Gas station design style rapid conversion method and device based on AI

Abstract

The invention discloses an AI-based gas station design style rapid conversion method and device, belongs to the technical field of building design, and particularly relates to rapid conversion of gas station (building) design styles. The problems that in the prior art, the designed building style needs to be modified, modeling needs to be conducted again to adjust the style, time and labor are consumed, and the drawing period is long are solved, and the problems that traditional design rendering is higher in dependence on people, the output effect is directly affected by personal aesthetic preference of designers, the limitation of the style is difficult to break through, and thinking imagination is limited are solved. The method comprises the following steps: performing structure locking, style migration and local optimization on an original design drawing in stages according to a style reference drawing, and generating a final effect image and modifying a record file. The AI-based gas station design style rapid conversion method and device are suitable for rapid conversion of gas station (building) design styles.

Description

Technical Field

[0001] The present invention relates to the technical field of architectural design, and in particular to the rapid conversion of gas station (architecture) design styles. Background Art

[0002] At present, the rendering of gas station (building) renderings mainly relies on manual 3D modeling, and then uses rendering software to set parameters such as materials, lighting, and scenes. This has the following problems: (1) If the architectural style of the design needs to be modified, the traditional method requires re-modeling and adjusting the style. The repeated modeling workload is large, time-consuming and labor-intensive. In addition, image rendering requires high equipment performance and a long drawing cycle. The traditional process takes 3-5 days. (2) Traditional design rendering is more dependent on people. The designer’s personal aesthetic preferences directly affect the output effect. It is difficult to break through the limitations of style and the thinking and imagination are limited. Summary of the Invention

[0003] The present invention proposes an AI-based method and device for quickly converting the design style of a gas station, which solves the problems existing in the prior art of requiring modification of the designed architectural style, requiring remodeling and style adjustment, which is time-consuming and labor-intensive, and has a long drawing cycle. In addition, traditional design rendering is more dependent on humans, and the designer's personal aesthetic preferences directly affect the output effect, making it difficult to break through the limitations of style and limiting thinking and imagination.

[0004] The AI-based method for rapid conversion of gas station design styles described in the present invention is implemented based on the StableDiffusion visual programming program ConfyUI, and includes: Input step: receiving the original design drawing and style reference drawing; Processing steps: Based on the style reference image, the original design image is subjected to structural locking, style transfer, and local optimization in stages to generate the final effect image and modification record file; Output step: Output the final effect image and modification record file generated by the processing step.

[0005] Furthermore, a preferred embodiment is provided, wherein the processing steps include: Structural locking step: Extract the edge contours of gas station scene objects and the image depth map from the original design drawing, and combine them with the image depth map as the structural control map; Style transfer steps: extract the style feature vector from the style reference image; render the structure control image based on the style feature vector to obtain the initial rendering; Local optimization step: redraw the design style of the area to be modified in the initial rendering to adjust the design style of the area to be modified and obtain the final rendering; Modification record step: record the modifications to the original design drawing during the local optimization step and generate a modification record file.

[0006] Furthermore, a preferred embodiment is provided, wherein the structural locking step comprises: Get the input original design drawing; The Controlnet Canny edge detector is used to extract the edge contours of the gas station scene objects in the original design drawing; Use the Depth detector to extract the image depth map from the original design image; The edge contour lines of objects in the gas station scene and the image depth map are combined as the output of the structure control map.

[0007] Furthermore, a preferred embodiment is provided, in the style transfer step: The style feature vector of the style reference image is extracted through the IP-Adapter module.

[0008] Furthermore, a preferred embodiment is provided, wherein the local optimization step comprises: Generate a region mask based on the semantic segmentation model, and lock the area to be modified in the initial rendering as the target area; Through the latent space restoration technology, only the latent code of the target area is adjusted in design style to complete the local redrawing. The other areas in the initial effect image remain unchanged to obtain the final effect image.

[0009] Furthermore, a preferred embodiment is provided, wherein the method further includes the steps of outputting multiple solutions and optimizing the solution: By using ComfyUI's batch image processing function, you can achieve batch replacement of the design style of the final effect image and form a variety of design options for selection.

[0010] The present invention also proposes an AI-based device for quickly converting gas station architectural design styles. The device is implemented based on the Stable Diffusion visual programming program ConfyUI and includes: Input layer: receives the original design image and style reference image; Processing layer: Based on the style reference image, the original design image is subjected to structural locking, style transfer, and local optimization in stages to generate the final effect image and modification record file; Output layer: The final effect image and modification record file generated by the output processing layer.

[0011] The present invention also proposes a computer device comprising: a processor and a memory, wherein the memory is used to store executable instructions of the processor, and the processor is configured to execute any one of the above-mentioned AI-based gas station design style rapid conversion methods by executing the executable instructions.

[0012] The present invention also proposes a computer storage medium, in which a computer program is stored. When the computer program is run, any one of the above-mentioned AI-based gas station design style rapid conversion methods is executed.

[0013] The present invention also proposes a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of any one of the above-mentioned AI-based gas station design style rapid conversion methods.

[0014] The present invention has the following beneficial effects: 1. The AI-based method and device for rapid gas station design style conversion described in this invention utilizes a node-based workflow (ComfyUI, an AI generation technology) and GPU acceleration to compress style adjustment to within 5 minutes. This compares to the 3-5 days required for traditional manual modeling, resulting in a more than 40-fold increase in efficiency. Furthermore, the method can be run on a standard office computer and supports batch generation.

[0015] 2. The AI-based method for rapid gas station design style conversion described in this invention uses an AI algorithm to extract the gas station structure (contour, size, layout), decoupling the design structure from the visual style, achieving "changing the style without changing the structure." Modifications do not require rebuilding the model, and materials, colors, and other elements can be adjusted directly based on the original image, significantly improving efficiency.

[0016] 3. The AI-based rapid gas station design style conversion method described in this invention combines edge (contour line) detection and depth control (image depth map) to achieve a key parameter error rate of less than 3%, comparable to manual modeling accuracy and high structural feature restoration.

[0017] 4. The AI-based method for rapid gas station design style conversion described in this invention breaks through the limitations of designers' personal aesthetics and can support multi-style hybrid adjustment. By quantitatively controlling style characteristics through parameterized controls (style transfer strength settings), AI automatically generates novel solutions that go beyond human experience.

[0018] 5. The AI-based method for rapid gas station design style conversion described in this invention breaks through the limitations of manual style conversion, supports preset styles and mixed adjustment of multiple styles, and can batch generate 4-8 sets of solutions at a time (through multiple solution output and optimization steps). The style diversity is more than 5 times that of traditional designs. AI can generate innovative combinations that exceed the existing experience library.

[0019] 6. The AI-based gas station design style rapid conversion method described in the present invention achieves precise regional modification through semantic segmentation and latent space repair, and has high accuracy. It is more efficient than traditional PS modification and has strong detail modification capabilities.

[0020] The AI-based method and device for rapid conversion of gas station design styles described in the present invention are suitable for rapid conversion of gas station (building) design styles. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 A schematic flow chart of a method for rapidly converting gas station design styles based on AI in one embodiment of the present invention; Figure 2 A schematic diagram of final effect images of different design styles in one embodiment of the present invention; Figure 3 A schematic diagram of a structure locking (structure feature extraction) step in one embodiment of the present invention; Figure 4 A schematic diagram of style transfer parameter configuration in one embodiment of the present invention; Figure 5 A schematic diagram of a local optimization (local redrawing) step in one embodiment of the present invention; Figure 6 FIG. 1 is a schematic diagram of an image batch processing function in one embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the technical solutions and advantages of the present invention more clearly described, the specific embodiments of the present invention will be further described in detail and completely in conjunction with the accompanying drawings. The various embodiments described below are only part of the preferred embodiments of the present invention, rather than all implementation plans; the various embodiments described below are intended to explain the present invention and cannot be understood as limiting the present invention; the reasonable combination of the technical features defined in the various embodiments of the present invention, as well as all other implementation plans obtained by ordinary technicians in this field based on the embodiments of the present invention without making creative work, all fall within the scope of protection of the present invention.

[0024] Implementation 1: An AI-based method for rapid conversion of gas station design styles, implemented based on the StableDiffusion visual programming program ConfyUI, includes: Input step: receiving the original design drawing and style reference drawing; Processing steps: Based on the style reference image, the original design image is subjected to structural locking, style transfer, and local optimization in stages to generate the final effect image and modification record file; Output step: Output the final effect image and modification record file generated by the processing step.

[0025] In this embodiment, the method is implemented based on the Stable Diffusion visual programming program ConfyUI to establish a rendering style replacement workflow; the method is collaboratively implemented by multiple workflow modules, and the method is encapsulated as a whole for subsequent application.

[0026] In this implementation, Stable Diffusion (SD) is an open-source generative artificial intelligence technology based on the Latent Diffusion Model (LDM). It can generate high-quality visual content (such as paintings, photos, and design drawings) based on text, images, or other conditional inputs. It is one of the core technologies in the current field of AI image generation.

[0027] In this implementation, ComfyUI is a Stable Diffusion image generation tool based on node-based visual programming. Its core is to achieve precise control of the AI ​​painting process through modular node connections. Its core positioning and features are as follows: Node-based workflow design: Users build the image generation pipeline by dragging and connecting nodes with different functions (such as model loading, prompt word encoding, and samplers). Each node represents a processing step in Stable Diffusion (such as text encoding, latent space diffusion, and image decoding). This design, similar to building a circuit board, makes the generation process completely transparent and traceable.

[0028] Modularity and scalability: Supports custom node combinations. Users can freely adjust parameters (such as sampling steps and scheduler type) or insert expansion modules such as ControlNet and LoRA to achieve complex operations such as multi-model series connection and local redrawing.

[0029] Efficient resource management: Compared with traditional WebUI, ComfyUI has low memory usage, faster startup and drawing speed, and can share model libraries (such as Checkpoint, LoRA, etc.) with WebUI.

[0030] In this implementation, ComfyUI is adopted to realize the rapid conversion of the style of the design rendering by writing workflow nodes in steps. Finally, multiple workflows are formed and encapsulated into a "rapid conversion of gas station rendering design style" workflow for practical application. At the same time, due to the strong openness of ComfyUI, it can be rewritten to a certain extent according to specific rules to make it more applicable.

[0031] In this embodiment, ComfyUI combines GPU acceleration technology to compress the style adjustment to within 5 minutes.

[0032] In this embodiment, the original design drawing is an original design drawing of a gas station.

[0033] In this embodiment, the modification record file records the modification records of the original design drawing in the processing steps.

[0034] Implementation method 2: The processing steps include: Structural locking step: Extract the edge contours of gas station scene objects and the image depth map from the original design drawing, and combine them with the image depth map as the structural control map; Style transfer steps: extract the style feature vector from the style reference image; render the structure control image based on the style feature vector to obtain the initial rendering; Local optimization step: redraw the design style of the area to be modified in the initial rendering to adjust the design style of the area to be modified and obtain the final rendering; Modification record step: record the modifications to the original design drawing during the local optimization step and generate a modification record file.

[0035] In this embodiment, the structure locking step may also be referred to as a structure extraction step.

[0036] In this embodiment, the style transfer step may also be referred to as a style decoupling step.

[0037] In this embodiment, the local optimization step may also be referred to as a local redrawing step.

[0038] In this embodiment, the area that needs to be modified is to see whether the design effect conforms to the design idea. The area where the local design effect does not conform to the design idea is the area that needs to be modified.

[0039] In this embodiment, during the local optimization step, the design compliance can also be verified in real time.

[0040] The real-time verification of design compliance includes detecting whether parameters such as the distance between the gas pump and the station building and the lane width are compliant.

[0041] The "regulations" for design compliance mentioned above refer to the "GB 50156-2021" standard.

[0042] If any violation is found (such as insufficient fire protection distance), the system will automatically redraw the part until the output plan fully complies with the "GB50156-2021" standard, achieving "modification means compliance" and avoiding blind spots in manual review.

[0043] In this embodiment, the modification record step: each generated effect image and the mask image of the local modification can be obtained through the image library, and then the modification record file can be obtained.

[0044] In this embodiment, the modification record file is used to understand the area modified each time and compare the effects of the solutions.

[0045] Implementation 3: The structure locking step includes: Get the input original design drawing; The Controlnet Canny edge detector is used to extract the edge contours of the gas station scene objects in the original design drawing; Use the Depth detector to extract the image depth map from the original design image; The edge contour lines of objects in the gas station scene and the image depth map are combined as the output of the structure control map.

[0046] In addition, in one embodiment, the image resolution of the original design drawing is 1920×1080 and adopts RGB format.

[0047] Implementation 4: In the style transfer step: The style feature vector of the style reference image is extracted through the IP-Adapter module.

[0048] In this embodiment, the style reference image is an input reference image for replacing the style, such as a Chinese architecture reference or a Hui style architecture reference.

[0049] In this embodiment, the style feature vector includes color distribution, material texture, and light and shadow pattern.

[0050] In addition, in one embodiment, the method further includes a style transfer strength setting step: Each style feature vector has a corresponding weight parameter. Setting the weight parameter of each style feature vector changes its reference strength to the design style: The higher the weight parameter of the style feature vector is set, the stronger its reference to the design style.

[0051] In this embodiment, style parameterization configuration is achieved through the style transfer strength setting step.

[0052] Implementation 5: The local optimization step includes: Generate a region mask based on the semantic segmentation model, and lock the area to be modified in the initial rendering as the target area; Through the latent space restoration technology, only the latent code of the target area is adjusted in design style to complete the local redrawing. The other areas in the initial effect image remain unchanged to obtain the final effect image.

[0053] In this embodiment, the semantic segmentation model is such as U-Net.

[0054] In this embodiment, the regional mask is called Mask in English.

[0055] In this embodiment, the areas that need to be modified include the canopy and the refueling island.

[0056] In this embodiment, the latent space restoration technology is called Latent Inpainting in English.

[0057] Implementation 6: The method further includes multiple solution output and optimization steps: By using ComfyUI's batch image processing function, you can achieve batch replacement of the design style of the final effect image and form a variety of design options for selection.

[0058] In this embodiment, the image batch processing function of ComfyUI is used to quickly realize batch replacement of the design style of the final effect image, forming a variety of design schemes for selection, greatly improving work efficiency and simplifying the process.

[0059] Implementation 7: An AI-based device for rapidly converting gas station architectural design styles, implemented based on the StableDiffusion visual programming program ConfyUI, comprising: Input layer: receives the original design image and style reference image; Processing layer: Based on the style reference image, the original design image is subjected to structural locking, style transfer, and local optimization in stages to generate the final effect image and modification record file; Output layer: The final effect image and modification record file generated by the output processing layer.

[0060] Embodiment 8: A computer device comprises: a processor and a memory, wherein the memory is used to store executable instructions of the processor, and the processor is configured to execute any one of the above-mentioned AI-based gas station design style rapid conversion methods by executing the executable instructions.

[0061] Implementation method 9: A computer storage medium having a computer program stored therein, wherein when the computer program is run, any one of the above-mentioned AI-based gas station design style rapid conversion methods is executed.

[0062] Embodiment 10: A computer program product, comprising a computer program / instruction, which, when executed by a processor, implements the steps of any one of the above-mentioned AI-based gas station design style rapid conversion methods.

[0063] This embodiment provides a computer device or system, the hardware device of this part is a general model and is not shown in the form of a diagram. The system includes a processor and a memory, wherein the processor and the memory can be connected through a bus or other means. The memory is a non-transient computer-readable storage medium that can be used to store non-transient software programs, non-transient computer executable programs and modules, and corresponding program instructions / modules. The processor executes various functional applications and data processing of the processor by running the non-transient software programs, instructions and modules stored in the memory, so as to realize the data space entity resolution data quality enhancement method in the above method embodiment.

[0064] The memory may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created by the processor, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory may optionally include a memory remotely located relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, an intranet, a mobile communication network, and combinations thereof.

[0065] One or more modules are stored in the memory. When the processor executes, the method steps in the embodiment are executed. In this way, the purpose of the invention can be achieved through the method, device and process of the present invention. The specific details of the above-mentioned computer equipment can be understood by referring to the corresponding descriptions and effects in the embodiment, and will not be repeated here.

[0066] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware through a computer program. The program can be stored in a computer-readable storage medium, and when executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (HDD), or a solid-state drive (SSD). The storage medium can also include a combination of the above-mentioned types of memory.

[0067] The above further describes the technical solution provided by the present invention in detail through several specific embodiments in order to highlight the advantages and benefits of the technical solution provided by the present invention. However, the several specific embodiments described above are not intended to limit the present invention. Any reasonable changes and improvements to the present invention, reasonable combinations of implementation methods and equivalent replacements based on the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. AI-based gas station design style rapid conversion method, characterized by: The method is implemented based on the StableDiffusion visual programming program ConfyUI, and the method includes: Input step: receiving the original design drawing and style reference drawing; Processing steps: Based on the style reference image, the original design image is subjected to structural locking, style transfer, and local optimization in stages to generate the final effect image and modification record file; Output step: Output the final effect image and modification record file generated by the processing step.

2. The AI-based gas station design style rapid conversion method according to claim 1 is characterized in that: The processing steps include: Structural locking step: Extract the edge contours of gas station scene objects and the image depth map from the original design drawing, and combine them with the image depth map as the structural control map; Style transfer steps: extract the style feature vector from the style reference image; render the structure control image based on the style feature vector to obtain the initial rendering; Local optimization step: redraw the design style of the area to be modified in the initial rendering to adjust the design style of the area to be modified and obtain the final rendering; Modification record step: record the modifications to the original design drawing during the local optimization step and generate a modification record file.

3. The AI-based gas station design style rapid conversion method according to claim 2 is characterized in that: The structure locking step includes: Get the input original design drawing; The Controlnet Canny edge detector is used to extract the edge contours of the gas station scene objects in the original design drawing; Use the Depth detector to extract the image depth map from the original design image; The edge contour lines of objects in the gas station scene and the image depth map are combined as the output of the structure control map.

4. The AI-based gas station design style rapid conversion method according to claim 2 is characterized in that: In the style transfer step: The style feature vector of the style reference image is extracted through the IP-Adapter module.

5. The AI-based gas station design style rapid conversion method according to claim 2 is characterized in that: The local optimization step includes: Generate a region mask based on the semantic segmentation model, and lock the area to be modified in the initial rendering as the target area; Through the latent space restoration technology, only the latent code of the target area is adjusted in design style to complete the local redrawing. The other areas in the initial effect image remain unchanged to obtain the final effect image.

6. The AI-based gas station design style rapid conversion method according to claim 2 is characterized in that: The method further includes the steps of outputting multiple solutions and optimizing the solution: Utilize ComfyUI's batch image processing function to achieve batch replacement of the design style of the final effect image, forming a variety of design options for selection.

7. AI-based gas station architectural design style rapid conversion device, characterized by: The device is implemented based on the StableDiffusion visual programming program ConfyUI, and the device includes: Input layer: receives the original design image and style reference image; Processing layer: Based on the style reference image, the original design image is subjected to structural locking, style transfer, and local optimization in stages to generate the final effect image and modification record file; Output layer: The final effect image and modification record file generated by the output processing layer.

8. A computer device comprising: A processor and a memory, characterized in that the memory is used to store executable instructions of the processor, and the processor is configured to execute the AI-based gas station design style rapid conversion method described in any one of claims 1-6 by executing the executable instructions.

9. A computer storage medium, characterized in that The storage medium stores a computer program, and when the computer program is run, the method for quickly converting the design style of a gas station based on AI according to any one of claims 1 to 6 is executed.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the steps of the AI-based gas station design style rapid conversion method according to any one of claims 1 to 6 are implemented.

Images