Checking method for three-dimensional image geometric design and checking system for three-dimensional image geometric design
Through the three-dimensional image geometric design inspection method, grayscale processing and machine learning technology are used to automatically confirm whether the geometric design of electronic product parts complies with the specifications, solving the problem of time-consuming and error-prone manual inspection in the existing technology, and achieving efficient and accurate design inspection.
Patent Information
- Application Number
- CN202310673230.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-07
- Publication Date
- 2024-12-10
AI Technical Summary
The prior art requires manual execution in the three-dimensional image geometric design inspection of internal parts of electronic products, which leads to time-consuming and error-prone, and lacks efficient and intelligent inspection methods.
Using the three-dimensional image geometric design inspection method, by capturing the three-dimensional image of an object, performing grayscale processing and geometric feature identification, combining logical judgment and machine learning methods, it automatically confirms whether the design meets the geometric design conditions, generates design drawings and automatically notes. remember.
Fast and accurate geometric design inspections are achieved, reducing manual errors and improving design efficiency and accuracy.
Smart Images
Figure CN119106459A_ABST
Abstract
Description
Technical Field
[0001] The present invention refers to a method for checking three-dimensional image geometric design and an inspection system for three-dimensional image geometric design, and in particular, refers to a method for checking three-dimensional image geometric design that can achieve intelligent geometric design inspection and an inspection system for three-dimensional image geometric design. Background Art
[0002] The existing electronic products contain a large number of structural design criteria for internal parts during the design stage. When the three-dimensional (3D) image of the internal parts of the electronic products is designed, the key points of the part design are checked. The existing inspection process needs to be performed manually, such as opening the image, measuring and comparing. However, the manual inspection process is time-consuming and prone to errors during comparison.
[0003] Therefore, the prior art needs to be improved. Summary of the invention
[0004] Therefore, the present invention provides a method for checking three-dimensional image geometric design and a system for checking three-dimensional image geometric design to achieve intelligent geometric design inspection.
[0005] An embodiment of the present invention discloses a method for checking a three-dimensional image geometric design, comprising capturing a three-dimensional image of an object; locating a structural feature of the three-dimensional image of the object; capturing a top view corresponding to the located structural feature; and performing a geometric check on the top view to confirm whether the top view corresponding to the object meets a geometric design condition.
[0006] The step of capturing the top view corresponding to the located structural feature includes: performing a grayscale processing on the top view to obtain a grayscale top view; and performing a geometric feature image recognition on the grayscale top view.
[0007] The step of performing the geometric check on the top view to confirm whether the top view corresponding to the object meets the geometric design conditions includes: redrawing at least one geometric feature corresponding to the top view based on a geometric feature image recognition result of the grayscale top view; classifying the at least one redrawn geometric feature; and confirming whether the top view corresponding to the object meets the geometric design conditions based on a logical judgment method corresponding to the at least one geometric feature.
[0008] The step of performing the geometric check on the top view to confirm whether the top view corresponding to the object meets the geometric design conditions includes: generating a plurality of design drawings corresponding to the top view based on the top view; and performing an image annotation on each of the plurality of design drawings; wherein the plurality of design drawings are generated based on an associated instruction of a computer-aided design application; and wherein the image annotation of the plurality of design drawings is implemented by a parametric method.
[0009] The step of performing the geometric check on the top view to confirm whether the top view corresponding to the object meets the geometric design conditions includes: applying a machine learning method to perform a machine learning training on the multiple design drawings containing the image annotations; and confirming whether the top view corresponding to the object meets the geometric design conditions based on the machine learning method.
[0010] An embodiment of the present invention further discloses an inspection system for three-dimensional image geometric design, comprising a processing device; and a memory device coupled to the processing device, for storing a program code to instruct the processing device to execute an inspection process for inspecting a three-dimensional image geometric design, wherein the process includes: capturing a three-dimensional image of an object; locating a structural feature of the three-dimensional image of the object; capturing a top view corresponding to the located structural feature; and performing a geometric inspection on the top view to confirm whether the top view corresponding to the object meets a geometric design condition.
[0011] The step of capturing the top view corresponding to the located structural feature includes: performing a grayscale processing on the top view to obtain a grayscale top view; and performing a geometric feature image recognition on the grayscale top view.
[0012] The step of performing the geometric check on the top view to confirm whether the top view corresponding to the object meets the geometric design conditions includes: redrawing at least one geometric feature corresponding to the top view based on a geometric feature image recognition result of the grayscale top view; classifying the at least one redrawn geometric feature; and confirming whether the top view corresponding to the object meets the geometric design conditions based on a logical judgment method corresponding to the at least one geometric feature.
[0013] The step of performing the geometric check on the top view to confirm whether the top view corresponding to the object meets the geometric design conditions includes: generating a plurality of design drawings corresponding to the top view based on the top view; and performing an image annotation on each of the plurality of design drawings; wherein the plurality of design drawings are generated based on an associated instruction of a computer-aided design application; and wherein the image annotation of the plurality of design drawings is implemented by a parametric method.
[0014] The step of performing the geometric check on the top view to confirm whether the top view corresponding to the object meets the geometric design conditions includes: applying a machine learning method to perform a machine learning training on the multiple design drawings containing the image annotations; and confirming whether the top view corresponding to the object meets the geometric design conditions based on the machine learning method. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Figure 1 Schematic diagram of an inspection system for three-dimensional image geometry design according to an embodiment of the present invention.
[0016] Figure 2 A schematic diagram of an inspection process according to an embodiment of the present invention.
[0017] Figure 3 FIG. 4 is a schematic diagram of an image recognition process according to an embodiment of the present invention.
[0018] Figure 4 Schematic diagram of a logic judgment method according to an embodiment of the present invention.
[0019] Figure 5 FIG. 4 is a schematic diagram of an image recognition process according to an embodiment of the present invention.
[0020] Figure 6 It is a schematic diagram of a logic judgment method according to an embodiment of the present invention.
[0021] Figure 7 FIG. 4 is a schematic diagram of an image recognition process according to an embodiment of the present invention.
[0022] Figure 8 Schematic diagram of a logic judgment method according to an embodiment of the present invention.
[0023] Fig. 9 Schematic diagram of an artificial intelligence recognition method according to an embodiment of the present invention.
[0024] Fig.10 It is a schematic diagram of generating different design drawings according to a top view of an object according to an embodiment of the present invention.
[0025] Fig.11It is a schematic diagram of a top view of an embodiment of the present invention.
[0026] Fig.12 and Fig.13 A schematic diagram of automatic annotation and machine learning according to an embodiment of the present invention.
[0027] Fig.14 and Fig.15 Schematic diagram of a binary annotation training according to an embodiment of the present invention.
[0028] Explanation of symbols:
[0029] 10: Check the system
[0030] 102: Processing device
[0031] 104: Memory device
[0032] 20: Check the process
[0033] 202-210,302-306,402-420,502-506,602-610,702-706,802-810,902-916: Steps
[0034] 30,50,70: Image recognition process
[0035] 40,60,80: Logical judgment method
[0036] 90: Artificial Intelligence Identification Method
[0037] A: Top view
[0038] B,C,D,E,F,G: Design drawings
[0039] C1,C2,C3,C4,C1',C2',C3': round
[0040] P1, P2: Positioning column
[0041] x1,x2,x3,x4,x1',x2',x3': coordinate values
[0042] y1,y2,y3,y4,y1',y2',y3': coordinate values DETAILED DESCRIPTION
[0043] Please refer to Figure 1 , Figure 1Schematic diagram of an inspection system 10 for a three-dimensional image geometric design according to an embodiment of the present invention. The inspection system 10 includes a processing device 102 and a memory device 104. The memory device 104 is coupled to the processing device 102 and is used to store a program code to instruct the processing device 102 to execute an inspection process 20 for inspecting a three-dimensional image geometric design. The inspection process 20 can be used to inspect a part of an electronic product, such as a hinge bracket. The inspection process 20 includes the following steps:
[0044] Step 202: Start;
[0045] Step 204: Locate a structural feature of a three-dimensional image of an object;
[0046] Step 206: Capture a top view corresponding to the located structural feature;
[0047] Step 208: Perform a geometric check on the top view to confirm whether the top view corresponding to the object meets a geometric design specification;
[0048] Step 210: End.
[0049] The inspection system 10 can inspect whether the geometric design of the electronic product parts meets the geometric design specifications according to the inspection process 20. For example, when the electronic product parts are hinge brackets, the geometric design specifications at least include: two positioning columns and screws cannot be on the same horizontal axis (X axis), etc. The geometric design specifications can be formulated by designers or manufacturers to avoid failure to produce parts due to non-compliance with the geometric design principles during the design process.
[0050] Therefore, the inspection system 10 of the embodiment of the present invention can capture the three-dimensional image of the hinge bracket, locate the structural features of the three-dimensional image of the hinge bracket, and capture the top view of the located structural features according to the inspection process 20. Then, the inspection system 10 performs a geometric inspection on the top view of the hinge bracket, for example, a plurality of circular holes on the hinge bracket and the corresponding circular hole coordinates, to confirm whether the hinge bracket meets the geometric design specification according to the top view of the hinge bracket.
[0051] In one embodiment, the inspection system 10 can use Creo toolkit to open the corresponding drawing of the part, center the image of the part in the file and open the top view, and capture and store a minimum rectangle that can contain the drawing.
[0052] For further information, please refer to Figure 3 , Figure 3 FIG. 3 is a schematic diagram of an image recognition process 30 according to an embodiment of the present invention. Figure 3In the embodiment of the present invention, the inspection system 10 can read the stored top view image file through the Phyton Open CV library (step 302), and perform a grayscale processing on the top view to obtain a grayscale top view, and then the inspection system 10 performs geometric feature image recognition on the grayscale top view (step 304). In the above embodiment, grayscale processing on the top view can remove color noise that affects image recognition and accelerate the image recognition process.
[0053] Next, the inspection system 10 can detect the circular hole through the HoughCircles function in the Phyton Open CV library and redraw at least one geometric feature corresponding to the top view, that is, redraw Figure 3 A circle of the top view in FIG. 1 is drawn, and the radius and center coordinates of the circle are output (step 306 ).
[0054] exist Figure 3 In the embodiment of the present invention, the top view of the hinge bracket includes circles C1, C2, C3, C4 and positioning posts P1, P2, wherein circles C1, C2, C3, C4 are screw holes of the hinge bracket, and circles C1, C2, C3, C4 have corresponding center coordinates (x1, y1), (x2, y2), (x3, y3), (x4, y4) in an XY coordinate axis. The inspection system 10 classifies the redrawn geometric features (i.e., classifies the circular hole data), and confirms whether the top view corresponding to the hinge bracket meets the geometric design specification according to the logic judgment method corresponding to the geometric features.
[0055] Please refer to Figure 4 , Figure 4 FIG. 4 is a schematic diagram of a logic judgment method 40 according to an embodiment of the present invention. The logic judgment method 40 is a logic judgment method for classifying data of a circular hole. Figure 4 As shown, step 402 is used to determine whether the circular hole is a screw hole or a positioning column according to the size of the circular hole. When the circular hole has a larger radius and is determined to be a screw hole (step 404), it is further confirmed in step 408 whether the Y-axis values of each circle are the same. If the Y-axis values of each circle are the same (step 410), the logic judgment fails (indicating that the design structure of the hinge bracket is unstable); on the contrary, if the Y-axis values of each circle are different (step 412), the logic judgment passes (i.e., the screw holes of the hinge bracket are not on the same X-axis, indicating that the design structure is stable).
[0056] On the other hand, when the radius of the circular hole is small and is determined to be a positioning post (step 406), it is further confirmed in step 414 whether the number of positioning posts is 2. If the number of positioning posts is not equal to 2, the logic judgment fails (i.e., there are too few or too many positioning posts); on the contrary, if the number of positioning posts is equal to 2, the logic judgment passes. Finally, in step 420, the logic judgment result on the geometric design of the hinge bracket is output.
[0057] Please refer to Figure 5 , Figure 5 FIG. 5 is a schematic diagram of an image recognition process 50 according to an embodiment of the present invention. Figure 5 In the embodiment of the present invention, the inspection system 10 can read the stored top view image file of the part through the Phyton Open CV library (step 502), and grayscale the top view to obtain a grayscale top view, and then the inspection system 10 performs geometric feature image recognition on the grayscale top view (step 504).
[0058] exist Figure 5 In the embodiment of the present invention, the top view of the part includes circles C1, C2, and C3 as screw holes of the hinge bracket, positioning columns P1 and P2, and circles C1, C2, and C3 include corresponding center coordinates (x1, y1), (x2, y2), and (x3, y3) in the XY coordinate axis. The inspection system 10 classifies the redrawn geometric features (i.e., classifies the circular hole data), and confirms whether the top view corresponding to the part meets the geometric design specification according to the logic judgment method corresponding to the geometric features.
[0059] Please refer to Figure 6 , Figure 6 FIG. 6 is a schematic diagram of a logic judgment method 60 according to an embodiment of the present invention. The logic judgment method 60 is a logic judgment method for detecting the number of circular holes in a part. Figure 6 As shown, step 602 is used to detect the number of circular holes and confirm whether the number of circular holes is greater than or equal to 2 (step 604). If the number of circular holes is less than 2, the logic judgment is that the design structure of this part does not pass (step 606); on the contrary, if the number of circular holes is greater than or equal to 2 (step 608), the logic judgment is passed (indicating that the design structure is stable).
[0060] Please refer to Figure 7 , Figure 7 FIG. 7 is a schematic diagram of an image recognition process 70 according to an embodiment of the present invention. Figure 7 In the embodiment of the present invention, the inspection system 10 can read the stored top view image file of the part through the Phyton Open CV library (step 702), and grayscale process the top view to obtain a grayscale top view, and then the inspection system 10 performs geometric feature image recognition on the grayscale top view (step 704).
[0061] exist Figure 7 In the example of , the inspection system 10 can compare three screw holes at a time (or can compare one at a time, taking the screw as the positioning center for capturing images). Figure 7 As shown, step 704 includes identifying circles and outputting drawings of circle center coordinates, and in step 706, the circles and circle center coordinates of the two drawings are marked, namely, circles C1, C2, C3 and their corresponding circle center coordinates (x1, y1), (x2, y2), (x3, y3), and circles C1', C2', C3' and their corresponding circle center coordinates (x1', y1'), (x2', y2'), (x3', y3'), and then further confirming whether the top view corresponding to the part meets the geometric design specifications.
[0062] Please refer to Figure 8 , Figure 8 FIG. 8 is a schematic diagram of a logic judgment method 80 according to an embodiment of the present invention. The logic judgment method 80 is used to judge whether a part is eccentric. Figure 8 As shown, step 802 is used to compare the center coordinates of the circles in the two drawings of step 706 to confirm whether the centers of the circles are consistent (step 804). If the centers of the circles in the two drawings are inconsistent, the logical judgment is that the design structure of this part fails (step 806); on the contrary, if the centers of the circles in the two drawings are consistent (step 808), the logical judgment is passed (indicating that the design structure is stable).
[0063] Different from the above-mentioned method of using image to identify the geometric design in the top view, the inspection system 10 can generate a plurality of design drawings corresponding to the top view according to the top view of the part; and perform an image annotation on each of the plurality of design drawings, wherein the design drawing is generated according to an association instruction of a computer-aided design application, such as a Creo relation instruction, and the image annotation of the design drawing is implemented by a parametric method, such as a Creo flexible modeling search and toolkit syntax. In this way, the inspection system 10 of the embodiment of the present invention can realize the inspection method of geometric design by automatically annotating the design drawing and applying the features of the geometric structure required for artificial intelligence (AI) training. In other words, the inspection system 10 of the embodiment of the present invention can apply a machine learning method to perform a machine learning training on the design drawing containing the image annotation, and confirm whether the top view corresponding to the object meets the geometric design conditions according to the machine learning method.
[0064] For example, see Fig. 9 , Fig. 9FIG. 9 is a schematic diagram of an artificial intelligence recognition method 90 according to an embodiment of the present invention. The artificial intelligence recognition method 90 can be used to check the geometric design of the hinge bracket, and includes the following steps:
[0065] Step 902: Start;
[0066] Step 904: Use AI to identify whether the screw holes of the hinge bracket are located on the same plane. If so, execute step 906; if not, execute step 908;
[0067] Step 906: Geometry check fails;
[0068] Step 908: Geometry check passed;
[0069] Step 910: Use AI to identify whether the number of the positioning posts of the hinge bracket is 2. If yes, execute step 914; if no, execute step 912;
[0070] Step 912: Geometry check fails;
[0071] Step 914: Geometry check passed;
[0072] Step 916: Output the inspection result.
[0073] It is worth noting that the inspection system 10 executes the artificial intelligence identification method 90 according to an AI training database to determine whether the inspected part (such as the hinge bracket) meets the geometric design conditions.
[0074] Please refer to Fig.10 , Fig.10 This is a schematic diagram of generating different design drawings according to the top view of an object according to an embodiment of the present invention. Fig.10 It can be seen that the inspection system 10 of the embodiment of the present invention can generate a design drawing BG different from the top view A according to the top view A of the part using the Creo associated command, wherein the design drawing BG are all parts related to the top view A.
[0075] In addition, please refer to Fig.11 , Fig.11 Schematic diagram of the top view of the embodiment of the present invention. Fig.11 It can be seen that the inspection system 10 of the embodiment of the present invention can automatically annotate the required circular hole features according to the AI training database. For example, according to the Creo flexible modeling search (Flexible modeling search) and toolkit syntax, the circular hole features can be annotated. Fig.11 The screw holes and positioning columns in the
[0076] Through the above Fig.10 and Fig.11The inspection system 10 of the embodiment of the present invention can use a machine learning method, such as Convolutional Neural Networks (CNN), and establish an AI training database corresponding to the number of positioning columns, the number of screw holes, and whether the screw holes are horizontal, etc. contained in different drawings.
[0077] Please refer to Fig.12 and Fig.13 , Fig.12 and Fig.13 Schematic diagram of automatic annotation and machine learning in an embodiment of the present invention. Fig.12 In the example, the inspection system 10 can perform grayscale processing on screw holes of different forms, and generate a large number of different design files for AI training according to the size and corresponding rules based on Creo Flexible Modeling Search and toolkit syntax.
[0078] Next, the inspection system 10 can use the machine learning method to Fig.13 The design files in the binary method are annotated and trained as a non-biased category (such as Fig.14 as shown) and an eccentric category (as Fig.15 As shown). In this way, the inspection system 10 applies the AI training database and the artificial intelligence recognition method 90 for AI training to perform a geometric structure inspection on whether the screw hole is eccentric on the captured top view.
[0079] It is worth noting that the above embodiments describe the concepts of the present invention, and those skilled in the art may make appropriate modifications accordingly and are not limited thereto. For example, geometric design specifications, tools for loading three-dimensional image files, machine learning methods, and modeling methods are not limited to the above examples, but can be adjusted according to user needs and are all within the scope of the present invention.
[0080] In summary, the embodiments of the present invention provide a method for inspecting three-dimensional image geometric design and an inspection system for three-dimensional image geometric design, which achieve intelligent geometric design inspection through a combination of image recognition and logical judgment, as well as automatic geometric annotation with machine learning.
[0081] The above description is only a preferred embodiment of the present invention. All equivalent changes and modifications made according to the claims of the present invention should fall within the scope of the present invention.
Claims
1. A method for checking three-dimensional image geometric design, It is characterized in that Contains: capturing a three-dimensional image of an object; Locating a structural feature of the three-dimensional image of the object; capturing a top view corresponding to the located structural feature; as well as A geometric check is performed on the top view to confirm whether the top view corresponding to the object meets a geometric design condition.
2. The method for checking the geometric design of a three-dimensional image according to claim 1, It is characterized in that The step of capturing the top view corresponding to the located structural feature comprises: Performing a grayscale processing on the top view to obtain a grayscale top view; and A geometric feature image recognition is performed on the grayscale top view.
3. The method for checking the geometric design of a three-dimensional image according to claim 2, It is characterized in that The step of performing the geometric inspection on the top view to confirm whether the top view corresponding to the object meets the geometric design condition comprises: Redrawing at least one geometric feature corresponding to the top view according to a geometric feature image recognition result of the grayscale top view; classifying the at least one geometric feature of the redrawn image; as well as According to a logic judgment method corresponding to the at least one geometric feature, it is confirmed whether the top view corresponding to the object meets the geometric design condition.
4. The method for checking a three-dimensional image geometric design according to claim 1, It is characterized in that The step of performing the geometric inspection on the top view to confirm whether the top view corresponding to the object meets the geometric design condition comprises: According to the top view, generating a plurality of design drawings corresponding to the top view; and Performing an image annotation on each of the plurality of design drawings; wherein the plurality of design drawings are generated according to an associated instruction of a computer-aided design application program; The image annotation of the plurality of design drawings is implemented by a parameterized method.
5. The method for checking the geometric design of a three-dimensional image according to claim 4, It is characterized in that The step of performing the geometric inspection on the top view to confirm whether the top view corresponding to the object meets the geometric design condition comprises: Applying a machine learning method to perform a machine learning training on the plurality of design drawings including the image annotations; as well as According to the machine learning method, it is confirmed whether the top view corresponding to the object meets the geometric design conditions.
6. An inspection system for three-dimensional image geometry design, It is characterized in that Contains: a processing device; as well as A memory device, coupled to the processing device, for storing a program code to instruct the processing device to execute a check process of checking a three-dimensional image geometric design, wherein the process includes: capturing a three-dimensional image of an object; Locating a structural feature of the three-dimensional image of the object; capturing a top view corresponding to the located structural feature; as well as A geometric check is performed on the top view to confirm whether the top view corresponding to the object meets a geometric design condition.
7. The inspection system for three-dimensional image geometric design according to claim 6, It is characterized in that The step of capturing the top view corresponding to the located structural feature comprises: Performing a grayscale processing on the top view to obtain a grayscale top view; and A geometric feature image recognition is performed on the grayscale top view.
8. The inspection system for three-dimensional image geometric design according to claim 7, It is characterized in that The step of performing the geometric inspection on the top view to confirm whether the top view corresponding to the object meets the geometric design condition comprises: Redrawing at least one geometric feature corresponding to the top view according to a geometric feature image recognition result of the grayscale top view; classifying the at least one geometric feature of the redrawn image; as well as According to a logic judgment method corresponding to the at least one geometric feature, it is confirmed whether the top view corresponding to the object meets the geometric design condition.
9. The inspection system for three-dimensional image geometric design according to claim 6, It is characterized in that The step of performing the geometric inspection on the top view to confirm whether the top view corresponding to the object meets the geometric design condition comprises: According to the top view, generating a plurality of design drawings corresponding to the top view; and Performing an image annotation on each of the plurality of design drawings; wherein the plurality of design drawings are generated according to an associated instruction of a computer-aided design application program; The image annotation of the plurality of design drawings is implemented by a parameterized method.
10. The inspection system for three-dimensional image geometric design according to claim 9, It is characterized in that The step of performing the geometric inspection on the top view to confirm whether the top view corresponding to the object meets the geometric design condition comprises: Applying a machine learning method to perform a machine learning training on the plurality of design drawings including the image annotations; as well as According to the machine learning method, it is confirmed whether the top view corresponding to the object meets the geometric design conditions.