CAD file generation method and system based on image recognition

Abstract

The invention relates to a CAD file generation method and system based on image recognition. The method comprises the steps that an image of a drawing is acquired; preprocessing the image; identifying the line, shape and labeling information of the graph corresponding to the image through an AI algorithm; in response to the fact that at least one graph lacks annotation information, the graph lacking the annotation information is sent to the user, and generation of the CAD file is stopped; in response to the fact that all the graphs do not lack the labeling information, the relative positions and the line directions between the lines are recognized through a geometric analysis algorithm; converting the obtained result into data which can be identified by drawing tool software to obtain a CAD file; the difference between the CAD file and the image is identified through the image processing technology, and then the CAD file is corrected through the AI model according to the difference. According to the method and the device, the problems that time consumed for drawing CAD files by technicians according to paper drawings or images is relatively long, and time consumed for drawing CAD files due to lack of technicians or limited abilities of the technicians is relatively long are solved.

Description

Technical Field

[0001] The present invention generally relates to the technical field of image recognition. More specifically, the present invention relates to a method and system for generating CAD files based on image recognition. Background Art

[0002] Before machining a component, it is necessary to design a machining program for the component, and the component is automatically machined by subtractive cutting using a numerical control machine tool or the like according to the machining program, or automatically manufactured by additive manufacturing using a 3D printer according to the machining program. Since it is more beneficial to program the component based on an engineering drawing file than a paper drawing or the like to obtain a machining program, a programmer can program according to the engineering drawing file of the component to obtain a machining program. The engineering drawing file includes 2D engineering drawing files and 3D engineering drawing files. The 2D engineering drawing file includes a CAD file, which is an electronic engineering drawing, and the format of the CAD file can be DWG format, DWT format, DWS format, etc. Moreover, the engineering drawing file is more convenient for modification, storage, and transmission than a paper drawing or the like, and users have a need to obtain corresponding engineering drawing files based on paper drawings of components, images of paper drawings, or pictures corresponding to 2D engineering drawing files.

[0003] In the prior art, in addition to programming components according to 2D engineering drawing files, components can also be programmed according to 3D engineering drawing files.

[0004] A user (order issuer) can post demand information for obtaining a machining program based on an electronic engineering drawing of a component on a research and development trading platform. A person receiving the order can design a corresponding machining program according to the platform order corresponding to the demand information. The platform order includes the electronic engineering drawing of the component, etc. After the person receiving the order finishes designing the machining program, the machining program, etc. can be uploaded or forwarded to the research and development trading platform for the order issuer to receive and review.

[0005] However, a user may only have a paper drawing of a component, an image of the paper drawing, or a picture corresponding to a 2D engineering drawing file. Currently, the user will ask a drafting technician to draw an engineering drawing file, etc. based on the paper drawing, the image of the paper drawing, or the picture corresponding to the 2D engineering drawing file. The engineering drawing file can be a CAD file, but this method takes a long time. Summary of the Invention

[0006] To solve the above one or more technical problems, the present invention proposes that when all graphics do not lack annotation information, the results obtained by AI algorithms and geometric analysis algorithms are converted into data that can be recognized by the corresponding drawing tool software, and then the corresponding CAD file is generated through the drawing tool software; the differences between the CAD file and the image are recognized, and then the CAD file is corrected by the AI model according to the differences.

[0007] To this end, the present invention provides solutions in the following aspects.

[0008] In the first aspect of the embodiments of the present invention, a method for generating a CAD file based on image recognition is provided, including: acquiring an image of a drawing; preprocessing the image, where the image includes lines, shapes, and annotation information of graphics corresponding to component elements; identifying the lines, shapes, and annotation information of the graphics through an AI algorithm; in response to at least one graphic lacking annotation information, sending the graphic lacking annotation information to the user and stopping generating the CAD file; in response to all graphics not lacking annotation information, identifying the relative positions between lines and the line directions through a geometric analysis algorithm based on AI; converting the results obtained through the AI algorithm and the geometric analysis algorithm into data recognizable by a corresponding drawing tool software, and then generating a corresponding CAD file through the drawing tool software; identifying the differences between the CAD file and the image through image processing technology, and then correcting the CAD file according to the differences through an AI model.

[0009] In one embodiment, preprocessing the image includes: in response to lines that are horizontal or vertical to the four sides of the photo in the normal case in the image being inclined, correcting the image to be upright through AI technology.

[0010] In one embodiment, preprocessing the image further includes: adjusting the brightness and contrast of the image, and denoising.

[0011] In one embodiment, identifying the lines, shapes, and annotation information of the graphics through an AI algorithm includes: identifying the lines, shapes, and annotation information of the graphics by using opencv and OCR technologies.

[0012] In one embodiment, the geometric analysis algorithm includes: an edge detection algorithm and a line matching algorithm.

[0013] In one embodiment, the image processing technology includes feature point matching.

[0014] In one embodiment, the differences include incorrect annotation information, missing graphics, and inaccurate line positions.

[0015] In one embodiment, it further includes: providing an interactive interface for the user to modify the correction of the AI model, and also providing an interactive interface for the user to adjust the CAD file.

[0016] In the second aspect of the embodiments of the present invention, a CAD file generation system based on image recognition is provided, including: a processor and a memory, where the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the methods of any of the above embodiments are implemented.

[0017] In the third aspect of the embodiments of the present invention, a computer-readable storage medium is provided, in which program instructions are stored, and when the program instructions are loaded and executed by a processor, the processor is caused to execute the method according to any of the embodiments in the first aspect of the embodiments of the present invention.

[0018] The beneficial effects of the present invention include: When at least one graphic in the image lacks annotation information, all the graphics lacking annotation information are sent to the user, and the generation of the CAD file corresponding to the image is stopped; in response to all the graphics not lacking annotation information, the results obtained by the AI algorithm and the geometric analysis algorithm are converted into data that can be recognized by the corresponding drawing tool software, and then the corresponding CAD file is generated through the drawing tool software. Before generating the CAD file, the graphics in the image do not lack annotation information, which improves the accuracy of the generated CAD file. According to the image of the drawing, the CAD file is generated by the software, which takes less time. Through image processing technology, the differences between the CAD file and the image are identified, and then the CAD file is corrected according to the differences and through the AI model, further improving the accuracy of the CAD file. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown by way of illustration and not limitation, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 is a flowchart schematically showing a method for generating a CAD file based on image recognition according to an embodiment of the present invention; Figure 2 is a block diagram schematically showing the structure of a CAD file generation system based on image recognition according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0021] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0022] Figure 1 It is a flowchart schematically showing a method for generating a CAD file based on image recognition according to an embodiment of the present invention. As Figure 1 shown, first, according to a first aspect of the present invention, there is provided a method for generating a CAD file based on image recognition, including: obtaining an image of a drawing; preprocessing the image, the image including the lines, shapes, and annotation information of the graphics corresponding to the component elements; identifying the lines, shapes, and annotation information of the graphics through an AI algorithm; in response to at least one graphic lacking annotation information, sending the graphic lacking annotation information to the user and stopping generating the CAD file; in response to all graphics not lacking annotation information, identifying the relative positions between the lines and the line directions through an AI-based geometric analysis algorithm; converting the results obtained through the AI algorithm and the geometric analysis algorithm into data recognizable by a corresponding drawing tool software, and then generating a corresponding CAD file through the drawing tool software; identifying the differences between the CAD file and the image through image processing technology, and then correcting the CAD file according to the differences through an AI model.

[0023] Specifically, as Figure 1 shown, it includes steps S1 to S6: In step S1, an image of a drawing is obtained.

[0024] Before processing a component, it is necessary to design a processing program for the component, and perform automatic subtractive cutting manufacturing on the component through a numerical control machine tool, etc., or perform automatic additive manufacturing through a 3D printer according to the processing program. Since it is more beneficial to program the component according to an engineering drawing file compared to a paper drawing, etc., to obtain a processing program, a programmer can program according to the engineering drawing file of the component to obtain a processing program. The engineering drawing file includes 2D engineering drawing files and 3D engineering drawing files. The 2D engineering drawing file includes a CAD file, which is an electronic engineering drawing, and the format of the CAD file can be a DWG format, a DWT format, a DWS format, etc. Moreover, the engineering drawing file is more convenient for modification, storage, and transmission compared to a paper drawing, etc., and users have a need to obtain a corresponding engineering drawing file based on a paper drawing of a component, an image of a paper drawing, or a picture corresponding to a 2D engineering drawing file.

[0025] In the prior art, in addition to programming components according to 2D engineering drawing files, components can also be programmed according to 3D engineering drawing files.

[0026] When a user only has paper drawings of parts, images of paper drawings, or pictures corresponding to 2D engineering drawing files, etc., the user is very likely to need the CAD file corresponding to the drawing. Because when obtaining the processing program for the part, programmers can program the part based on the CAD file to obtain the processing program, and engineering drawing files are more convenient for modification, storage, and transmission compared to paper drawings, etc.

[0027] To obtain the CAD file corresponding to the drawing, if the user only has the paper drawing of the part, the drawing of the part needs to be scanned into an image, or a photo image of the drawing can be taken using a photographic device. The drawing can be the six views and sectional views of the part. The six views include the front view, top view, side view, bottom view, the other side view, and the back view. The format of the image can be JPEG format, etc.

[0028] In step S2, preprocess the image. The image includes the lines, shapes, and annotation information of the graphics corresponding to the part elements.

[0029] Graphics are composed of lines. The annotation information of the graphics includes dimension annotations and symbols, etc. For example, when the graphic is a circle and the dimension annotation of the circle is the radius, the symbol of the radius can be added to the dimension annotation, which can be r.

[0030] The lines, graphics, or annotation information in the image may not be clear, and there may be interference information in the image, such as noise, etc. And normally, some lines in the image are perpendicular or parallel to the edge of the photo. After taking the photo or scanning, they are neither perpendicular nor parallel to the edge of the photo, but are inclined at a certain angle. Therefore, it is necessary to preprocess the image to reduce the possibility of errors in the generated CAD file.

[0031] In step S3, identify the lines, shapes, and annotation information of the graphics through an AI algorithm.

[0032] The lines, shapes, and annotation information of the graphics can be identified through some existing technologies. For example, the lines, shapes, and annotation information of the graphics can be identified through artificial intelligence (AI) algorithms to facilitate subsequent searching for graphics lacking annotation information.

[0033] In step S4, in response to at least one graphic lacking annotation information, send the graphic lacking annotation information to the user and stop generating the CAD file.

[0034] When at least one graphic in the image lacks annotation information, send all the graphics lacking annotation information to the user and stop generating the CAD file corresponding to the image.

[0035] Compare the annotation information corresponding to the graphics with the standard annotation information of the same type of graphics. The standard annotation information of the graphics includes dimension annotations indicating the size of the graphics and positioning dimension annotations indicating the position of the graphics. For example, if a graphic is a rectangle, the standard annotation information of the same type of graphics includes the dimension annotation of the length of the rectangle, the dimension annotation of the width of the rectangle, and the positioning dimension annotation. If the annotation information of this graphic only includes the dimension annotation of the length of the rectangle and the positioning dimension annotation, then this graphic lacks annotation information. If the annotation information of a graphic is missing compared to the standard annotation information of the same type of graphics, the size and / or position of this graphic cannot be accurately determined. To make the generated CAD file more accurate, send the graphic lacking annotation information to the user and stop generating the CAD file. After sending the graphic lacking annotation information to the user, the user can re-upload the image corresponding to the modified drawing.

[0036] In step S5, in response to all graphics not lacking annotation information, identify the relative positions and line directions between the lines through an AI-based geometric analysis algorithm; convert the results obtained through the AI algorithm and the geometric analysis algorithm into data that can be recognized by the corresponding drawing tool software, and then generate the corresponding CAD file through the drawing tool software.

[0037] When all the graphics in the image do not lack annotation information, the relative positions and line directions between the lines can be identified through existing technologies. For example, the relative positions and line directions between the lines can be identified through an AI-based geometric analysis algorithm. After obtaining the relative positions and line directions between the lines, combined with the lines, shapes, and annotation information of the graphics identified through the AI algorithm, the results obtained through the AI algorithm and the geometric analysis algorithm can be converted into data that can be recognized by the drawing tool software corresponding to the image. This data can be the drawing instructions of the drawing tool software corresponding to the image, and then the CAD file corresponding to the image can be generated through the drawing tool software. For example, the results identified through the AI algorithm and the geometric analysis algorithm can be converted into a data format that can be recognized by the drawing tool software, and this data can be interacted with the API of the drawing tool software through a Python script to transmit the data, and then the CAD file corresponding to the image can be generated through the drawing tool software. The drawing tool software can be AutoCAD, etc.

[0038] In step S6, through image processing technology, identify the differences between the CAD file and the image, and then correct the CAD file according to the differences through an AI model.

[0039] Under normal circumstances, the line information, graphic information, and dimension information corresponding to the component elements in the CAD file are consistent with the line information, graphic information, and dimension information in the corresponding image. When an abnormal situation occurs, that is, when the line information, graphic information, and dimension information corresponding to the component elements in the CAD file are inconsistent with the line information, graphic information, and dimension information in the image, existing technologies such as image processing technology can be used to identify the differences between the CAD file and the image. Furthermore, based on this difference and through an AI model, the CAD file can be corrected so that the line information, graphic information, and dimension information corresponding to the component elements in the CAD file are consistent with the line information, graphic information, and dimension information in the image.

[0040] Using the trained AI model and inputting the difference information of the CAD file, the model will output corresponding correction operations, such as transformation parameters, shape change parameters, etc. Applying these correction operations to the CAD file can make the CAD file consistent with the original image. Among them, the training process of the AI model is as follows: Using machine learning or deep learning technology to train the AI model, which can map the differences between the CAD file and the original image to correction operations. The training data can include the original CAD file, the corresponding original image, and the corrected CAD file.

[0041] The AI model can improve the accuracy of correcting the CAD file through multiple trainings.

[0042] In one embodiment, preprocessing the image includes: preprocessing the image, including: in response to the lines that are normally horizontal or vertical to the four sides of the photo in the image being inclined, correcting the image to be upright through AI technology.

[0043] When the lines that are normally horizontal or vertical to the four sides of the photo in the image are inclined, it is not conducive to identifying the shape of the graphic, etc. It may cause traditional feature extraction algorithms to be unable to accurately extract key feature points or feature descriptors, and may also cause confusion between the graphic and other types of graphics, and thus may result in errors in identifying the category of the graphic. Therefore, the image can be corrected to be upright through technology based on angle detection and rotation. This technology uses computer vision technology to detect the main lines, edges, and angles in the image and automatically rotates the image to align it or make it perpendicular to the reference direction, which can be the edge of the photo. The main algorithms include the Hough transform, SIFT feature point matching, convolutional neural network based on deep learning, etc.

[0044] The image can also be rectified to be upright through a rectification technique based on feature point matching. This technique uses computer vision technology to detect feature points (such as corner points, edges, textures, etc.) in the image, and automatically rotates or rectifies the image through feature point matching or registration. Among them, the main algorithms include SIFT, SURF, ORB, etc.

[0045] In one embodiment, preprocessing the image further includes: adjusting the brightness and contrast of the image, and denoising.

[0046] By adjusting the brightness and contrast of the image, and denoising, the image becomes clearer. A clearer image is more conducive to identifying the lines, shapes, and annotation information of the graph. Among them, the brightness of the image can be increased or decreased through linear transformation or histogram equalization; the contrast of the image can be increased or decreased by adjusting the gray range of the image. For example, linear stretching or adaptive histogram equalization can be used to adjust the contrast of the image; the image can be denoised through wavelet denoising methods. Wavelet denoising is a denoising method based on wavelet transform. The image can also be denoised through bilateral filtering, mean shift filtering, and adaptive filtering, etc.

[0047] In one embodiment, identifying the lines, shapes, and annotation information of the graph through an AI algorithm includes: using opencv and OCR technology to identify the lines, shapes, and annotation information of the graph. Specifically, the edge detection algorithm of OpenCV (such as Canny edge detection) can be used to detect the edges of the graph, which helps to find the contours of the lines and shapes; the contour extraction function of OpenCV (such as findContours) can also be used to extract the contours of the graphs in the image. For the extracted contours, the shape matching function of OpenCV (such as matchShapes) or the method based on feature descriptors (such as Hu moments, Zernike moments) can be used for shape recognition. Among them, by comparing the extracted contours with the preset shape templates, the shape type represented by the contours can be determined. If there is annotated text (such as dimension information) on the graph, OCR technology (such as Tesseract OCR) can be used to identify and extract the text information. Finally, the text is associated with the corresponding graph to obtain graph-related information.

[0048] In another embodiment, the lines, shapes, and annotation information of the graph can also be identified through neural networks, OCR technology, and edge detection algorithms.

[0049] In one embodiment, the geometric analysis algorithm includes: an edge detection algorithm and a line matching algorithm.

[0050] Through an AI-based geometric analysis algorithm, the relative positions and directions of lines can be identified. Specifically: Classic edge detection algorithms (such as Canny edge detection) or deep learning-based edge detection models (such as U-Net, FCN, etc.) can be used to extract line information in the image. The extracted lines can be matched to determine the relative position and direction information between them. Specifically, traditional geometric calculation methods (such as Hough transform, line fitting, etc.) or deep learning-based matching models can be used to classify and match the lines. Calculate the relative position information between the lines. Specifically, based on the line matching results, information such as the angle and intersection situation between the lines can be calculated. For example, geometric calculation methods (such as vector operations, matrix transformations, etc.) can be used to achieve the calculation of the relative position. Traditional geometric calculation methods (such as vector operations, angle calculations, etc.) or deep learning-based classification models can be used to identify the line directions.

[0051] In one embodiment, the image processing technology includes feature point matching.

[0052] Through image processing technology, the differences between the CAD file and the image are identified. Specifically: First, feature extraction is performed on the CAD file and the original image respectively. For example, some common feature extraction algorithms, such as SIFT, SURF, or ORB, etc., can be used to extract key points and corresponding descriptors. Then, a feature matching algorithm, such as Nearest Neighbor matching or a filter-based matching algorithm (such as FLANN), can be used to correspond and match the feature points in the CAD file with the feature points in the original image. Then, reliable matching points can be screened according to the matching score or distance threshold. Furthermore, the accuracy of the matching can be evaluated by comparing the distances or similarities between the feature points, and the matching points with high similarity are retained. Then, according to the screened matching points, connection lines or marks can be drawn to visualize the differences between the CAD file and the original image. Finally, according to the visualization results of the differences, the elements in the CAD file that are different from the image are determined.

[0053] In another embodiment, a deep learning-based comparison algorithm can be used to compare the generated CAD file with the original image to obtain the differences between the CAD file and the original image. Among them, the deep learning models used include convolutional neural networks (CNNs) and generative adversarial networks (GANs). Appropriate model structures can be selected according to actual needs and dataset complexity. The loss function during model training can be the mean square error or the adversarial loss function to guide the model to learn how to compare the differences between the CAD file and the original image.

[0054] In one embodiment, the differences include incorrect annotation information, missing graphics, inaccurate line positions, etc.

[0055] In one embodiment, an interactive interface for the user to modify the AI model is provided, and an interactive interface for the user to adjust the CAD file is also provided.

[0056] Since there may be cases of correction errors in the CAD file after correction compared with the original image, there may also be cases where some differences are not corrected, and there may also be cases where some differences between the CAD file and the image are not recognized. Therefore, an interactive interface for the user to modify the AI model is provided, and an interactive interface for the user to adjust the CAD file is also provided to improve the consistency between the CAD file and the original image.

[0057] It should be noted that some users require to obtain the CAD file within a shorter time, and some users can also obtain the CAD file after a longer time. When there are multiple tasks to be processed on the software platform simultaneously, one task corresponds to the workload of generating a CAD file from a drawing image. Different users corresponding to these multiple tasks may have different requirements for the time to obtain the CAD file. The software platform provides an interface for changing the processing order of these multiple tasks. In this interface, relevant staff can adjust the processing order of these multiple tasks by the software platform according to the different requirements of different users corresponding to these multiple tasks for the time to obtain the CAD file, so that the user can obtain the CAD file within the required time. The shorter the time requirement of a user corresponding to a task for obtaining the CAD file, the more preferably this task should be processed by the software platform.

[0058] Figure 2 It is a schematic block diagram showing the structure of a CAD file generation system based on image recognition according to this embodiment.

[0059] The present invention also provides a CAD file generation system based on image recognition. As Figure 2 shown, the system includes a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a method for generating a CAD file based on image recognition according to the first aspect of the present invention is implemented.

[0060] The system also includes other components well-known to those skilled in the art such as a communication bus and a communication interface. Their settings and functions are known in the art, so they will not be described in detail here.

[0061] In the present invention, the foregoing memory may be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, the computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as, resistive random access memory (RRAM), dynamic random access memory (DRAM), static random-access memory (SRAM), enhanced dynamic random access memory (EDRAM), high-bandwidth memory (HBM), hybrid memory cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application, module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions that can be stored or otherwise held by such a computer-readable medium.

[0062] In the description of this specification, "a plurality of" and "several" mean at least two, for example, two, three or more, etc., unless otherwise specifically defined.

[0063] Although this specification has shown and described multiple embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and scope of the present invention. It should be understood that various alternatives to the embodiments of the present invention described herein may be employed in practicing the present invention.

Claims

1. A method for generating CAD files based on image recognition, characterized in that Comprising: Obtaining an image of the drawing; Preprocessing the image, the image including lines, shapes and annotation information of the corresponding graphics of component elements; Identifying the lines, shapes and annotation information of the graphics through an AI algorithm; In response to at least one graphic lacking annotation information, sending the graphic lacking annotation information to the user and stopping generating the CAD file; In response to all graphics not lacking annotation information, through an AI-based geometric analysis algorithm, identifying the relative positions between lines and the line directions; converting the results obtained through the AI algorithm and the geometric analysis algorithm into data recognizable by the corresponding drawing tool software, and then generating the corresponding CAD file through the drawing tool software; Through image processing technology, identifying the differences between the CAD file and the image, and then correcting the CAD file according to the differences through an AI model.

2. The method for generating a CAD file based on image recognition according to claim 1, wherein, Preprocessing the image includes: In response to the lines that are normally horizontal or vertical around the photo in the image being inclined, correcting the image to be upright through AI technology.

3. A method for generating a CAD file based on image recognition according to claim 1, characterized in that, Preprocessing the image further includes: Adjusting the brightness and contrast of the image, and denoising.

4. A method for generating a CAD file based on image recognition according to claim 1, characterized in that, Identifying the lines, shapes and annotation information of the graphics through an AI algorithm includes: Identifying the lines, shapes and annotation information of the graphics by using opencv and OCR technology.

5. A method for generating a CAD file based on image recognition according to claim 1, characterized in that The geometric analysis algorithm includes: an edge detection algorithm and a line matching algorithm.

6. A method for generating a CAD file based on image recognition according to claim 1, wherein, The image processing technology includes feature point matching.

7. A method for generating a CAD file based on image recognition according to claim 1, wherein The differences include incorrect annotation information, missing graphics and inaccurate line positions.

8. A method for generating a CAD file based on image recognition according to claim 1, characterized in that, Also comprising: Providing an interactive interface for the user to modify the correction of the AI model, and also providing an interactive interface for the user to adjust the CAD file.

9. An image recognition-based CAD file generation system, characterized in that Comprising: A processor and a memory, the memory storing computer program instructions, and when the computer program instructions are executed by the processor, implementing a method for generating a CAD file based on image recognition according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, Wherein program instructions are stored, and when the program instructions are loaded and executed by the processor, the processor is caused to execute a method for generating a CAD file based on image recognition according to any one of claims 1 to 8.

Images