A computer vision-based core wire cutting method and system

Abstract

The application discloses a core linear cutting method based on computer vision, which comprises the following steps: collecting the image of a core to be cut, and establishing a three-dimensional model of the core through the image of the core; manufacturing a three-dimensional boundary model of a plunger sample with a corresponding size; displaying the three-dimensional model of the core and the three-dimensional boundary model of the plunger sample in the same three-dimensional coordinate system, and screening an optimal cutting position; moving the three-dimensional boundary model of the plunger sample to the optimal cutting position, and generating a cutting path; and cutting the core according to the cutting path. The technical scheme has the beneficial effects that the cutting effect can be directly displayed by first modeling the core to be cut in three dimensions and then cutting the three-dimensional model of the core through the three-dimensional boundary model of the plunger sample, the cut core plunger can meet the standard, the damage of the core plunger to the whole core can be reduced, and the utilization value of the remaining core can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of core processing, in particular to a core wire cutting method and system based on computer vision. BACKGROUND

[0002] After a full core sample is drilled, the full core is cut to obtain a plurality of samples (optimal sample diameter of 6-8mm) suitable for micron CT displacement scanning specifications. Since the sample specifications are strict, accurate positioning is required during cutting.

[0003] Currently, in the process of cutting small plunger samples, wire cutting (such as Chinese patent application No. CN201510975780.6) or sample drilling machine is generally used to drill samples. In practice, the core cutting focus is placed on the cut plunger part, and the morphology of the remaining core is ignored. The remaining core is damaged due to the integrity of the shape, and the utilization value is greatly reduced. At the same time, when positioning the cutting position, the operator relies on visual observation and experience. Even if there is a clear indication of the core cutting part, it is difficult to intuitively design the cutting. This leads to inaccurate cutting of the target sample and waste of the remaining core sample. SUMMARY

[0004] Therefore, it is necessary to provide a core wire cutting method and system based on computer vision to accurately position the core cutting position during wire cutting, achieve the expected rock sample preparation target, reduce the damage of core cutting to the whole core, and improve the utilization value of the remaining core.

[0005] To achieve the above purpose, the present application provides a core wire cutting method based on computer vision, comprising:

[0006] S1, acquiring an image of a core to be cut, and establishing a core three-dimensional model through the image of the core;

[0007] S2, according to the actual experimental requirements, a plunger sample three-dimensional boundary model of corresponding size is made;

[0008] S3, the core three-dimensional model and the plunger sample three-dimensional boundary model are displayed in the same three-dimensional coordinate system, the plunger sample three-dimensional boundary model is moved to different cutting positions in the core three-dimensional model, the core three-dimensional model is cut through the plunger sample three-dimensional boundary model, and each different position cutting plunger sample three-dimensional model and the remaining core three-dimensional model are generated. Compare each plunger sample three-dimensional model and each remaining core three-dimensional model to screen out the best cutting position;

[0009] S4, moving the plunger sample three-dimensional boundary model to the optimal cutting position, and projecting the plunger sample three-dimensional boundary model to a horizontal plane to generate a cutting path;

[0010] S5, cutting the core according to the cutting path.

[0011] In some embodiments, in step S1, the image of the core to be cut is collected, and the three-dimensional model of the core is established by the image of the core, specifically comprising:

[0012] S11, collecting the image of the core to be cut;

[0013] S12, removing the background in the image in the image of the core;

[0014] S13, establishing a three-dimensional model of the core by a plurality of core images after removing the background.

[0015] In some embodiments, in step S12, the background in the image is removed in the image of the core, specifically comprising the following steps:

[0016] S121, selecting a plurality of core images, pre-processing each core image, and marking the position information of the core on the core image;

[0017] S122, establishing a U-net neural network detection model;

[0018] S123, training the U-net neural network detection model by the marked core image set;

[0019] S124, importing the core image of the core to be cut into the trained U-net neural network detection model for core detection, and cutting each core image into an image containing only the core.

[0020] In some embodiments, in step S13, the three-dimensional model of the core is established by a plurality of core images after removing the background, specifically:

[0021] After generating a 3D model by using monocular vision 3D modeling technology on a plurality of core images after removing the background, the 3D model is imported into 3DS MAX software to establish a three-dimensional model of the core.

[0022] In some embodiments, in step S4, the plunger sample three-dimensional boundary model is projected to a horizontal plane to generate a cutting path, specifically:

[0023] The plunger sample three-dimensional boundary model is projected to a horizontal plane, and a cutting path cad drawing is obtained according to the projection of the plunger sample three-dimensional boundary model on the horizontal plane.

[0024] In some embodiments, step S5, cutting the core according to the cutting path, specifically includes:

[0025] S51. Generate 3B codes that can be recognized by wire cutting equipment based on the cutting path CAD drawing;

[0026] S52. Fix the core to the wire cutting equipment, import the 3B code into the wire cutting equipment, and cut the core.

[0027] This invention also provides a computer vision-based core wire cutting system, comprising:

[0028] The core modeling module is used to acquire images of the core to be cut and to build a three-dimensional model of the core based on the images.

[0029] The plunger size design module is used to create a three-dimensional boundary model of a plunger sample of the appropriate size according to actual experimental requirements.

[0030] The wire cutting design module is used to display the core 3D model and the plunger sample 3D boundary model in the same 3D coordinate system, move the plunger sample 3D boundary model to different cutting positions in the core 3D model, cut the core 3D model through the plunger sample 3D boundary model, generate the plunger sample 3D model after cutting at each different position and the remaining core 3D model, and compare each plunger sample 3D model and each remaining core 3D model to select the optimal cutting position;

[0031] The wire cutting path generation module is used to move the three-dimensional boundary model of the plunger sample to the optimal cutting position and project the three-dimensional boundary model of the plunger sample onto a horizontal plane to generate a cutting path.

[0032] A wire cutting device used to cut the core along the cutting path.

[0033] In some embodiments, the core modeling module includes a core image acquisition mechanism, which includes a turntable, a rotation drive, and a camera. The turntable is used to place the core, the rotation drive is connected to the turntable and is used to drive the turntable to rotate, and the camera is used to capture images of the core on the turntable.

[0034] In some embodiments, the rotation drive includes a rotation drive motor and a reducer, wherein the output shaft of the rotation drive motor is connected to the input end of the reducer, and the output end of the reducer is connected to the turntable.

[0035] In some embodiments, the turntable is padded with anti-reflective paper.

[0036] Compared with the prior art, the technical scheme provided by the present application has the beneficial effects that: by first modeling the core to be cut in three dimensions, and then cutting the core three-dimensional model through the plunger sample three-dimensional boundary model, the cutting effect can be intuitively displayed, the specification of the cut core plunger can be guaranteed to meet the standard, the damage of the core cutting to the whole core can be reduced, and the utilization value of the remaining core can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 is a flowchart of an embodiment of the core linear cutting method based on computer vision provided by the present application;

[0038] Figure 2 is a flowchart of step S1 in Figure 1

[0039] Figure 3 is a part of the core image photographed in an embodiment of the present application;

[0040] Figure 4 is a flowchart of step S12 in Figure 2

[0041] Figure 5 is a schematic diagram of marking the original image of the core;

[0042] Figure 6 is the image containing only the core formed after cutting each core image in Figure 3

[0043] Figure 7 is a three-dimensional model of the core established after importing each image in Figure 6

[0044] Figure 8 is a schematic diagram of moving the plunger sample three-dimensional boundary model to a certain cutting position in the core three-dimensional model;

[0045] Figure 9 is a flowchart of step S5 in Figure 1

[0046] Figure 10 is a cutting path cad diagram in an embodiment of the present application;

[0047] Figure 11 is a schematic diagram of the three-dimensional structure of the core image acquisition mechanism in an embodiment of the present application;

[0048] In the figure: 100-core image acquisition mechanism, 110-rotating disc, 120-rotating driving part, 121-rotating driving motor, 122-reducing part. DETAILED DESCRIPTION​​​​​

[0049] Preferred embodiments of the present application are described in detail below with reference to the attached drawing figures, wherein the drawing figures are a part of this specification, and wherein the attached drawing figures together with the present specification illustrate, explain, and enable the present application.

[0050] Referring to Figure 1 , the present application provides a core linear cutting method based on computer vision, comprising:

[0051] S1, collecting images of a core to be cut, and establishing a three-dimensional core model through the images of the core;

[0052] Referring to Figure 2 , in step S1, collecting images of a core to be cut, and establishing a three-dimensional core model through the images of the core specifically comprises:

[0053] S11, collecting images of a core to be cut;

[0054] When collecting core images, as many original pictures as possible are collected at different angles. In the present embodiment, some of the original pictures collected are as shown in Figure 3 .

[0055] S12, removing the background in the images from the images of the core;

[0056] Referring to Figure 4 , in step S12, removing the background in the images from the images of the core specifically comprises the following steps:

[0057] S121, selecting a plurality of core images, pre-processing each core image, and marking the position information of the core on the core image. In the present embodiment, the Labelme tool is used to mark the boundary of the core in each core image (as shown in Figure 5 ).

[0058] S122, establishing a U-net neural network detection model;

[0059] S123, training the U-net neural network detection model through the marked core image set. The model training has been completed before the collection of the core photos to be cut, and the model is trained through a large number of previous core photos, so that the model has good recognition for new core photos.

[0060] S124, importing the core images of the core to be cut into the trained U-net neural network detection model for core detection, so as to identify the boundary of the core in each core image, and then cutting each core image into an image containing only the core (as shown in Figure 6 ).

[0061] S13, a three-dimensional model of the core is established by using a plurality of core images after background removal. In this embodiment, the three-dimensional model is generated by using monocular vision 3D modeling technology, and then the three-dimensional model is imported into 3DS MAX software to establish a three-dimensional model of the core (as shown in Figure 7 ).

[0062] S2, a three-dimensional boundary model of the plunger sample with a corresponding size is prepared according to actual experimental requirements.

[0063] S3, the core three-dimensional model and the plunger sample three-dimensional boundary model are displayed in the same three-dimensional coordinate system, and the plunger sample three-dimensional boundary model is moved to different cutting positions in the core three-dimensional model (as shown in Figure 8 ), the core three-dimensional model is cut by the plunger sample three-dimensional boundary model to generate a plurality of plunger sample three-dimensional models after cutting at different positions and a plurality of remaining core three-dimensional models, and the plurality of plunger sample three-dimensional models and the plurality of remaining core three-dimensional models are compared to screen out the best cutting position. In this embodiment, when comparing the plurality of plunger sample three-dimensional models, a plunger sample three-dimensional model with a relatively complete structure should be selected, and when comparing the plurality of remaining core three-dimensional models, a remaining core three-dimensional model with a larger volume of open hole part should be selected, so as to facilitate subsequent use of the remaining core.

[0064] S4, the plunger sample three-dimensional boundary model is moved to the best cutting position, and the plunger sample three-dimensional boundary model is projected onto a horizontal plane to generate a cutting path.

[0065] Specifically, the plunger sample three-dimensional boundary model is projected onto a horizontal plane, and a cutting path cad drawing is obtained according to the projection of the plunger sample three-dimensional boundary model on the horizontal plane.

[0066] S5, the core is cut according to the cutting path.

[0067] Please refer to Figure 9 , in the step S5, the core is cut according to the cutting path, which specifically includes:

[0068] S51, a 3B code that can be recognized by a wire cutting device is generated according to the cutting path cad drawing.

[0069] In this embodiment, Figure 10 the 3B code corresponding to the cutting path cad drawing is as follows:

[0070]

[0071] S52, fixing the core on the wire cutting device, importing the 3B code into the wire cutting device, cutting the core, and achieving the expected rock sample preparation target.

[0072] The application further provides a core wire cutting system based on computer vision, comprising:

[0073] a core modeling module for collecting images of a core to be cut and establishing a three-dimensional model of the core through the images of the core;

[0074] a plunger size design module for making a three-dimensional boundary model of a plunger sample with a corresponding size according to actual experimental requirements;

[0075] a wire cutting design module for displaying the three-dimensional model of the core and the three-dimensional boundary model of the plunger sample in a same three-dimensional coordinate system, moving the three-dimensional boundary model of the plunger sample to different cutting positions in the three-dimensional model of the core, cutting the three-dimensional model of the core through the three-dimensional boundary model of the plunger sample, generating three-dimensional models of the plunger sample after cutting at different positions and three-dimensional models of the remaining core, and comparing the three-dimensional models of the plunger sample and the three-dimensional models of the remaining core to screen out a best cutting position;

[0076] a wire cutting path generation module for moving the three-dimensional boundary model of the plunger sample to the best cutting position and projecting the three-dimensional boundary model of the plunger sample to a horizontal plane to generate a cutting path;

[0077] a wire cutting device for cutting the core according to the cutting path.

[0078] For the convenience of collecting core images at different angles, please refer to Figure 11 In a preferred embodiment, the core modeling module comprises a core image collection mechanism 100, which comprises a turntable 110, a rotating driving member 120 and a camera (not shown), the core is placed on the turntable 110, the rotating driving member 120 is connected with the turntable 110 and is used to drive the turntable 110 to rotate, and the camera is used to shoot the core on the turntable 110.

[0079] In a preferred embodiment, please refer to Figure 11 The rotating driving member 120 comprises a rotating driving motor 121 and a speed reducer 122, the output shaft of the rotating driving motor 121 is connected with the input end of the speed reducer 122, and the output end of the speed reducer 122 is connected with the turntable 110. In this embodiment, the speed reducer 122 is a speed reducer.

[0080] For the convenience of reducing reflection when shooting, please refer to Figure 3 andFigure 6 In a preferred embodiment, the turntable 110 is provided with anti-glare paper, and a regular pattern is drawn on the anti-glare paper, which facilitates subsequent removal of the background in the core photo through an algorithm.

[0081] In summary, the present application can directly show the cutting effect by first modeling the core to be cut in three dimensions, and then cutting the core three-dimensional model through the plunger sample three-dimensional boundary model, so as to ensure that the cut core plunger specification meets the standard, reduce the damage of core cutting to the whole core, and improve the utilization value of the remaining core.

[0082] The above merely describes the preferred embodiments of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A computer vision-based method of core wireline cutting, characterized by, The method comprises the following steps: S1, collecting an image of a core to be cut, and establishing a three-dimensional model of the core through the image of the core; S2, according to actual experimental requirements, a three-dimensional boundary model of a plunger sample of a corresponding size is made; S3, the three-dimensional model of the core and the three-dimensional boundary model of the plunger sample are displayed in the same three-dimensional coordinate system, the three-dimensional boundary model of the plunger sample is moved to different cutting positions in the three-dimensional model of the core, the three-dimensional model of the core is cut through the three-dimensional boundary model of the plunger sample, three-dimensional models of the plunger sample after cutting at different positions and three-dimensional models of the remaining core are generated, and the three-dimensional models of the plunger sample and the three-dimensional models of the remaining core are compared to screen out the best cutting position; S4, the three-dimensional boundary model of the plunger sample is moved to the best cutting position, and the three-dimensional boundary model of the plunger sample is projected onto a horizontal plane to generate a cutting path; S5, the core is cut according to the cutting path.

2. The computer vision-based core wireline cutting method of claim 1, wherein, In step S1, the image of the core to be cut is collected, and the three-dimensional model of the core is established through the image of the core, specifically comprising the following steps: S11, collecting an image of a core to be cut; S12, removing the background in the image in the image of the core; S13, establishing a three-dimensional model of the core through a plurality of core images after removing the background.

3. The computer vision-based core wire sawing method of claim 2, wherein, In step S12, the background in the image of the core is removed, specifically comprising the following steps: S121, selecting a plurality of core images, pre-processing each core image, and marking the position information of the core on the core image; S122, establishing a U-net neural network detection model; S123, training the U-net neural network detection model through the marked core image set; S124, importing the core image of the core to be cut into the trained U-net neural network detection model for core detection, and cutting each core image into an image containing only the core.

4. The computer vision-based core wire sawing method of claim 2, wherein, In step S13, the three-dimensional model of the core is established through a plurality of core images after removing the background, specifically comprising: After generating a 3D model using monocular vision 3D modeling technology on the plurality of core images after removing the background, the 3D model is imported into 3DS MAX software to establish a three-dimensional model of the core.

5. The computer vision-based core wire sawing method of claim 1, wherein, In step S4, the three-dimensional boundary model of the plunger sample is projected onto a horizontal plane to generate a cutting path, specifically comprising: The three-dimensional boundary model of the plunger sample is projected onto a horizontal plane, and a cutting path cad drawing is obtained according to the projection of the three-dimensional boundary model of the plunger sample on the horizontal plane.

6. The computer vision-based core wire sawing method of claim 5, wherein, In step S5, the core is cut according to the cutting path, specifically comprising: S51, generating a 3B code recognizable by a wire cutting device according to the cutting path cad drawing; S52, fixing the core on the wire cutting device, importing the 3B code into the wire cutting device, and cutting the core.

7. A computer vision-based core wire sawing system, characterized by, The method comprises the following steps: A core modeling module is used to collect an image of a core to be cut, and a three-dimensional model of the core is established through the image of the core; A plunger size design module is used to make a three-dimensional boundary model of a plunger sample of a corresponding size according to actual experimental requirements; The line cutting design module is used to display the core three-dimensional model and the plunger sample three-dimensional boundary model in the same three-dimensional coordinate system, move the plunger sample three-dimensional boundary model to different cutting positions in the core three-dimensional model, cut the core three-dimensional model by the plunger sample three-dimensional boundary model, generate the plunger sample three-dimensional models after cutting at different positions and the remaining core three-dimensional models, and compare the plunger sample three-dimensional models and the remaining core three-dimensional models to screen the best cutting position. The line cutting path generation module is used to move the plunger sample three-dimensional boundary model to the best cutting position, project the plunger sample three-dimensional boundary model to a horizontal plane, and generate a cutting path. The line cutting equipment is used to cut the core according to the cutting path.

8. The computer vision-based core wire sawing system of claim 7, wherein, The core modeling module comprises a core image acquisition mechanism, the core image acquisition mechanism comprising a turntable, a rotating driving member and a camera, the turntable being used to place the core, the rotating driving member being connected with the turntable and being used to drive the turntable to rotate, and the camera being used to shoot the core on the turntable.

9. The computer vision-based core wire sawing system of claim 8, wherein, The rotating driving member comprises a rotating driving motor and a speed reduction member, an output shaft of the rotating driving motor being connected with an input end of the speed reduction member, and an output end of the speed reduction member being connected with the turntable.

10. The computer vision-based core line cutting system according to claim 8, characterized in that the turntable is provided with anti-glare paper.

Images