intermediate stage image, the structural information of which includes the outer contour of the foundation pit, auxiliary lines for dividing the support area, and the void area inside the support. Construct a final stage image, which includes structural information such as the outer contour of the foundation pit, the center lines of the corner braces and anti-braces structural components, and the hollow areas inside the supports.
3. The method according to claim 2, characterized in that, The semantic processing of key structural information in the images at each stage to obtain semantic images corresponding to each stage includes: Semantic processing of key structural information in images at each stage is performed based on the Euclidean distance between different pixels in the three-dimensional pixel space to obtain semantic images corresponding to each stage.
4. The method according to any one of claims 1 to 3, characterized in that, The generative adversarial network model consists of a generator and a discriminator. The generator consists of a global generator and several local detail enhancers, and the discriminator adopts a multi-pixel scale discriminator architecture.
5. The method according to claim 4, characterized in that, The process involves training a generative adversarial network model based on the design images of each single-support structure and each multi-support structure, and selecting the optimal hyperparameters to obtain an automated design model for the corner bracing and anti-bracing structures within the foundation pit, including: The single-support structure design images and the multi-support structure design images are mixed to obtain a mixed dataset, and the multi-support structure design images are used as the multi-support structure dataset. After pre-training the generative adversarial network model using the hybrid dataset, the generative adversarial network model is further enhanced using the multi-support structure dataset. The generative adversarial network model's generation capability is evaluated based on image generation quality assessment criteria, and the optimal model hyperparameters are selected based on the evaluation results.
6. The method according to claim 1, characterized in that, After automating the design of the corner bracing and counter-bracing structure within the foundation pit using the automated design model of the foundation pit corner bracing and counter-bracing structure, the method further includes: Gaussian blurring is performed on the generated image obtained from the automated design model of the corner brace and counterbracing structure in the foundation pit. The image after Gaussian blurring is converted into a grayscale image and binarized. The outline of the key structural information of the binarized image is drawn to obtain the outline of the corner brace and counterbracing structure components in the foundation pit. The local twisted lines in the outline of the corner brace and counterbracing structure components in the foundation pit are straightened.
7. The method according to claim 1, characterized in that, Before semantically processing the key structural information in the images at each stage to obtain the semantically represented images for each stage, the method further includes: Remove non-critical structural information from the images at each stage.
8. An automated design device for corner bracing and counter-bracing structures within a foundation pit, characterized in that, include: The image filtering module is used to filter out the design images of corner braces and anti-braces in the foundation pit from the design drawings of the support structure in the foundation pit. The design process construction module is used to construct images of each stage in the three-stage design process of the corner brace and counter-brace structure based on the design images of the corner brace and counter-brace structure in the foundation pit. The semanticization module is used to semantically process the key structural information in the images at each stage to obtain the semantic images corresponding to each stage. The training set construction module is used to decompose the semantic images corresponding to each stage into multiple single-support structure design images, and then combine and stitch the multiple single-support structure design images into a multi-support structure design image. The training module is used to build a generative adversarial network model. Based on the design images of each single-support structure and each multi-support structure, the generative adversarial network model is trained. The optimal hyperparameters are selected to obtain an automated design model for the corner bracing and counter-bracing structures inside the foundation pit. The automated design model for the corner bracing and counter-bracing structures inside the foundation pit is then used for the automated design of the corner bracing and counter-bracing structures inside the foundation pit.
9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the method according to any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 7.