1. A tunnel boring machine characterized in that: comprises a crawler chassis (1), a control cabinet (2) and a movable base body (6) which are sequentially arranged from the ground to the top, wherein the control cabinet (2) is fixedly connected with the top of the crawler chassis (1), a sliding mechanism is arranged between the movable base body (6) and the control cabinet (2), the sliding mechanism comprises a first guide rail (3) arranged at the top of the control cabinet (2), a second guide rail (4) which is vertical to the first guide rail (3) is connected on the first guide rail (3) in a sliding way, the lower part of the second guide rail (4) is connected with a first lead screw (5) which is parallel to the first guide rail (3), the end part of the first lead screw (5) is connected with a first motor (11) which is used for driving the first lead screw (5) to rotate, the lower part of the movable base body (6) is connected with a second lead screw (7) which is parallel to the second guide rail (4), the end connection of second lead screw (7) has and is used for driving second lead screw (7) pivoted second motor (12), the upper portion of activity pedestal (6) is provided with swivel plate (8), be provided with in the activity pedestal (6) and drive swivel plate (8) pivoted third motor (13), the end rotation of swivel plate (8) is connected with flexible arm (9), the end connection of flexible arm (9) has drill bit mechanism (10), be provided with visual system (15) that are used for discerning the location drilling in drill bit mechanism (10).

2. A tunnel boring machine according to claim 1, wherein: the drill bit mechanism (10) comprises a support frame (10-1) and a drill bit (10-2), one end of the support frame (10-1) is movably connected with a telescopic arm (9), a drill bit guide rail (10-3) is arranged at the other end of the support frame (10-1), a mounting plate (10-4) used for mounting the drill bit (10-2) is connected onto the drill bit guide rail (10-3) in a sliding mode, a drill bit lead screw (10-5) parallel to the drill bit guide rail (10-3) is connected to the lower portion of the mounting plate (10-4), and the end portion of the drill bit lead screw (10-5) is connected with a servo motor (14) used for driving the drill bit lead screw (10-5) to rotate.

3. A tunnel boring machine according to claim 2, wherein: and a hinged rod (16) and a hydraulic cylinder (17) for adjusting the inclination angle of the support frame (10-1) are connected between the support frame (10-1) and the telescopic arm (9).

4. A tunnel boring machine according to claim 2, wherein: be provided with oil pressure station, generator and controller in the control cabinet (2), first motor (11), second motor (12), third motor (13) and servo motor (14) all are connected with the output of controller, visual system (15) are connected with the input of controller.

5. A method of construction of a tunnel boring machine, characterized in that a tunnel boring machine according to claims 1-4 is used, which method comprises the steps of:

moving to a tunnel drilling operation working area through the crawler chassis (1);

secondly, the third motor (13) drives the rotary disk seat (8) to rotate, so that the drill bit mechanism (10) rotates to an approximate working angle;

thirdly, the drill bit mechanism (10) is extended to an approximate working height under the action of the telescopic arm (9);

fourthly, a vision system (15) in the drill bit mechanism (10) identifies and positions the target drill hole;

driving the movable seat body (6) to move through the sliding mechanism, and further adjusting the position of the drill bit (10-2);

sixthly, driving the drill bit (10-2) to move through the servo motor (14) to enable the drill bit (10-2) to be aligned with the target drilling hole;

seventhly, adjusting the inclination angle of the drill bit (10-2) through a hydraulic cylinder (17) to enable the drill bit (10-2) to be vertical to the surface of the target drilling hole;

and step eight, the drill bit (10-2) works to perform drilling operation.

6. The method for constructing a tunnel boring machine according to claim 5, wherein the step four in which the vision system (15) in the drill head mechanism (10) identifies and locates the target borehole comprises:

step 401, performing edge detection on a target drilling mark image;

step 402, carrying out color detection on the target drilling hole mark image;

and 403, precisely positioning the target drilling mark image by adopting a depth convolution neural network.

7. The method as claimed in claim 6, wherein the step 401 of edge detection of the target hole-drilling mark image comprises:

step 40101, filtering the target borehole marking image by using an optimized median filtering algorithm;

the optimized median filtering algorithm is judged by adding noise points into the median filtering algorithm, and the noise points in a window are removed when the median is obtained, and the specific process comprises the following steps:

noise point determination

Wherein, X_ijIs a point to be judged in the window, S is a signal point, N is a noise point, X_minIs the minimum value, X, within the window_maxIs the maximum value in the window, D is the set threshold;

noise point removal

Wherein, X_medTo remove the median value after the noise point in the window, [ X ]_ij]Is the complete set of points in the window, [ X ]_S]A noise point set in a window is defined, p is the number of all points in the window, and q is the number of noise points in the window;

step 40102, performing edge enhancement on the target borehole marking image by adopting a gradient amplitude value;

step 40103, edge position detection is carried out on the target borehole marking image by adopting a Canny operator;

smoothing the target borehole marking image with a Gaussian filter; calculating the magnitude and direction of the gradient by using the finite difference of the first-order partial derivatives; carrying out non-maximum suppression on the gradient amplitude; edges are detected and connected using a dual threshold algorithm.

8. The tunnel boring machine construction method according to claim 6, wherein the specific process of performing color detection on the target boring mark image in step 402 comprises:

step 40201, converting the target drilling mark image from the RGB space to the HSV space;

40202, setting fuzzy ranges of hue components H, saturation components S and brightness components V in an HSV space;

40203, processing the target drilling hole mark image in the HSV space into a binary image according to the fuzzy range;

40204, drawing and connecting the periphery of each connected region in the binary image;

step 40205, screening the connected region according to the length-width ratio and the area of the drilling mark to obtain a candidate region image of the drilling mark;

step 40206, segmenting the candidate region image from the target borehole marker image, so as to facilitate subsequent accurate identification and positioning.

9. The method as claimed in claim 8, wherein the step 403 of precisely positioning the target hole mark image using the deep convolutional neural network comprises:

40301, constructing a depth convolution neural network structure for accurately positioning a target borehole marking image;

the deep convolutional neural network structure comprises three levels of cascaded convolutional neural networks CNN;

the first layer comprised 1 CNN with 39 x 39 inputs;

the second layer comprises 2 CNNs, the input sizes are all 15 x 15, every two CNNs form a pair, a pair of CNNs predict an anchor point, and the prediction results are averaged;

the third layer comprises 2 CNNs, the input sizes are all 15 x 15, and every two CNNs form a pair;

step 40302, training a deep convolutional neural network;

selecting a plurality of candidate area images as a training sample set, marking drilled holes or non-drilled holes, and inputting a deep convolutional neural network for training;

40303, accurately positioning the target borehole marker image by using a depth convolution neural network;

and inputting the candidate area image to perform target drilling mark identification and positioning to obtain a positioning and identification result of drilling or non-drilling.

Background

With the rapid development of Chinese economy, the urbanization process is accelerated continuously, the living standard of people is improved continuously, the traveling mode of people is changed greatly, rail transit becomes one of important choices for traveling of people, and due to the rapidness and economy of the rail transit, the investment of the nation is increased continuously to meet the requirements of people, and more rails are waiting for construction. In order to complete construction more quickly and better, rail transit practitioners continuously try to improve the working efficiency and complete construction tasks on schedule, but as a part of rail construction, a drilling construction process in tunnel construction is still a relatively backward manual operation mode, and manual operation has multiple defects: labor intensive, operator fatigue, inability to continue drilling at the same rate, resulting in inefficiencies: the manual operation is not consistent in the angle, the propelling speed and the depth of the hand-held electric hammer during drilling at each time, so that the drilling quality is not guaranteed, and the repeatability is poor. The manual work is high in labor cost: the manual operation people are careless or the cooperation of multiple people is inconsistent, safety accidents are easy to happen, the operation is unsafe, air flow in the channel is insufficient, light illumination is insufficient, and the operation environment is poor due to dust generated in drilling. Therefore, a method is urgently needed to improve the working efficiency and solve the problems of difficulty in manual hole making, low efficiency, poor precision, long construction period, inconvenience in operation space and the like in the tunnel punching process.

Disclosure of Invention

The technical problem to be solved by the invention is to provide a tunnel drilling machine and a construction method thereof aiming at the defects in the prior art, wherein the tunnel drilling machine is simple in structure, reasonable in design, convenient to implement, capable of being effectively applied to tunnel drilling, wide in drilling operation range and high in working efficiency, and can effectively solve the problems of difficulty in manual hole making, low efficiency, poor precision, long construction period, inconvenience in operation space and the like in the tunnel drilling process by combining a construction method, and the tunnel drilling machine is remarkable in effect and convenient to popularize.

In order to solve the technical problems, the invention adopts the technical scheme that: a tunnel drilling machine comprises a crawler chassis, a control cabinet and a movable base body which are sequentially arranged from the ground to the top, wherein the control cabinet is fixedly connected to the top of the crawler chassis, a sliding mechanism is arranged between the movable base body and the control cabinet and comprises a first guide rail arranged at the top of the control cabinet, a second guide rail which is perpendicular to the first guide rail is connected to the first guide rail in a sliding manner, the lower part of the second guide rail is connected with a first lead screw which is parallel to the first guide rail, the end part of the first lead screw is connected with a first motor which is used for driving the first lead screw to rotate, the lower part of the movable base body is connected with a second lead screw which is parallel to the second guide rail, the end part of the second lead screw is connected with a second motor which is used for driving the second lead screw to rotate, a rotary disc seat is arranged at the upper part of the movable base body, and a third motor which is used for driving the rotary disc seat to rotate is arranged in the movable base body, the end part of the rotary disk seat is rotatably connected with a telescopic arm, the end part of the telescopic arm is connected with a drill bit mechanism, and a visual system for identifying, positioning and drilling is arranged in the drill bit mechanism.

The tunnel drilling machine comprises a support frame and a drill bit, wherein one end of the support frame is movably connected with a telescopic arm, a drill bit guide rail is arranged at the other end of the support frame, a mounting plate used for mounting the drill bit is slidably connected onto the drill bit guide rail, a drill bit lead screw parallel to the drill bit guide rail is connected to the lower portion of the mounting plate, and the end portion of the drill bit lead screw is connected with a servo motor used for driving the drill bit lead screw to rotate.

Foretell tunnel boring machine, be connected with the hinge bar between support frame and the flexible arm and be used for adjusting the pneumatic cylinder at support frame inclination.

Foretell tunnel boring machine, be provided with oil pressure station, generator and controller in the control cabinet, first motor, second motor, third motor and servo motor all are connected with the output of controller, visual system is connected with the input of controller.

The invention also discloses a construction method of the tunnel drilling machine, which adopts the tunnel drilling machine and comprises the following steps:

moving to a tunnel drilling operation working area through the crawler chassis;

driving the rotary disk seat to rotate through the third motor, so that the drill bit mechanism rotates to an approximate working angle;

thirdly, the drill bit mechanism extends to the approximate working height under the action of the telescopic arm;

fourthly, a vision system in the drill bit mechanism identifies and positions the target drilled hole;

driving the movable seat body to move through the sliding mechanism, and further adjusting the position of the drill bit;

sixthly, driving the drill bit to move through the servo motor, and enabling the drill bit to be aligned with the target drilling hole;

seventhly, adjusting the inclination angle of the drill bit through a hydraulic cylinder to enable the drill bit to be vertical to the surface of the target drilling hole;

and step eight, the drill bit works to perform drilling operation.

In the construction method of the tunnel boring machine, the specific process of identifying and positioning the target borehole by the vision system in the drill mechanism in the fourth step includes:

step 401, performing edge detection on a target drilling mark image;

step 402, carrying out color detection on the target drilling hole mark image;

and 403, precisely positioning the target drilling mark image by adopting a depth convolution neural network.

In the above construction method of the tunnel boring machine, the specific process of performing edge detection on the target boring mark image in step 401 includes:

step 40101, filtering the target borehole marking image by using an optimized median filtering algorithm;

the optimized median filtering algorithm is judged by adding noise points into the median filtering algorithm, and the noise points in a window are removed when the median is obtained, and the specific process comprises the following steps:

noise point determination

Wherein, X_ijIs a point to be judged in the window, S is a signal point, N is a noise point, X_minIs the minimum value, X, within the window_maxIs the maximum value in the window, D is the set threshold;

noise point removal

Wherein, X_medTo remove the median value after the noise point in the window, [ X ]_ij]Is the complete set of points in the window, [ X ]_S]A noise point set in a window is defined, p is the number of all points in the window, and q is the number of noise points in the window;

step 40102, performing edge enhancement on the target borehole marking image by adopting a gradient amplitude value;

step 40103, edge position detection is carried out on the target borehole marking image by adopting a Canny operator;

smoothing the target borehole marking image with a Gaussian filter; calculating the magnitude and direction of the gradient by using the finite difference of the first-order partial derivatives; carrying out non-maximum suppression on the gradient amplitude; edges are detected and connected using a dual threshold algorithm.

In the above construction method of the tunnel boring machine, the specific process of performing color detection on the target boring mark image in step 402 includes:

step 40201, converting the target drilling mark image from the RGB space to the HSV space;

40202, setting fuzzy ranges of hue components H, saturation components S and brightness components V in an HSV space;

40203, processing the target drilling hole mark image in the HSV space into a binary image according to the fuzzy range;

40204, drawing and connecting the periphery of each connected region in the binary image;

step 40205, screening the connected region according to the length-width ratio and the area of the drilling mark to obtain a candidate region image of the drilling mark;

step 40206, segmenting the candidate region image from the target borehole marker image, so as to facilitate subsequent accurate identification and positioning.

In the above construction method of the tunnel boring machine, the specific process of accurately positioning the target boring mark image by using the depth convolution neural network in step 403 includes:

40301, constructing a depth convolution neural network structure for accurately positioning a target borehole marking image;

the deep convolutional neural network structure comprises three levels of cascaded convolutional neural networks CNN;

the first layer comprised 1 CNN with 39 x 39 inputs;

the second layer comprises 2 CNNs, the input sizes are all 15 x 15, every two CNNs form a pair, a pair of CNNs predict an anchor point, and the prediction results are averaged;

the third layer comprises 2 CNNs, the input sizes are all 15 x 15, and every two CNNs form a pair;

step 40302, training a deep convolutional neural network;

selecting a plurality of candidate area images as a training sample set, marking drilled holes or non-drilled holes, and inputting a deep convolutional neural network for training;

40303, accurately positioning the target borehole marker image by using a depth convolution neural network;

and inputting the candidate area image to perform target drilling mark identification and positioning to obtain a positioning and identification result of drilling or non-drilling.

Compared with the prior art, the invention has the following advantages:

1. the tunnel drilling machine is simple in structure, reasonable in design and convenient to achieve.

2. The crawler chassis is used as the travelling mechanism of the tunnel drilling machine, so that the tunnel drilling machine can be better adapted to the complex ground environment in the tunnel.

3. According to the invention, the crawler chassis, the control cabinet and the movable seat body are designed into an up-down structure, so that the floor area of the tunnel drilling machine is effectively reduced, meanwhile, the mounting height of the telescopic arm is increased, and the designed telescopic length of the telescopic arm is reduced.

4. According to the invention, the hydraulic cylinder is designed between the support frame and the telescopic arm, so that the inclination angle of the support frame is adjusted, the angle of the drill bit is further adjusted, and the drill bit is ensured to be vertical to the surface to be drilled.

5. The method adopts the combination of a computer vision algorithm and a deep convolution neural network to determine the drilling position, firstly carries out edge detection and then color detection on the target drilling mark image, and finally adopts the deep convolution neural network to realize accurate identification and positioning on the target drilling mark image.

6. The invention can be effectively applied to tunnel drilling, has wide drilling operation range and high working efficiency, can effectively solve the problems of difficult manual hole making, low efficiency, poor precision, long construction period, inconvenient operation space and the like in the tunnel drilling process by combining a construction method, has obvious effect and is convenient to popularize.

In conclusion, the tunnel drilling machine is simple in structure, reasonable in design, convenient to implement, capable of being effectively applied to tunnel drilling, wide in drilling operation range and high in working efficiency, can effectively solve the problems of difficulty in manual hole making, low efficiency, poor precision, long construction period, inconvenience in operation space and the like in the tunnel drilling process by combining a construction method, and is remarkable in effect and convenient to popularize.

The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.

Drawings

FIG. 1 is a schematic structural view of the present invention;

FIG. 2 is a schematic structural view of the drill mechanism of the present invention;

FIG. 3 is a flow chart of the construction method of the present invention;

FIG. 4 is an image effect diagram of the target borehole marking in the construction method of the present invention.

Description of reference numerals:

1-a crawler chassis; 2-control cabinet; 3-a first guide rail;

4-a second guide rail; 5-a first lead screw; 6-a movable seat body;

7-a second lead screw; 8, rotating the disc seat; 9-a telescopic arm;

10-a drill bit mechanism; 10-1-a support; 10-2-bit;

10-3-drill guide; 10-4-mounting plate; 10-5-drill screw;

11 — a first motor; 12-a second motor; 13-a third motor;

14-a servo motor; 15-a vision system; 16-a hinged lever;

and 17, a hydraulic cylinder.

Detailed Description

As shown in fig. 1, the tunnel boring machine of the present invention comprises a crawler chassis 1, a control cabinet 2 and a movable seat 6 which are sequentially arranged from the ground to the top, wherein the control cabinet 2 is fixedly connected to the top of the crawler chassis 1, a sliding mechanism is arranged between the movable seat 6 and the control cabinet 2, the sliding mechanism comprises a first guide rail 3 arranged at the top of the control cabinet 2, a second guide rail 4 which is perpendicular to the first guide rail 3 is slidably connected to the first guide rail 3, a first lead screw 5 which is parallel to the first guide rail 3 is connected to the lower portion of the second guide rail 4, a first motor 11 for driving the first lead screw 5 to rotate is connected to the end of the first lead screw 5, a second lead screw 7 which is parallel to the second guide rail 4 is connected to the lower portion of the movable seat 6, a second motor 12 for driving the second lead screw 7 to rotate is connected to the end of the second lead screw 7, the drilling machine is characterized in that a rotary disk seat 8 is arranged on the upper portion of the movable seat body 6, a third motor 13 which drives the rotary disk seat 8 to rotate is arranged in the movable seat body 6, a telescopic arm 9 is rotatably connected to the end portion of the rotary disk seat 8, a drill bit mechanism 10 is connected to the end portion of the telescopic arm 9, and a vision system 15 which is used for recognizing, positioning and drilling is arranged in the drill bit mechanism 10.

During specific implementation, the crawler chassis 1 is used as a travelling mechanism of the tunnel drilling machine, so that the tunnel drilling machine can be better adapted to the complex ground environment in the tunnel. The crawler chassis 1, the control cabinet 2 and the movable seat body 6 are designed to be of an upper structure and a lower structure, so that the floor area of the tunnel drilling machine is effectively reduced, meanwhile, the mounting height of the telescopic arm 9 is increased, the design telescopic length of the telescopic arm 9 is reduced, and the cost is saved.

In this embodiment, as shown in fig. 2, the drill mechanism 10 includes a support frame 10-1 and a drill 10-2, one end of the support frame 10-1 is movably connected to the telescopic arm 9, the other end of the support frame 10-1 is provided with a drill guide rail 10-3, the drill guide rail 10-3 is slidably connected to an installation plate 10-4 for installing the drill 10-2, a drill screw 10-5 parallel to the drill guide rail 10-3 is connected to the lower portion of the installation plate 10-4, and a servo motor 14 for driving the drill screw 10-5 to rotate is connected to an end of the drill screw 10-5.

In this embodiment, a hinged rod 16 and a hydraulic cylinder 17 for adjusting the inclination angle of the support frame 10-1 are connected between the support frame 10-1 and the telescopic arm 9.

During specific implementation, the hydraulic cylinder 17 is connected with the oil pressure station, the inclination angle of the support frame 10-1 is adjusted through the hydraulic cylinder 17, the angle of the drill bit 10-2 is further adjusted, and the drill bit 10-2 is guaranteed to be perpendicular to the surface of a target drilling hole.

In this embodiment, an oil pressure station, a generator and a controller are arranged in the control cabinet 2, the first motor 11, the second motor 12, the third motor 13 and the servo motor 14 are all connected with an output end of the controller, and the vision system 15 is connected with an input end of the controller.

As shown in fig. 3, the construction method of the tunnel boring machine of the present invention includes the steps of:

moving to a tunnel drilling operation working area through the crawler chassis 1;

step two, the third motor 13 drives the rotary disk seat 8 to rotate, so that the drill bit mechanism 10 rotates to an approximate working angle;

thirdly, the drill bit mechanism 10 is extended to the approximate working height under the action of the telescopic arm 9;

fourthly, the vision system 15 in the drill bit mechanism 10 identifies and positions the target drill hole;

driving the movable seat body 6 to move through the sliding mechanism, and further realizing the position adjustment of the drill 10-2;

sixthly, driving the drill bit 10-2 to move through the servo motor 14, and enabling the drill bit 10-2 to be aligned with a target drill hole;

seventhly, adjusting the inclination angle of the drill bit 10-2 through the hydraulic cylinder 17 to enable the drill bit 10-2 to be perpendicular to the surface of the target drilling hole;

and step eight, the drill 10-2 works to perform drilling operation.

In this embodiment, the specific process of identifying and positioning the target borehole by the vision system 15 in the drill mechanism 10 in step four includes:

step 401, performing edge detection on a target drilling mark image;

restoring the geometric characteristics of the target drilling mark image by using edge detection, thereby preliminarily determining the position of the drilling mark in the two-dimensional image; the target borehole marker image is shown in fig. 4;

step 402, carrying out color detection on the target drilling hole mark image;

based on the identification algorithm based on the edge information, the identification based on the color detection algorithm is further carried out, so that a more accurate drilling hole label circumscribed rectangle is obtained;

and 403, precisely positioning the target drilling mark image by adopting a depth convolution neural network.

Computer vision algorithms cannot accurately locate the target under complex conditions, and therefore further accurate location is achieved by means of deep learning.

In this embodiment, the specific process of performing edge detection on the target drilling mark image in step 401 includes:

step 40101, filtering the target borehole marking image by using an optimized median filtering algorithm;

the optimized median filtering algorithm is judged by adding noise points into the median filtering algorithm, and the noise points in a window are removed when the median is obtained, and the specific process comprises the following steps:

noise point determination

Wherein, X_ijIs a point to be judged in the window, S is a signal point, N is a noise point, X_minIs the minimum value, X, within the window_maxIs the maximum value in the window, D is the set judgment threshold;

in specific implementation, the value range of D is 10-20; the noise points are generally distributed at a minimum value end or a maximum value end, but when the image noise density is high, points near an extreme value may also be noise points, so that a judgment threshold value D is set;

noise point removal

Wherein, X_medTo remove the median value after the noise point in the window, [ X ]_ij]Is the complete set of points in the window, [ X ]_S]A noise point set in a window is defined, p is the number of all points in the window, and q is the number of noise points in the window;

by removing the noise points, all signal points in the window are adopted when the median is calculated, and the noise transmission is effectively avoided.

Step 40102, performing edge enhancement on the target borehole marking image by adopting a gradient amplitude value;

step 40103, edge position detection is carried out on the target borehole marking image by adopting a Canny operator;

smoothing the target borehole marking image with a Gaussian filter; calculating the magnitude and direction of the gradient by using the finite difference of the first-order partial derivatives; carrying out non-maximum suppression on the gradient amplitude; edges are detected and connected using a dual threshold algorithm.

In this embodiment, the specific process of performing color detection on the target borehole marker image in step 402 includes:

step 40201, converting the target drilling mark image from the RGB space to the HSV space;

40202, setting fuzzy ranges of hue components H, saturation components S and brightness components V in an HSV space;

in specific implementation, fuzzy ranges of hue components H, saturation components S and brightness components V in the HSV space are shown in Table 1;

TABLE 1 fuzzy Range

40203, processing the target drilling hole mark image in the HSV space into a binary image according to the fuzzy range;

40204, drawing and connecting the periphery of each connected region in the binary image;

step 40205, screening the connected region according to the length-width ratio and the area of the drilling mark to obtain a candidate region image of the drilling mark;

step 40206, segmenting the candidate region image from the target borehole marker image, so as to facilitate subsequent accurate identification and positioning.

In specific implementation, since the resolution of the picture returned by the camera in the vision system 15 is high, which may seriously affect the real-time performance of the settlement of the positioning result, the candidate region image is segmented in the target borehole marker image and used as the input of the deep convolutional neural network.

In this embodiment, the specific process of accurately positioning the target borehole marker image by using the deep convolutional neural network in step 403 includes:

40301, constructing a depth convolution neural network structure for accurately positioning a target borehole marking image;

the deep convolutional neural network structure comprises three levels of cascaded convolutional neural networks CNN;

the first layer comprised 1 CNN with 39 x 39 inputs;

the second layer comprises 2 CNNs, the input sizes are all 15 x 15, every two CNNs form a pair, a pair of CNNs predict an anchor point, and the prediction results are averaged;

the third layer comprises 2 CNNs, the input sizes are all 15 x 15, and every two CNNs form a pair;

step 40302, training a deep convolutional neural network;

selecting a plurality of candidate area images as a training sample set, marking drilled holes or non-drilled holes, and inputting a deep convolutional neural network for training;

40303, accurately positioning the target borehole marker image by using a depth convolution neural network;

and inputting the candidate area image to perform target drilling mark identification and positioning to obtain a positioning and identification result of drilling or non-drilling.

The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention, and all simple modifications, changes and equivalent structural changes made to the above embodiment according to the technical spirit of the present invention still fall within the protection scope of the technical solution of the present invention.