Water treatment engineering effect picture generation method and system based on AI drawing software

Abstract

The invention discloses a water treatment engineering effect picture generation method and system based on AI drawing software, and relates to the technical field of intelligent drawing, and the key points of the technical scheme are as follows: obtaining basic image information of a water treatment engineering, and learning and training by taking the basic image information as a data set to obtain an AI drawing model; obtaining a plane layout diagram of the water treatment project, and identifying diagram information in the plane layout diagram; obtaining effect picture requirement information, and generating cue words in combination with the picture surface information; and inputting the cue word into the AI drawing model, and generating a final effect picture after setting related parameters of the effect picture. According to the method, the AI drawing software is fully utilized, a secondary development technical means is combined, direct interaction between a scheme designer and the software is achieved, the scheme designer can adjust the effect picture by himself / herself according to basic data such as an engineering design file and a scheme assumption of himself / herself, and efficient and rapid generation of the water treatment engineering effect picture is achieved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent drawing technology, and more specifically, to a method and system for generating water treatment engineering renderings based on AI drawing software. Background Art

[0002] The renderings of municipal water treatment projects can visually present the overall or local effects of the project after completion. They are an important tool for project scheme display, scheme comparison and communication among all parties. They are used in multiple stages of project planning, reporting, design, etc.

[0003] At present, the conventional method for generating renderings of municipal water treatment projects mainly includes the following steps: (1) constructing three-dimensional models of various entities and elements, such as buildings, structures, roads, walls, green spaces, etc., based on the plan layout of the municipal water treatment project; (2) performing rendering in three-dimensional rendering software, such as 3ds Max software, to obtain a preliminary rendering; (3) using two-dimensional image processing software, such as Photoshop, to further process and modify the sky, surrounding environment, etc. to obtain the final rendering.

[0004] However, in the process of 3D modeling and rendering, the existing technology needs to manually build the models of each monomer and then render them. Therefore, the time for generating conventional renderings is generally calculated in days. If it involves a particularly complex modeling process, the time will be longer. The process of modifying the renderings will also involve model modification and re-rendering. In addition, since the establishment and rendering of 3D models require learning and accumulation of experience, most engineering renderings are produced by professional rendering companies, but they do not have enough understanding of engineering floor plans and cannot understand the design intent well. There is a process of communication and coordination, and complicated modifications will occur in practice. In addition, traditional 3D renderings are charged by the sheet, and according to the complexity of the model, a single sheet is generally around 1,000 yuan, which is relatively costly. Therefore, how to study and design a method and system for generating water treatment engineering renderings based on AI painting software that can overcome the above-mentioned defects is a problem that we urgently need to solve. Summary of the invention

[0005] In order to solve the deficiencies in the prior art, the purpose of the present invention is to provide a method and system for generating water treatment project renderings based on AI painting software, make full use of AI painting software, combine secondary development technical means, and realize direct interaction between scheme designers and software. Scheme designers can adjust the renderings by themselves according to their own engineering design documents, scheme ideas and other basic information, so as to realize efficient and rapid generation of water treatment project renderings.

[0006] The above technical objectives of the present invention are achieved through the following technical solutions:

[0007] In the first aspect, a method for generating a water treatment engineering rendering based on AI painting software is provided, comprising the following steps:

[0008] Obtain basic image information of water treatment projects, and use the basic image information as a data set to learn and train an AI painting model;

[0009] Obtain the plan layout of the water treatment project and identify the drawing information in the plan layout;

[0010] Obtain rendering requirements and generate prompt words based on drawing information;

[0011] Input the prompt words into the AI ​​painting model, set the relevant parameters of the rendering, and generate the final rendering.

[0012] Furthermore, the basic image information includes:

[0013] On-site images and renderings of water treatment projects;

[0014] Physical images of various equipment in water treatment projects, Revit 3D model renderings and effect diagrams of various equipment;

[0015] As well as real-life images and videos of water treatment project site surveys.

[0016] Furthermore, the training process of the AI ​​painting model is specifically as follows:

[0017] Generate adversarial networks to synthesize scene data that lacks samples, so as to enhance and synthesize data sets;

[0018] The enhanced and synthesized dataset is divided into training set, validation set and test set. The training set is used for model learning, the validation set is used to adjust model parameters and prevent model overfitting, and the test set is used to evaluate the generalization ability of the model.

[0019] Select the Adam optimizer to update the model parameters and use the learning rate decay strategy to adjust the learning rate;

[0020] Among them, the generator in the generative adversarial network generates new image samples, and the discriminator distinguishes whether the image is real or generated by the generator;

[0021] Also, the generator uses a deep convolutional neural network with a network depth of 6 layers. Each layer uses the ReLU activation function. The last layer uses the tanh function to output image data. Batch normalization is used after each convolution layer.

[0022] The discriminator uses a deep convolutional neural network with a network depth of 5 layers. The LeakyReLU activation function is used to increase nonlinearity. The last layer uses the sigmoid activation function to judge the authenticity of the image.

[0023] Furthermore, the process of obtaining the plan layout of the water treatment project and identifying the drawing information in the plan layout is specifically as follows:

[0024] Obtain water treatment plan and layer information: Use CAD secondary development tools and programming languages ​​to obtain water treatment plan, legends, and the expression content of each layer;

[0025] Identify the project redline boundary and calculate the area: In the layer information corresponding to the redline, use the Polyline object to find the polygon representing the redline boundary and calculate the area through the Area property;

[0026] Identify the position and size of each layer: Get the position and size through its Position and ScaleFactors properties. The layers include structures, buildings, red lines, walls, roads, and green areas.

[0027] Identify characteristic information: Use DBText and MText functions to read text information in the text descriptions and annotations in the drawings to identify the characteristic information of each structure, building, road and green space.

[0028] Furthermore, the rendering requirement information includes: the scope of the rendering display, the viewing angle of the rendering, the surrounding environment of the factory, the names of structures and buildings, the names of common equipment included in the structures and buildings, and the requirements and characteristics of the rendering style.

[0029] Furthermore, the process of obtaining the rendering requirement information and generating prompt words in combination with the drawing information is specifically as follows:

[0030] Perform calculation and analysis on the scope and viewing angle in the rendering requirement information and the names of structures and buildings in the same drawing information to obtain detailed equipment information included in the rendering;

[0031] Based on the equipment details contained in the renderings, the rendering requirements information and the records in the drawing information, AI tools are used to generate prompt words.

[0032] Furthermore, the related parameters include effect image size, file format, resolution and number of iterations.

[0033] Secondly, a water treatment engineering rendering generation system based on AI painting software is provided, including:

[0034] The model training module is used to obtain basic image information of water treatment projects and use the basic image information as a data set to learn and train an AI painting model;

[0035] An information recognition module is used to obtain the plan layout of the water treatment project and recognize the drawing information in the plan layout;

[0036] A prompt word generation module is used to obtain the rendering requirement information and generate prompt words in combination with the drawing information;

[0037] The rendering generation module is used to input the prompt words into the AI ​​painting model, set the relevant parameters of the rendering, and then generate the final rendering.

[0038] In a third aspect, a computer terminal is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a method for generating a water treatment project rendering based on AI painting software as described in any one of the first aspects is implemented.

[0039] In a fourth aspect, a computer-readable medium is provided, on which a computer program is stored, and the computer program is executed by a processor to implement a method for generating a water treatment project rendering based on AI painting software as described in any one of the first aspects.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] 1. The method for generating water treatment engineering renderings based on AI painting software provided by the present invention makes full use of AI painting software and combines secondary development technical means to realize direct interaction between scheme designers and software. Scheme designers can adjust the renderings by themselves according to their own engineering design documents, scheme ideas and other basic data, so as to realize efficient and rapid generation of water treatment engineering renderings;

[0042] 2. The present invention does not require the three-dimensional modeling stage, and can quickly establish the renderings of the municipal water treatment project. The engineering personnel can operate directly, which further saves time and can achieve rapid response of the renderings;

[0043] 3. After the AI ​​painting model training is completed, the present invention can generate renderings of more types of projects at a lower cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:

[0045] Figure 1 is a flow chart of Embodiment 1 of the present invention;

[0046] Figure 2 It is a system block diagram in Example 2 of the present invention. DETAILED DESCRIPTION

[0047] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0048] Example 1: A method for generating a water treatment project rendering based on AI painting software, such as Figure 1 As shown, the following steps are included:

[0049] S1: Obtain basic image information of water treatment projects, and use the basic image information as a data set to learn and train an AI painting model;

[0050] S2: Obtaining a plan layout of a water treatment project and identifying drawing information in the plan layout;

[0051] S3: Obtaining rendering requirement information and generating prompt words in combination with drawing information;

[0052] S4: Input the prompt words into the AI ​​painting model, set the relevant parameters of the rendering, and generate the final rendering.

[0053] In step S1, the basic image information includes but is not limited to: on-site images and renderings of the water treatment project; physical images of various equipment in the water treatment project, Revit three-dimensional model renderings and renderings of various equipment; and real-life images and videos of the on-site survey of the water treatment project.

[0054] In step S1, the model training process is completed using machine learning and deep learning frameworks, such as TensorFlow, PyTorch, sklearn, etc.

[0055] 1. Dataset enhancement and synthesis: Generative adversarial networks (GANs) are used to synthesize scene data that lack samples. First, the architecture of GANs is defined, where the generator is responsible for generating new image samples and the discriminator is responsible for distinguishing whether the image is real or generated by the generator.

[0056] The parameters are set as follows: The generator uses a deep convolutional neural network (DCNN) with a network depth of 6 layers, each layer uses the ReLU activation function, and the last layer uses the tanh function to output image data. Batch normalization is used after each convolution layer to stabilize the training process.

[0057] The discriminator also uses a deep convolutional neural network with a network depth of 5 layers. The LeakyReLU activation function is used to increase nonlinearity and avoid the gradient vanishing problem. The last layer uses the sigmoid activation function to judge the authenticity of the image.

[0058] The training parameters use the Adam optimizer, the learning rate is set to 0.0002, and the beta1 parameter is 0.5.

[0059] During the training process, in order to balance the training speed of the generator and the discriminator, the generator is updated twice for every update of the discriminator.

[0060] Data preprocessing: All input images are resized to a uniform size (e.g., 256x256 pixels) and normalized to the range of [-1, 1].

[0061] 2. Dataset division: The purpose of data set division is to ensure that the model can learn on diverse data, which is conducive to the model forming good generalization ability.

[0062] Dataset division strategy: standard training set, validation set, and test set division methods are used, where the training set is used for model learning, the validation set is used to adjust model parameters and prevent model overfitting, and the test set is used to evaluate the generalization ability of the model. The specific division ratio is: 70% for the training set, 15%-20% for the validation set, and 15%-10% for the test set.

[0063] Random sampling and stratified sampling: In order to ensure the diversity and representativeness of the data set, random sampling is used to divide the data to ensure that each subset can reflect the characteristics of the entire data set as much as possible. For specific types of data, such as scenes under different weather conditions, stratified sampling is used to ensure that the proportion of each type of data in each subset is consistent with the proportion in the entire data set.

[0064] Random Seed: Set a fixed random seed (such as 42) to ensure that the partitioning of the dataset is repeatable, so that the experimental results are comparable.

[0065] Category labels for stratified sampling: Category labels are set based on factors such as the complexity of the scene (such as simple, medium, and complex), weather conditions (such as sunny, rainy, and foggy), etc., to ensure that the distribution of data in each category in each subset is balanced.

[0066] Dividing the data set in this way can not only improve the efficiency of model training, but also enhance the robustness and prediction accuracy of the model when facing unknown data. At the same time, such detailed parameter settings and division strategy descriptions also increase the transparency and practicability of the technical solution, helping to demonstrate its scientificity and systematicity.

[0067] 3. Optimizer selection: Select the Adam optimizer to update the model parameters. It combines the advantages of the Momentum and RMSprop optimizers, can adaptively adjust the learning rate under different conditions, and is suitable for complex data sets and model structures.

[0068] Learning rate setting: The initial learning rate is set to 0.001, which is a common starting value and suitable for most cases. In order to deal with the gradient disappearance or explosion problem that may be encountered during training, a learning rate decay strategy is adopted. After every 10 epochs, the learning rate is halved until the end of training.

[0069] Batch Size: Set the batch size to 32. This value can achieve a good balance on most hardware, which can use the GPU for efficient calculations without causing memory overflow.

[0070] Epochs: Set the number of iterations to 100 epochs. According to experience, this number of iterations is enough for the model to converge in most cases. At the same time, set the Early Stopping mechanism. If the loss on the validation set does not decrease significantly within 20 consecutive epochs, the training will be terminated early to avoid overfitting.

[0071] 4. Model performance verification

[0072] Evaluation indicators: To comprehensively evaluate the performance of the model, multiple indicators are used for evaluation, including but not limited to accuracy, precision, recall, F1 score, and visual quality evaluation of the generated images. Visual quality can be evaluated manually or using pixel-level similarity indicators such as SSIM (Structural Similarity Index) and PSNR (Peak Signal-to-Noise Ratio).

[0073] Accuracy: measures the proportion of samples correctly predicted by the model to the total samples. It is one of the most intuitive performance indicators. The target accuracy of this model verification is above 90%.

[0074] Precision and recall: Precision measures the proportion of positive samples correctly identified by the model to the total number of positive samples identified by the model; recall measures the proportion of positive samples correctly identified by the model to all actual positive samples. These two indicators are particularly suitable for unbalanced data sets. The target precision and recall of this model verification reached 80%-90%.

[0075] F1 score: It is the harmonic mean of precision and recall, and can provide a single metric to evaluate the overall performance of the model. The F1 score of this model validation reached 80%-90%.

[0076] SSIM and PSNR: These two indicators are used to evaluate the quality of generated images. SSIM measures the structural similarity of images, and PSNR measures the peak signal-to-noise ratio of images. They can provide a quantitative measure of the similarity between the generated image and the target image. The SSIM target value for this model validation is over 0.85, and the PSNR target value is between 30dB and 50dB.

[0077] 5. Save the model

[0078] Model saving time: There are two forms: optimal model saving and regular saving.

[0079] Best Model Saved: The model with the best performance on the validation set is saved. This is because this model shows the best generalization ability on unseen data.

[0080] Save regularly: To prevent unexpected situations during training (such as hardware failure or power outage), you can set the model to be saved regularly after a certain number of epochs. In this way, even if the training is interrupted, you can resume the training from the most recent save point without having to start from scratch.

[0081] Use the ModelCheckpoint callback function to save the best model at the end of each epoch based on the performance on the validation set (such as validation loss).

[0082] The model is saved in HDF5 (.h5) format (a file format for storing large arrays, widely used to save model data with a hierarchical structure, suitable for frameworks such as TensorFlow and Keras).

[0083] In step S2, the plan layout of the water treatment project places structures, buildings, red lines, walls, roads, green areas, text descriptions, etc. into different layers, and places various annotation information such as names, boundary lines, and site elevations into different layers, and the legend of the plan layout corresponds to the layers.

[0084] Obtain water treatment plan and layer information: Use CAD secondary development tools and programming languages ​​such as C# or Python to obtain water treatment plan, legends, and the expression content of each layer.

[0085] The following takes C# programming language as an example, uses the Database class to access the DWG file database, performs data operations through Transaction transactions, and uses the LayerTable class to obtain all information on layers such as structures, buildings, red lines, walls, roads, and green areas.

[0086] Identify the project's redline boundary and calculate the area: In the layer information corresponding to the redline, use the Polyline object to find the polygon representing the redline boundary and calculate the area through the Area property. When dealing with polygon closure and area calculation, the tolerance value (Tolerance) can be set to 0.01 units to handle slight geometric inaccuracies.

[0087] Identify the position and size of each layer: Get the position and size through its Position and ScaleFactors properties. Layers include but are not limited to structures, buildings, red lines, walls, roads, and greenery.

[0088] Identify characteristic information: Use DBText and MText functions to read the text information in the text description and annotation in the drawing, and identify the characteristic information of each structure, building, road and greening, such as the name and field elevation of each building and structure; the elevation and slope of the road; the type of greening vegetation; the method of fence, etc.

[0089] The required information for the renderings includes: the scope of the rendering, such as the entire factory, part of the complex building, etc.; the perspective of the rendering, such as a bird's-eye view, a horizontal view, etc.; the surrounding environment of the factory, such as fields, cities, night, day, etc.; the names of structures and buildings, and the names of common equipment contained in the structures and buildings; the requirements and characteristics of the rendering style, facade style, theme, facade color preference, etc.

[0090] In step S3, the process of generating prompt words is specifically as follows: the scope and viewing angle in the rendering requirement information and the names of structures and buildings in the same drawing information are calculated and analyzed to obtain detailed equipment information contained in the rendering; based on the detailed equipment information contained in the rendering, the rendering requirement information and the records in the drawing information, prompt words are generated using AI tools.

[0091] When the current AI image generation model is applied to the field of municipal water treatment projects, the conventional operation is to input prompt words for generation. However, the equipment layout principles and logic of the water treatment project renderings are not embedded between the prompt words. Therefore, there is a problem that the water treatment project equipment that meets the requirements cannot be generated, and it is difficult to meet the professional requirements of the water treatment project renderings.

[0092] In addition, in the renderings generated by these models, the types and styles of equipment often do not meet the actual needs, and there are situations such as missing equipment and wrong equipment styles. In addition, the relative positions of the equipment, such as horizontal and vertical positions, do not conform to the conventions, such as common sense errors such as submersible sewage pumps being located on the ground.

[0093] To this end, the process of obtaining the detailed equipment information contained in the rendering is as follows:

[0094] ① Obtain the names of structures and buildings in the drawing information, and generate detailed equipment information of the corresponding structures / buildings, including equipment name, horizontal position, and vertical position, according to the correspondence between structures / buildings and equipment commonly used in water treatment projects, the commonly used horizontal layout position of equipment, and the commonly used vertical layout position of equipment, based on the structure direction and compass direction in the drawing information. For example, pretreatment tank-pump-southeast-bottom of the tank, pretreatment tank-gate-west-tank wall, pretreatment tank-grid-southwest-on the tank and in the tank; A2O biochemical tank-submersible mixer-southeast-bottom of the tank, A2O biochemical tank-aeration plate-west-bottom of the tank, A2O biochemical tank-submersible flow pusher-south-center of the tank; secondary sedimentation tank-scraper-southeast-center of the tank; high-efficiency sedimentation tank-mixer-west-center of the tank, high-efficiency sedimentation tank-scraper-south-bottom of the tank; denitrification deep bed filter-submersible sewage pump-north-bottom of the tank, denitrification deep bed filter-submersible sewage pump-north-bottom of the tank, Bed filter - mixer - southwest - in the pool; dosing room - storage tank - south - indoor floor, dosing room - metering pump - west - indoor floor, dosing room - dosing device - north - indoor floor; online monitoring room - instrument - south - indoor floor; dewatering room - filter press - south - indoor high floor floor, dewatering room - screw pump, dewatering room - sludge silo - south - outdoor ground; blower room - blower, blower room - crane - south - upper part of the room, blower room - axial flow fan - south - upper part of the wall, etc.

[0095] ② According to the scope and viewing angle in the rendering requirements, calculate the detailed equipment information contained in the structure / building in ① to obtain the detailed equipment information contained in the rendering. The calculation rules are:

[0096] a) According to the rendering range, filter the detailed information of the equipment contained in the structure / building. If the rendering range is the whole plant, it includes all the structures and buildings; if the rendering range is the exterior of a certain structure / building, keep the structure / building and the nearest 3-5 structures / buildings around it; if the rendering range is the interior of a certain structure / building, only keep the structure / building.

[0097] b) According to the perspective of the rendering, filter the detailed equipment information of the structures / buildings that have been filtered in the previous step. Keep the detailed equipment information of the structures / buildings that are consistent with the perspective of the rendering. For example, in the south bird's-eye view, keep the detailed equipment information of the south, southwest, and southeast sides; in the west pedestrian perspective, keep the detailed equipment information of the west, southwest, and northwest sides.

[0098] The present invention can utilize AI tools (such as Chat-GPT-4, Gemini, etc.) to generate prompt words according to the rendering requirements. The prompt words should be as clear and concise as possible, and should include all the rendering requirements, especially some keywords, such as "light blue facade", "modern sense", "bright tones", "water treatment equipment" and "green area".

[0099] In step S4, in the trained model, input the prompt word and set the relevant parameters of the rendering, such as the rendering size, file format, resolution, number of iterations, etc., to generate the rendering; ③ Check the generated rendering and obtain the final rendering by refining the prompt word or adjusting the parameters.

[0100] Embodiment 2: A system for generating water treatment engineering renderings based on AI painting software, the system is used to implement the method for generating water treatment engineering renderings based on AI painting software described in Embodiment 1, such as Figure 2 As shown, it includes a model training module, an information recognition module, a prompt word generation module and an effect diagram generation module.

[0101] Among them, the model training module is used to obtain the basic image information of the water treatment project, and use the basic image information as the data set to learn and train the AI ​​painting model; the information recognition module is used to obtain the plan layout of the water treatment project and identify the drawing information in the plan layout; the prompt word generation module is used to obtain the rendering requirement information and generate prompt words based on the drawing information; the rendering generation module is used to input the prompt words into the AI ​​painting model, set the relevant parameters of the rendering, and generate the final rendering.

[0102] The present invention also records a computer terminal, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method for generating a water treatment project rendering based on AI painting software as described in Example 1 is implemented.

[0103] The present invention also records a computer-readable medium on which a computer program is stored. The computer program is executed by a processor to implement the method for generating water treatment engineering renderings based on AI painting software as described in Example 1.

[0104] Working principle: The present invention uses existing data of water treatment projects, such as on-site photos, renderings, equipment photos, etc., to train an AI painting model that meets the requirements of municipal water treatment project renderings. According to the floor plan of the water treatment project, the information of the project-related monomers and elements is obtained, prompt words are generated using AI tools, and renderings are generated using the trained AI painting model. The present invention makes full use of AI painting software, combined with secondary development technical means, to achieve direct interaction between scheme designers and software. Scheme designers can adjust the renderings by themselves according to their own engineering design documents, scheme ideas and other basic data, and achieve efficient and rapid generation of water treatment project renderings.

[0105] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0106] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0107] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0108] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0109] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for generating water treatment engineering renderings based on AI painting software, characterized in that: The following steps are involved: Obtain basic image information of water treatment projects, and use the basic image information as a data set to learn and train an AI painting model; Obtain the plan layout of the water treatment project and identify the drawing information in the plan layout; Obtain rendering requirements and generate prompt words based on drawing information; Input the prompt words into the AI ​​painting model, set the relevant parameters of the rendering, and generate the final rendering.

2. The method for generating water treatment engineering renderings based on AI painting software according to claim 1 is characterized in that: The basic image information includes: On-site images and renderings of water treatment projects; Physical images of various equipment in water treatment projects, Revit 3D model renderings and effect diagrams of various equipment; As well as real-life images and videos of water treatment project site surveys.

3. The method for generating water treatment engineering renderings based on AI painting software according to claim 1 is characterized in that: The training process of the AI ​​painting model is specifically as follows: Generate adversarial networks to synthesize scene data that lacks samples, so as to enhance and synthesize data sets; The enhanced and synthesized dataset is divided into training set, validation set and test set. The training set is used for model learning, the validation set is used to adjust model parameters and prevent model overfitting, and the test set is used to evaluate the generalization ability of the model. Select the Adam optimizer to update the model parameters and use the learning rate decay strategy to adjust the learning rate; Among them, the generator in the generative adversarial network generates new image samples, and the discriminator distinguishes whether the image is real or generated by the generator; Also, the generator uses a deep convolutional neural network with a network depth of 6 layers. Each layer uses the ReLU activation function. The last layer uses the tanh function to output image data. Batch normalization is used after each convolution layer. The discriminator uses a deep convolutional neural network with a network depth of 5 layers. The LeakyReLU activation function is used to increase nonlinearity. The last layer uses the sigmoid activation function to judge the authenticity of the image.

4. The method for generating water treatment engineering renderings based on AI painting software according to claim 1 is characterized in that: The process of obtaining the plan layout of the water treatment project and identifying the drawing information in the plan layout is specifically as follows: Obtain water treatment plan and layer information: Use CAD secondary development tools and programming languages ​​to obtain water treatment plan, legends, and the expression content of each layer; Identify the project redline boundary and calculate the area: In the layer information corresponding to the redline, use the Polyline object to find the polygon representing the redline boundary, and calculate the area through the Area property; Identify the position and size of each layer: Get the position and size through its Position and ScaleFactors properties. The layers include structures, buildings, red lines, walls, roads, and green areas. Identify characteristic information: Use DBText and MText functions to read text information in the text descriptions and annotations in the drawings to identify the characteristic information of each structure, building, road and green space.

5. The method for generating water treatment engineering renderings based on AI painting software according to claim 1 is characterized in that: The rendering requirements include: the scope of the rendering, the viewing angle of the rendering, the surrounding environment of the factory, the names of structures and buildings, and the requirements and characteristics of the rendering style.

6. The method for generating water treatment engineering renderings based on AI painting software according to claim 1 is characterized in that: The process of obtaining the rendering requirement information and generating prompt words in combination with the drawing information is specifically as follows: Perform calculation and analysis on the scope and viewing angle in the rendering requirement information and the names of structures and buildings in the same drawing information to obtain detailed equipment information included in the rendering; Based on the detailed equipment information, rendering requirement information and drawing information contained in the rendering, AI tools are used to generate prompt words.

7. The method for generating water treatment engineering renderings based on AI painting software according to claim 1 is characterized in that: The relevant parameters include effect image size, file format, resolution and number of iterations.

8. A water treatment engineering rendering generation system based on AI painting software, characterized by: include: The model training module is used to obtain basic image information of water treatment projects and use the basic image information as a data set to learn and train an AI painting model; An information recognition module is used to obtain the plan layout of the water treatment project and recognize the drawing information in the plan layout; A prompt word generation module is used to obtain the rendering requirement information and generate prompt words in combination with the drawing information; The rendering generation module is used to input the prompt words into the AI ​​painting model, set the relevant parameters of the rendering, and then generate the final rendering.

9. A computer terminal comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, it implements the method for generating water treatment project renderings based on AI painting software as described in any one of claims 1 to 7.

10. A computer readable medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the method for generating water treatment project renderings based on AI painting software as described in any one of claims 1 to 7.

Images