A coal flow monitoring method based on binocular vision

Abstract

The present invention discloses a coal flow monitoring method based on binocular vision, including step 1, selecting two binocular cameras for calibration to make a detector, step 2, reasonably arranging a detector bracket equipped with the detector, step 3, obtaining the coal flow information above the scraper conveyor based on the detector, and calculating the coal flow Q of the comprehensive caving working face, wherein the binocular camera is selected as an MV-SUA630C-M industrial-grade camera, and the lens of the MV-SUA630C-M industrial-grade camera is selected as an M2808-1K-4 lens. The present invention calibrates the binocular camera to make a detector, and reasonably arranges the bracket equipped with the detector. When the comprehensive caving working face is caving coal, the coal flow information above the scraper conveyor can be obtained in real time. At the same time, by reconstructing the coal flow information in three dimensions, the three-dimensional point cloud data of the coal flow is obtained, and then after obtaining the coal flow scanning volume, the coal flow size per unit time can be calculated, so that the staff can quickly know the specific coal flow.

Description

Technical Field

[0001] The invention specifically relates to a coal flow monitoring method based on binocular vision. Background Art

[0002] The fully mechanized top coal caving working face refers to the comprehensive mechanized top coal caving and mining working face, which is usually mined in sufficiently thick coal seams. This type of working face is characterized by the use of hydraulic supports, high-power scraper conveyors, double-drum coal mining machines and other equipment for mechanized operation, and plays an important role in coal mining;

[0003] At present, top coal caving operations in fully mechanized caving working faces mainly include two methods:

[0004] The first is manual top coal caving, which relies on the binocular vision and work experience of the coal caving workers. It is a relatively traditional way of caving coal. Affected by environmental factors such as coal dust, light, and the narrow space behind the support, the control accuracy of manual top coal caving is low, which cannot meet the requirements of the top coal recovery rate of the fully mechanized top coal caving technology, and even has a serious impact on the life, health and safety of the operators.

[0005] The second method is to use the control system to operate the opening and closing of the coal-laying port. The opening and closing time of the coal-laying port is obtained through statistical laws and workers' experience. A pre-set coal-laying program is written into the control system, and coal is automatically laid according to the pre-set coal-laying time. This method reduces manual participation and ensures the safety of workers. However, on the other hand, due to the influence of the thickness and inclination of the underground coal seam, the opening and closing time of the coal-laying port of the hydraulic support and the coal-laying action need to change with the advancement of the mining face, and the coal-laying port information of a single hydraulic support cannot be obtained;

[0006] In summary, the current fully mechanized caving working face uses two traditional methods, manual visual inspection or control system, to caving coal. It is impossible to accurately monitor the coal flow rate, and under-cavitation or over-cavitation often occurs, resulting in insufficient coal flow or overload on the scraper conveyor, which is not conducive to production.

[0007] Therefore, it is necessary to invent a coal flow monitoring method based on binocular vision to solve the above problems. Summary of the invention

[0008] The purpose of the present invention is to provide a coal flow monitoring method based on binocular vision. By calibrating the binocular camera to make a detector, and reasonably arranging the bracket equipped with the detector, when the coal is released on the comprehensive caving working face, the coal flow information above the scraper conveyor can be obtained in real time. At the same time, the coal flow information is reconstructed in three dimensions to obtain three-dimensional point cloud data of the coal flow. After obtaining the coal flow scanning volume, the coal flow size per unit time can be calculated, so that the staff can quickly know the real-time coal flow of the comprehensive caving working face, thereby accurately monitoring the coal flow of the comprehensive caving working face, avoiding under-release and over-release. The invention stabilizes the coal flow on the scraper conveyor, which is beneficial to production, so as to solve the above-mentioned deficiencies in the technology.

[0009] In order to achieve the above object, the present invention provides the following technical solution: a coal flow monitoring method based on binocular vision, comprising the following steps:

[0010] Step 1: Select two binocular cameras for calibration to make a detector;

[0011] Step 2: Reasonably arrange the detector bracket equipped with the detector;

[0012] Step 3: Based on the detector, the coal flow information above the scraper conveyor is obtained and the coal flow rate Q of the fully mechanized caving working face is calculated.

[0013] In the aforementioned coal flow monitoring method based on binocular vision, in step 1, the binocular camera is selected as the MV-SUA630C-M industrial-grade camera, and the lens of the MV-SUA630C-M industrial-grade camera is selected as the M2808-1K-4 lens, and an active light source is added, and the active light source is selected as SDI 3005.

[0014] In the above-mentioned coal flow monitoring method based on binocular vision, in step 1, the binocular camera is calibrated, and the specific steps of making a detector are as follows:

[0015] 1.1. Prepare a calibration plate with a black and white chessboard paper inside;

[0016] 1.2. Place the calibration plate at the scanning position and adjust the two binocular cameras so that the calibration plate appears within the center of the field of view of both cameras at the same time. At this time, fix the binocular cameras.

[0017] Among them, the fixed positions of the binocular cameras are one on the left and one on the right on the same straight line;

[0018] 1.3. Adjust the position of the calibration plate to change its image position in the binocular camera, thereby obtaining multiple sets of chessboard images at different positions and angles, and storing them in the disk;

[0019] 1.4. The binocular camera on the left is calibrated separately. The chessboard image acquired by the binocular camera on the left is imported into the MATLAB program. The specifications of the calibration plate are input. All vertices on the calibration plate are extracted using the MATLAB program. Then the parameters of the binocular camera on the left are calibrated. The three-dimensional position simulation diagram and parameters of the calibration plate are obtained, and then the calibration results of the binocular camera on the left are obtained. The specific results are as follows;

[0020] The internal parameter matrix is:

[0021]

[0022] The radial distortion is:

[0023] [-0.060.47]

[0024] The correction matrix is:

[0025]

[0026] 1.5. Import the chessboard image acquired by the right binocular camera into the MATLAB program, input the calibration plate specifications, use the MATLAB program to extract all vertices on the calibration plate, and then calibrate the parameters of the right binocular camera. The three-dimensional position simulation diagram and parameters of the calibration plate are obtained, and then the calibration results of the right binocular camera are obtained. The specific results are as follows;

[0027] The internal parameter matrix is:

[0028]

[0029] The radial distortion is:

[0030] [-0.1040.917]

[0031] The correction matrix is:

[0032]

[0033] 1.6. Input the chessboard images acquired by the two binocular cameras into the MATLAB program at the same time. After completing the vertex acquisition, the three-dimensional position simulation diagram of the calibration plate is obtained. At this time, the rotation matrix and translation matrix of the two binocular cameras are calculated to complete the calibration of the two binocular cameras. The specific results are as follows; the rotation matrix R of the two binocular cameras is:

[0034]

[0035] The translation matrix of the two binocular cameras is: T = [210.802312.995810.1175];

[0036] 1.7. After the two binocular cameras have been calibrated, they can be directly installed side by side to make a detector.

[0037] In the above-mentioned binocular vision-based coal flow monitoring method, in step 2, the specific steps of reasonably arranging the detector bracket equipped with the detector are:

[0038] 2.1. Calculate the maximum distance between the detector bracket installation position and the coal bracket. The specific calculation formula is as follows:

[0039]

[0040] Where, v is the running speed of the scraper conveyor, and v = 1.5m / s,

[0041] Δt s It is the time interval between the top coal leaving the coal discharge port and the monitoring system scanning the coal flow.

[0042] N a is the serial number on the target detector bracket, where the detector bracket at the fully mechanized caving working face is numbered 1, and increases outwards in sequence.

[0043] N f It is the serial number on the coal caving support. The serial number of the detector support at the fully mechanized caving working face is 1, and it increases outwards in sequence.

[0044] W y is the support width of the hydraulic support;

[0045] 2.2. Analyze the dust concentration of coal flow and get |N f -N a |The relationship is:

[0046] 4≤|N f -N a |;

[0047] 2.3. Analyze the distance of coal flow movement and get |N f -N a Another relational expression for | is:

[0048] 1≤|N f -N a |;

[0049] 2.4. Based on the analysis of the dust concentration of the coal flow and the distance of the coal flow movement, the specific location for installing the detector bracket is determined.

[0050] In the above-mentioned binocular vision-based coal flow monitoring method, in step 2.2, the specific steps for analyzing the dust concentration are:

[0051] 2.2.1. Determine the hydraulic support that needs to be caving coal at the caving working face. Assume that the hydraulic support is support No. 80, define the serial number of the hydraulic support as 80, and set dust concentration detection points from the 15th support on the left to the 15th support on the right with the hydraulic support as the center;

[0052] 2.2.2. After coal discharge begins, record the changes in wind direction and dust concentration at the working face during the coal discharge process.

[0053] In the above-mentioned binocular vision-based coal flow monitoring method, in step 2.3, the specific process of analyzing the distance of coal flow movement is as follows:

[0054] The process of top coal discharge is divided into the first stage in which the top coal is broken and moves downward from the rear of the tail beam under the influence of gravity, the second stage in which the opening and closing of the coal discharge port is controlled by controlling the contraction of the plug plate, and the third stage in which the top coal is broken and contacts the upper surface of the rear scraper conveyor.

[0055] And in the third stage: let v x is the lateral velocity of the top coal when it leaves the coal discharge port,

[0056] v y is the lateral velocity of the top coal when it leaves the coal discharge port,

[0057] v z is the movement speed of the top coal,

[0058] v is the running speed of the scraper conveyor,

[0059] μ is the surface friction coefficient of the scraper conveyor, and its specific value is 0.07.

[0060] The relationship between the distance d of coal flow movement and the distance d is as follows:

[0061]

[0062] Among them, v0 is the initial lateral velocity when the top coal contacts the scraper conveyor, which is 0m / s.

[0063] g is the acceleration due to gravity, which is 9.8 m / s 2 ,

[0064] Substituting into the above formula we obtain: d = 1.64m.

[0065] In the above-mentioned binocular vision-based coal flow monitoring method, in step 3:

[0066] Assume the scanning frequency of the binocular camera is f, and f = 18FPS,

[0067] The time consumed by a single scan is Δt, and Δt=0.5s,

[0068] The width of a single identification area is w, and the length of a single identification area on a scraper conveyor is z s ,

[0069] When the detector is installed at a height of 90 cm, the width of the detector's single identification area is w = 150 cm, and the length of the single identification area on the scraper conveyor is z s =90cm;

[0070] And based on the detector to obtain the coal flow information above the scraper conveyor, the coal flow Q of the fully mechanized caving working face is calculated. The specific steps are as follows:

[0071] 3.1. The coal flow information above the scraper conveyor obtained by the detector is reconstructed in three dimensions to obtain three-dimensional point cloud data of the coal flow;

[0072] 3.2. Starting from t=0, with Δt as the time interval, the 3D point cloud data is segmented according to the scanning time, and a total of t / Δt segment data is obtained;

[0073] 3.3. Take any period of time Δt i Analyze the three-dimensional point cloud data in the section, and let the coal flow scanning volume corresponding to this section be V;

[0074] 3.4. The coal flow scanning volume V is calculated according to its area length z s Divided into multiple small units, the calculation formula of the coal flow scanning volume V is as follows:

[0075]

[0076] Among them, V i The volume of each small unit;

[0077] Δz is the length of each small unit;

[0078] 3.5. Let the cross-sectional area of ​​the coal flow obtained by the detector scanning the coal flow be S i , using the differential method to calculate the cross-sectional area of ​​the coal flow is S i , the specific process is:

[0079] 3.5.1. Assume that the bottom is the scraper conveyor, the top is the cross-sectional profile of the coal flow, and establish a rectangular coordinate system with the optical center of the detector as the origin;

[0080] 3.5.2. Use point cloud data to divide the cross-sectional area of ​​the coal flow into n regions, and obtain the cross-sectional area S of the coal flow based on the starting and ending coordinates of the interval. i , the specific formula is as follows:

[0081]

[0082] Among them, x is the coal flow profile coordinate,

[0083] y is the coordinate of the bottom contour of the scraper conveyor;

[0084] 3.6. According to the calculated coal flow cross-sectional area S i , the volume of all small units V i Calculate the sum and get the coal flow scanning volume V. The specific formula is as follows:

[0085]

[0086] 3.7. Combining formulas (1), (2) and (3), we can get the formula for the coal flow scanning volume V, which is as follows:

[0087]

[0088] 3.8. Calculate the additional coal flow volume V obtained by the detector in a single time x , and the calculation formula is as follows:

[0089] V x =VV c

[0090] Among them, V c The repeated partial volumes of the coal flow volume database acquired for the detector;

[0091] 3.9. The additional coal flow volume V obtained by a single pass of all detectors x The total volume of coal released is obtained by summing up, and the coal flow size per unit time, that is, the coal flow rate, is calculated as follows, and the calculation formula is as follows:

[0092]

[0093] Compared with the prior art, the present invention has the following beneficial effects:

[0094] The present invention calibrates a binocular camera to make a detector, and reasonably arranges a bracket equipped with the detector. When coal is released on the fully-mechanized caving working face, the coal flow information above the scraper conveyor can be obtained in real time. At the same time, the coal flow information is reconstructed in three dimensions to obtain three-dimensional point cloud data of the coal flow. After obtaining the coal flow scanning volume, the coal flow rate per unit time can be calculated, so that the staff can quickly know the real-time coal flow rate of the fully-mechanized caving working face, thereby accurately monitoring the coal flow rate of the fully-mechanized caving working face and avoiding under-release or over-release. The invention ensures that the coal flow on the scraper conveyor is stable, which is beneficial to production. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0096] Figure 1 is a flow chart of the present invention;

[0097] Figure 2 This is a simulation diagram of the binocular camera calibration of the present invention;

[0098] Figure 3 It is a dust concentration measurement change line graph of the present invention;

[0099] Figure 4 It is a schematic diagram of the free-falling motion of the coal flow of the present invention;

[0100] Figure 5 It is a schematic diagram of friction rolling of the coal flow of the present invention;

[0101] Figure 6 It is a schematic diagram of coal discharge monitoring of the present invention;

[0102] Figure 7 A schematic diagram of the single scan volume division of the present invention;

[0103] Figure 8 It is a schematic diagram of calculating the cross-sectional area of ​​coal flow according to the present invention;

[0104] Fig. 9 It is a schematic diagram of the experimental tool of the present invention;

[0105] Fig.10 A schematic diagram of the difference between the horizontal scanning profile and the actual profile of the present invention;

[0106] Fig.11 This is a scanning error analysis diagram of coal flow with different block sizes according to the present invention;

[0107] Fig.12 A schematic diagram of different brightness of the present invention;

[0108] Fig.13 This is a diagram demonstrating the experimental results of the external lighting impact of the present invention. DETAILED DESCRIPTION

[0109] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0110] The present invention provides Figure 1-13 A method for monitoring coal flow based on binocular vision is shown, comprising the following steps:

[0111] Step 1: Select two binocular cameras for calibration to make a detector;

[0112] The binocular camera is selected as the MV-SUA630C-M industrial camera. The MV-SUA630C-M industrial camera has a programmable I / O function, supports external trigger flash synchronization photography, adopts a USB3.0 transmission interface, has a faster transmission speed, and reaches a transmission bandwidth of 5Gbps. In combination with a high-speed image sensor, the minimum transmission time can be 1.2 milliseconds, meeting the high-efficiency requirements for data transmission speed. It can be used in environments with low visibility and lack of natural light underground, and has a high resolution, can accurately capture the details of the coal flow, and has a high frame rate, which can ensure the capture of images of instantaneous changes in the coal flow.

[0113] The lens of the MV-SUA630C-M industrial camera is the M2808-1K-4 lens, which has the advantages of small size, light weight, adjustable focal length and aperture, can flexibly adapt to the underground environment, and has good field of view and depth perception performance;

[0114] In step 1, since natural light cannot be relied upon in the underground environment, an active light source needs to be selected to ensure that the binocular camera obtains clear depth information of the coal flow. In view of this, the active light source is selected as SDI 3005, which uses an integrated LED as the power supply, provides an illumination brightness of 100LM, an illumination uniformity of up to 90%, an operating power of 35W, and an external power supply standard of 220V. It has stable performance and can adapt to the underground environment;

[0115] like Figure 2 As shown, in step 1, the binocular camera is calibrated, and the specific steps of making the detector are:

[0116] 1.1. Prepare a calibration plate with a black and white chessboard paper inside;

[0117] 1.2. Place the calibration plate at the scanning position and adjust the two binocular cameras so that the calibration plate appears within the center of the field of view of both cameras at the same time. At this time, fix the binocular cameras.

[0118] In step 1.2, the binocular cameras are fixed in the left and right positions on the same straight line;

[0119] 1.3. Adjust the position of the calibration plate to change its image position in the binocular camera, thereby obtaining multiple sets of chessboard images at different positions and angles, and storing them in the disk;

[0120] 1.4. The binocular camera on the left is calibrated separately. The chessboard image acquired by the binocular camera on the left is imported into the MATLAB program. The specifications of the calibration plate are input. All vertices on the calibration plate are extracted using the MATLAB program. Then the parameters of the binocular camera on the left are calibrated. The three-dimensional position simulation diagram and parameters of the calibration plate are obtained, and then the calibration result of the binocular camera on the left is obtained.

[0121] In step 1.4, the calibration results of the left binocular camera are as follows:

[0122] The internal parameter matrix is:

[0123]

[0124] The radial distortion is:

[0125] [-0.060.47]

[0126] The correction matrix is:

[0127]

[0128] 1.5. Import the chessboard image acquired by the right binocular camera into the MATLAB program, input the calibration plate specifications, use the MATLAB program to extract all vertices on the calibration plate, and then calibrate the parameters of the right binocular camera. The obtained calibration plate 3D position simulation diagram and parameters, and then obtain the calibration result of the right binocular camera;

[0129] In step 1.5, the calibration results of the right binocular camera are as follows:

[0130] The internal parameter matrix is:

[0131]

[0132] The radial distortion is:

[0133] [-0.1040.917]

[0134] The correction matrix is:

[0135]

[0136] Step 1.6: Input the chessboard images acquired by the two binocular cameras into the MATLAB program at the same time. After completing the vertex acquisition, the three-dimensional position simulation diagram of the calibration plate is obtained. At this time, the rotation matrix and translation matrix of the two binocular cameras are calculated to complete the calibration of the two binocular cameras.

[0137] In step 1.6, the rotation matrix R of the two stereo cameras is:

[0138]

[0139] The translation matrix of the two binocular cameras is: T = [210.802312.995810.1175];

[0140] 1.7. After the two binocular cameras have been calibrated, they can be directly installed side by side to form a detector;

[0141] In step 1, since the binocular camera obtains the spatial information of the object by shooting from different angles with two cameras, this is to obtain the point cloud information of the three-dimensional space through the image information of the two-dimensional plane, that is, the binocular vision three-dimensional reconstruction converts the phase plane coordinates of the object into the world coordinates. The premise for completing this process is that the parameters of the visual model such as the camera spacing, imaging plane angle, lens focal length, optical center, etc. must be known in advance. The calibration of the binocular camera refers to the process of determining the internal and external parameters of the binocular camera in order to accurately measure the three-dimensional position of the object in the image or perform other computer vision tasks. Therefore, the binocular camera needs to be calibrated before it can work normally.

[0142] Step 2: Reasonably arrange the detector bracket equipped with the detector;

[0143] In step 2, when the coal-laying support starts to lay coal, the top broken coal will roll down to the coal-laying port along the hydraulic support shielding beam and tail beam to form a coal body to be laid out. Then the hydraulic support coal-laying plug plate is retracted, and the top coal rolls down from the coal-laying port and flows to the rear scraper conveyor. Therefore, in order to ensure the stability of the coal flow in the laid coal area, installing the detector support in the laid coal support area is the best solution in the prior art, which can make the detector support work stably and reliably, and is more conducive to the operation of the coal-laying amount monitoring system.

[0144] The installation height of the detector on the detector bracket is set to 90-95cm, which can completely scan the coal flow on the scraper conveyor. The installation angle θ is set to 0-5°, which can obtain coal flow in different forms, and the monitoring effect is optimal. When the height of the detector from the plane to be detected is 30cm, the size of the detector recognition area is a rectangle with a length of l and a width of w of 60×45cm. As the height of the detector from the plane to be detected increases, the length l and width w of the recognition area will also increase, and the size of the area and the height meet the following formula:

[0145]

[0146] Where h is the distance between the optical center of the detector and the surface of the scraper conveyor, and q is a non-zero constant;

[0147] And in step 2, the specific steps of reasonably arranging the detector bracket equipped with the detector are:

[0148] 2.1. Calculate the maximum distance between the detector bracket installation position and the coal bracket. The specific calculation formula is as follows:

[0149]

[0150] Wherein, v is the running speed of the scraper conveyor, and v = 1.5m / s;

[0151] αt s It is the time interval between the top coal leaving the coal discharge port and the monitoring system scanning the coal flow.

[0152] N a is the serial number on the target detector bracket, where the detector bracket at the fully mechanized caving working face is numbered 1, and increases outwards in sequence.

[0153] N f It is the serial number on the coal caving support, among which the detector support at the fully mechanized caving working face is numbered 1 and increases outwards in sequence;

[0154] W y is the support width of the hydraulic support;

[0155] 2.2. Analyze the dust concentration and get |N f -N a |The relationship is:

[0156] 4≤|N f -N a |

[0157] In step 2.2, since the binocular camera is set to MV-SUA630C-M industrial-grade camera, the minimum dust concentration for shooting within two meters is 10 mg / m3. The dust concentration in the coal-laying area is higher between 1-3 racks away from the coal-laying port, and the measured value is between 10 and 65 mg / m3. 3 After the 4th frame, the dust concentration is 10mg / m 3 Below, and continues to decrease, so considering the actual dust situation in the working surface, the detector installation bracket sequence number is best after the 4th frame, that is, 4≤|N f -N a |;

[0158] like Figure 3 As shown in the figure, the specific steps for analyzing dust concentration are:

[0159] 2.2.1. Determine the hydraulic support that needs to be caving coal at the caving working face. Assume that the hydraulic support is support No. 80, define the serial number of the hydraulic support as 80, and set dust concentration detection points from the 15th support on the left to the 15th support on the right with the hydraulic support as the center;

[0160] 2.2.2. After the coal caving begins, record the wind direction and dust concentration changes at the working face during the coal caving process;

[0161] In step 2.2.2, the coal discharge port was actually detected as the point with the highest dust concentration, reaching 63.5 mg / m 3 In the area without coal placement, the dust concentration value is lower than 5mg / m 3 The area is affected by the wind direction, the dust diffuses slowly, and the overall impact of dust is small. The dust concentration after stabilization is low, while the coal-laying area is seriously affected by dust. After the dust reduction treatment of the working face, the dust concentration gradually stabilized at the position of the 76th hydraulic support. The dust concentration was measured to be stable at 7mg / m 3 ;

[0162] 2.3. Analyze the distance of coal flow movement and get |N f -N a Another relational expression for | is:

[0163] 1≤|N f -N a |

[0164] In step 2.3, during the actual coal caving process, the measured top coal changes from a relative motion state to a relative static state. The movement distance during this stage is 1.64 m. When the detector scans the coal flow in relative motion, the result obtained is multiple constantly changing coal flows. The calculated volume error is too large, which is not conducive to the operation of the coal caving amount monitoring system. Therefore, it is necessary to install a detector bracket in the hydraulic support area after the relative motion ends. That is, the relationship between the detector bracket and the coal caving bracket is: 1≤|N f -N a |.

[0165] The specific process of analyzing the distance of coal flow movement is as follows:

[0166] The process of top coal release is divided into the first stage, the second stage and the third stage in sequence;

[0167] The first stage is: the broken top coal starts to move downward from the rear of the tail beam under the influence of gravity;

[0168] In the first stage, since the top coal is located above the tail beam of the hydraulic support, the support pressure caused by mining, the mine pressure, and the repeated disturbance of the coal-laying support will cause the complete coal seam to be broken into a broken coal body. After the coal-laying port of the tail beam of the hydraulic support is opened, the process is no longer subject to the pressure from the roof. Due to the guidance of the tail beam of the hydraulic support, the interaction between coal blocks, and the obstruction of the gangue in the goaf, the broken top coal will slide to the coal-laying port position, entering the second stage;

[0169] The second stage is: controlling the opening and closing of the coal discharge port by controlling the contraction of the plug plate;

[0170] In the second stage, the coal placing speed can be adjusted. At this time, the crushed top coal is no longer restricted by the space behind the hydraulic support tail beam, and the mutual squeezing effect between the crushed top coal is reduced. It falls to the scraper conveyor mainly under the action of gravity.

[0171] The third stage is: the broken top coal contacts the upper surface of the rear scraper conveyor;

[0172] In the third stage, the movement state of the coal flow can be divided into two situations: relative movement and relative stillness. The relative movement state will increase the difficulty of point cloud scanning, resulting in large calculation errors. In order to reduce the error and ensure the accuracy of the data, it is necessary to accurately locate the relative stillness of the coal flow. Specifically, Figure 4-5 As shown:

[0173] After the coal flow reaches the rear scraper conveyor from the coal discharge port, it experiences two stages: free fall and friction rolling. In the free fall stage, the coal body to be discharged rolls downward under the action of gravity. The top coal at the coal discharge port leaves the coal discharge port under the impact of gravity G and the rear rock, and obtains the initial lateral velocity v x and vertical velocity v y , and the lateral velocity v in this stage x remains constant, the vertical velocity v y Under the action of gravity, it gradually increases until the coal body contacts the rear scraper conveyor. Since the vertical distance between the coal discharge port and the scraper conveyor is small, the vertical velocity v y The increase was small;

[0174] Then it enters the friction rolling stage. After the top coal contacts the scraper conveyor, its downward kinetic energy is converted into the ability to destroy the coal body. At this time, the vertical speed v y Reduced to 0, at the lateral speed v x Under the action of the hydraulic support, the coal flows toward the direction of the hydraulic support and along the direction of the scraper conveyor. At this time, the movement speed of the top coal is v z is 0, the running speed of the scraper conveyor is v, and the two move relative to each other. Under the action of the scraper and friction of the rear scraper conveyor, v z Gradually increase until v z =v, the coal body and the rear scraper conveyor reach relative stillness, and the friction rolling stage ends;

[0175] At this time, let the surface friction coefficient of the scraper conveyor be μ, and μ = 0.07, the movement distance of the coal flow from relative motion to relative stillness, that is, the distance d of the coal flow movement, is as follows:

[0176]

[0177] Among them, v0 is the initial lateral velocity when the top coal contacts the scraper conveyor, which is 0m / s, and g is the gravitational acceleration, which is 9.8m / s 2 , substituting into the above formula, we get d = 1.64m;

[0178] Then, the actual situation on site was observed, and it was found that after the top coal left the coal discharge port, it was in a stationary state at a position of a hydraulic support away from the coal discharge port. The width of the hydraulic support was 2m. According to the observation results, the calculation results were consistent with the actual situation on site, that is, the d value was correct.

[0179] 2.4. Based on the analysis of the dust concentration of the coal flow and the distance of the coal flow movement, the specific location of the detector bracket installation is obtained;

[0180] In step 2.4, the detector bracket is installed at the position of the 4th frame from the coal discharge port in the coal discharge area. At the same time, in order to ensure the timeliness of the monitoring system to obtain data, the detector bracket should be installed as close to the coal discharge port as possible. Therefore, the detector is directly installed at the position of the 4th frame from the coal discharge port, which is the most reasonable and optimal position.

[0181] In step 2, since the working space of underground coal mines is usually small and long, the installation of the detector bracket must not interfere with the normal operation of other equipment. Therefore, the detector bracket equipped with binocular cameras must be reasonably distributed to optimize the monitoring effect.

[0182] Step 3: Obtain the coal flow information above the scraper conveyor based on the detector and calculate the coal flow Q of the fully mechanized caving working face;

[0183] like Figure 6-8 As shown in Figure 1, when the detector starts to monitor the coal flow, the scraper conveyor moves to the right at a speed v.

[0184] At this time, let the scanning frequency of the binocular camera be f, the time consumed for a single scan be Δt, the width of a single recognition area be w, and the length of a single recognition area on the scraper conveyor be z. s , and since the detector bracket installation position has been confirmed, w and z s The values ​​are known, from v, Δt, z s The relationship between the three can be obtained in the following three situations:

[0185] The first case: z s >v·Δt, when the single recognition length is long or the scanning frequency is fast enough, the detector recognizes the same coal flow on the scraper conveyor multiple times, that is, Δt i-1 With Δt iWhen the moment detector obtains that there is an intersection in the coal flow area, the same coal flow characteristics will be obtained repeatedly. It is necessary to extract and divide the same part and the newly added part of the data obtained by the detector each time. The same part is overwritten according to the data obtained at the previous moment, calculate the volume of the newly added part of the coal flow, and repeat this process;

[0186] The second case: z s = v·Δt. When this condition is met, the length of the coal flow data obtained by the detector each time just matches the movement length of the scraper conveyor. By splicing the obtained coal flow data, the complete coal flow data of the scraper conveyor can be obtained. This is the most ideal scanning state under this condition, and only the volume of the coal flow obtained by the detector once needs to be calculated;

[0187] The third case: z s < v·Δt. This condition means that the detector has a slow scanning speed and a small scanning area, and the obtained coal flow data is incomplete. In response to this situation, if the missing data interval is small, a prediction algorithm can be used to infer the coal flow in the missing interval based on the known coal flow. If the missing data interval is large, it is necessary to improve the recognition speed of the detector or replace the equipment with better performance;

[0188] Since the binocular camera is selected as the MV-SUA630C-M industrial camera, that is, the scanning frequency f of the detector = 18FPS, and the single-shot imaging data is synthesized by 9 frames of data, and the time consumed for each scan is Δt = 0.5s;

[0189] And when the installation height of the detector is 90cm, the width w of the single-shot recognition area of the detector = 150cm, then the length z of the single-shot recognition area on the scraper conveyor s = 90cm. From this, it can be obtained that when v ≤ 180cm / s, the first case and the second case are satisfied, and the speed v of the scraper conveyor in the actual field meets the above conditions;

[0190] In view of this, based on the binocular camera to obtain the coal flow information above the scraper conveyor, the specific steps to calculate the coal flow of the fully mechanized caving face are as follows:

[0191] 3.1. Reconstruct the three-dimensional coal flow information obtained by the detector above the scraper conveyor to obtain the three-dimensional point cloud data of the coal flow;

[0192] 3.2. Starting from t = 0, at intervals of Δt, segment the three-dimensional point cloud data according to the scanning time, and a total of t / Δt segments of data are obtained;

[0193] 3.3. Take the three-dimensional point cloud data within any period of time Δt i for analysis, and let the scanned volume of the coal flow corresponding to this segment be V;

[0194] 3.4. The coal flow scanning volume V is calculated according to its area length z s Divided into multiple small units, the calculation formula of the coal flow scanning volume V is as follows:

[0195]

[0196] Among them, V i The volume of each small unit;

[0197] Δz is the length of each small unit;

[0198] 3.5. Let the cross-sectional area of ​​the coal flow obtained by the detector scanning the coal flow be S i , using the differential method to calculate the cross-sectional area of ​​the coal flow is S i , the specific calculation process is:

[0199] 3.5.1. Assume that the bottom is the scraper conveyor, the top is the cross-sectional profile of the coal flow, and establish a rectangular coordinate system with the optical center of the detector as the origin;

[0200] 3.5.2. Use point cloud data to divide the cross-sectional area of ​​the coal flow into n regions, and obtain the cross-sectional area S of the coal flow based on the starting and ending coordinates of the interval. i , the specific formula is as follows:

[0201]

[0202] Among them, x is the coordinate of the coal flow profile,

[0203] y is the coordinate of the bottom contour of the scraper conveyor;

[0204] In step 3.5, since at time Δt i In the coal flow scanning volume unit V i It is an irregular shape, and its volume is directly related to the length of the detector's single identification area division Δz and the coal flow cross section. Δz is known, so it is necessary to calculate the coal flow cross section area S i Solve, and the coal flow cross-sectional area S i is the irregular surface area. For irregular surfaces, the above differential method is used for calculation;

[0205] In step 3.5.2, a single area is divided by the front and rear point cloud data, and together with the scraper conveyor, it forms a trapezoidal area. Therefore, the area of ​​a single differential area can be calculated by combining the trapezoidal area formula, and then the coal flow cross-sectional area S can be obtained by adding them together. i ;

[0206] 3.6. According to the calculated coal flow cross-sectional area S i , the volume of all small units V iCalculate the sum and get the coal flow scanning volume V. The specific formula is as follows:

[0207]

[0208] That is:

[0209]

[0210] 3.7. Combining formulas (1), (2) and (3), we can get the formula for the coal flow scanning volume V, which is as follows:

[0211]

[0212] 3.8. Calculate the additional coal flow volume V obtained by the detector in a single time x , and the calculation formula is as follows:

[0213] V x =VV c

[0214] Among them, V c The repeated partial volumes of the coal flow volume database acquired for the detector;

[0215] In step 3.8, due to the fast scanning speed of the detector, the obtained coal flow volume data has repeated parts, so feature recognition is needed to identify the repeated parts;

[0216] 3.9. The additional coal flow volume V obtained by a single pass of all detectors x The total volume of coal released is obtained by summing up, and the coal flow size per unit time, that is, the coal flow rate, is calculated as follows, and the calculation formula is as follows:

[0217]

[0218] In this step, the length of the single identification area on the scraper conveyor is z. s , the detector single recognition area division length Δz, and the volume V of the small unit i Calculate the coal flow mass M, the specific formula is as follows:

[0219]

[0220] Among them, ρ is the coal flow density, which can be used by the staff to quickly calculate the coal flow quality and have a clear and intuitive understanding of the coal release amount;

[0221] In step 3, the calculation and detection of the coal flow rate in the above steps can ensure the accuracy and reliability of the coal flow rate monitoring results.

[0222] At the same time, in view of steps 1-3, steps 4-5 can also be added to improve the coal placement accuracy and realize the visual operation of coal placement, specifically:

[0223] Step 4: Perform error analysis on the obtained coal flow Q. The specific steps are as follows:

[0224] In this step, the coal flow scanning volume V is affected by the running speed v of the scraper conveyor, the time Δt consumed by a single scan, and the coal flow cross-sectional area S i The acquisition of these data depends on the hardware foundation of the detector and the scraper conveyor, as well as the outer contour of the coal flow itself. Specifically:

[0225] The detector measurement accuracy error is about ±3%.

[0226] The coal flow on the scraper conveyor has a certain porosity, and there is an error between the measured volume and the actual volume;

[0227] The coal flow is composed of coals of different sizes, with complex surface contours. The accuracy of point cloud data description is limited, and small features cannot be obtained;

[0228] The above errors are unavoidable in actual coal caving, but the angular resolution of the detector will also determine the accuracy of the online intelligent monitoring system for coal caving in the fully mechanized caving working face.

[0229] Specifically, angular resolution refers to the minimum angular interval between two adjacent targets that the online intelligent monitoring system for coal caving in the fully mechanized caving working face can distinguish. When the detector position is fixed, the measuring angle interval of the detector is fixed, and its angular resolution is the main factor determining the number of point clouds obtained. The relationship between the two is: the smaller the angular resolution, the more unit intervals obtained by angular resolution division, and the more point clouds obtained by measurement;

[0230] In view of this, if Fig. 9 As shown, the error analysis of the obtained coal flow Q is mainly to analyze the relationship between the coal flow Q and the detector angular resolution. The specific steps are:

[0231] 4.1. Assume that the area measured by the detector is S c , the actual area of ​​the coal flow cross section is S j , then the two have the following relationship:

[0232] S j =S c +aS - -bS +

[0233] Among them, S - is the missing area, indicating the surface features of the coal flow that cannot be described by the scanning profile;

[0234] a is the amount of missing area, indicating that the scanning profile describes too many surface features of the coal flow;

[0235] b is the amount of surplus area;

[0236] 4.2. Use volume relative error ε V Describing the error of coal flow Q, the specific formula is as follows:

[0237]

[0238] Among them, V s Actual measured volume for coal flow;

[0239] In step 4.2, the measured volume V and the relative error ε S The size change is mainly determined by the missing area and the surplus area, which are related to the detector angular resolution δθ, and the length of one side of the missing area and the surplus area is determined by the distance between the two point clouds. When the distance is smaller, the enclosed missing area or surplus area is smaller. Therefore, when the detector angular resolution δθ gradually decreases, the point cloud density increases, the unit area range decreases, and the scanning contour described by the point cloud becomes closer to the coal flow contour. The relative error ε s The smaller;

[0240] In step 4, it is necessary to determine the accuracy threshold of the coal release amount monitoring system. It is known from the prior art that the relative error of camera monitoring is between 10% and 18%. Therefore, this application sets the volume relative error threshold to 15%. This step analyzes the error, which can help to further optimize the coal release amount monitoring and improve the coal release accuracy in the later stage.

[0241] Step 5: Visualize the obtained coal flow. The specific process is as follows:

[0242] 5.1. Use Qt as UI design in PyCharm;

[0243] 5.2. Establish a TCP connection between the detector and QT so that QT can directly display the scanning screen of the detector;

[0244] 5.3. When the detector monitors the coal flow at the coal discharge port, the collected point cloud data of the coal flow is stored and calculated, and the 2D and 3D images of the coal flow are displayed until the collection stops;

[0245] In step 5, PyCharm is a Python integrated development environment with a complete set of tools, including the PyQT5 library. QT is a cross-platform C++ graphical user interface application development framework. In PyCharm, QT can be run, and when QT is connected to the detector TCP, it can be used to view the 2D and 3D images of the coal flow. When QT is used to render the surface contour of the coal flow, it is convenient. This step can further display the coal flow obtained by the detector intuitively, so that workers can observe the surface contour of the coal flow and the transportation status of the scraper conveyor in real time. At the same time, on the display screen, a message box on the right can be provided to display the coal release amount, coal release time and other information of the comprehensive caving working face in real time as needed, and the above information can be obtained by scanning the detector. This step realizes the visualization of coal release monitoring, and can intuitively monitor the real-time situation of coal release, thereby improving the practicality and intelligence level.

[0246] Verification experiment

[0247] like Fig. 9 As shown, in order to verify the effectiveness of the present invention, the coal flow monitoring method system of the fully mechanized caving working face, coal gangue samples, and other tools are provided for experiments;

[0248] Among them, the gangue samples are from the top coal of the Longwanggou 61607 fully mechanized caving working face. Gangue is the monitoring object, and the coal samples are divided into different block sizes;

[0249] Among them, the required scraper conveyor is replaced by a plastic film, aiming to restore the monitoring process of coal flow and remove the influence of different coal gangue reflectivity on the monitoring results in the experiment;

[0250] Other tools include, but are not limited to: computers, measuring buckets, searchlights;

[0251] The specific experimental process is:

[0252] K1. Select a cylindrical coal sample with a known volume as the scanning object, place it under the detector in two postures, vertical and horizontal, scan it, and record the scanning volume. Repeat the scan three times in each posture. According to the experimental results, Table 1-2 is obtained:

[0253] Table 1 (Scanning volume when coal sample is standing upright)

[0254]

[0255] As can be seen from Table 1, the three sets of data obtained by the detector scanning three coal samples of different volumes have small variances, which indicates that the data deviation obtained by coal flow monitoring is small, and the present invention has high stability. The average relative error obtained by scanning the three sets of coal samples is usually 0.51%, which is low and meets the requirements of coal flow monitoring.

[0256] Table 2 (Scanning volume when coal sample is placed horizontally)

[0257]

[0258] As shown in Table 2, in this measurement, it is found that the scanning volume of the horizontally placed coal sample increases significantly, and the average relative error exceeds 14%, which is much larger than the result of scanning the vertical coal sample. Fig.10 As shown in the figure, this is because after the coal sample is inverted, only the upper surface contour of the coal sample is scanned, and the lower surface contour of the coal sample is unknown. In the calculation process, the space below the coal sample should be calculated. Specifically, the scanned contour is calculated, and its cross-sectional area is 22.32cm 3 , while in actual situations, the cross-sectional area of ​​the coal sample is 19.63 cm 3 , the additional calculated volume is 2.67cm 3 , accounting for 13.69% of the actual cross-sectional area of ​​the coal sample. The scanning results of the horizontally placed coal sample were re-analyzed, and the actual errors of the three scans were 0.51%, 0.55%, and 0.51% respectively;

[0259] Therefore, combining the experimental results of the two scans, i.e., Table 1-2, it can be concluded that the system error of the detector is between 0.50% and 0.55%, and because the shape and placement of the coal sample change, the error between the scanned coal sample volume and the actual coal sample volume also changes accordingly;

[0260] K2, three states of coal flow: stacked scattered state, flat state, and pile state. The stacking length on the scraper conveyor is 1.5m each time. After the stacking is completed, the detector is started to scan and the volume of the scanned coal flow is recorded. After the scan is completed, the stacked coal flow is measured with a measuring bucket to obtain the actual volume of the coal flow and recorded on the experimental table. The coal flow in different stacking states is scanned three times repeatedly. According to the experimental results, Table 3 is obtained:

[0261] Table 3

[0262]

[0263] As shown in Table 3, the average relative error of the three types of coal piles of the same length using normal coal flow is between 4.78% and 5.44%, and the error difference is not large. The same length scraper conveyor lays coal flow in different stacking states, among which the scattered state lays the least coal flow, followed by flat laying and pile state. In this regard, when the coal flow volume scanned by the coal discharge monitoring system is lower than the set value, it can be inferred that the scraper conveyor stacking state is a scattered state;

[0264] The results show that different accumulation states of coal flow have little effect on the scanning results of the coal discharge monitoring system, and the influence of different accumulation states on the scanning can be ignored in actual operation;

[0265] K3. Use a measuring bucket to obtain 10L of coal flow with different block sizes, pile them up on the scraper conveyor in a flat manner, use a detector to scan the piled coal flow, and record the scanned coal flow volume. Repeat the scanning of coal flow with different block sizes three times. According to the experimental results, Table 4 is obtained:

[0266] Table 4

[0267] Coal flow density True volume (L) Measurement 1 (L) Measurement 2 (L) Measurement 3 (L) Coal powder 10 10.15 10.14 10.14 Small block 10 10.31 10.34 10.36 Medium 10 10.71 10.75 10.75 Large size 10 10.91 10.94 10.95

[0268] As can be seen from Table 4, when the present invention scans coal flows of different particle sizes, the change between measurement 1 and measurement 3 under the same particle size is small, but the change between different particle sizes is large. Fig.11 As shown in the figure, the relative errors in each group of measurements will be compared and analyzed to discuss the effect of block size on the same error in coal flow monitoring.

[0269] according to Fig.11 It can be seen that the scanning results of coal flow with different block sizes are very different. The relative error of the scanning results of coal flow with coal powder block size is less than 1%, and the error of the result is small, while the relative error of the scanning results of coal flow with large block size is about 9%, and the error of the result is large. As the block size of the coal flow increases, the relative error increases, and the two show a positive correlation.

[0270] Therefore, highly broken coal flow is more conducive to coal caving monitoring, and the error of its scanning results is lower;

[0271] K4. In a dark environment in the laboratory, arrange a 10L coal flow of coal powder particles on the scraper conveyor, adjust the optical machine in the binocular camera to high brightness and low brightness respectively, scan the coal flow under the coal powder particles, record the scanned coal flow volume, and then use a 30W bulb to supplement the laboratory light, repeat the above process, record the experimental data, and repeat the experiment three times under different lighting conditions;

[0272] In the above experiment:

[0273] In step K1, the regular object is first scanned to verify the feasibility of the coal flow monitoring method of the present invention and measure the system error of the detector;

[0274] In step K2, the scattered state, the flat state, and the pile state are the coal flow accumulation states obtained during the investigation of the Longwanggou 61607 fully mechanized caving working face. The coal flow in the same three states is piled up in the laboratory to verify the influence of the coal flow accumulation state on the monitoring of the present invention;

[0275] In step K3, the coal flows of different speeds are specifically: coal powder, small size, medium size, and large size, which are the coal flow size conditions obtained from the investigation of the Longwanggou 61607 fully mechanized caving working face. By monitoring the above four types of coal flow size, the influence of the coal flow size state on the monitoring of the present invention can be verified;

[0276] In step K4, experiments are conducted under different power optical machines and external lighting conditions to verify the influence of lighting on coal flow monitoring. Fig.12 As shown in Figure 2, different lighting conditions are divided into dark environment + low brightness, dark environment + high brightness, bright environment + low brightness, and bright environment + high brightness. The experimental results are as follows: Fig.13 As shown;

[0277] according to Fig.13 It can be concluded that when the optical machine is in low brightness, the point cloud data obtained by the detector scanning is seriously missing, and the calculated volumes are 1.35L and 3.32L, which are very different from the actual situation; while the point cloud data obtained by the optical machine scanning in high brightness is complete and continuous, and the surface contour of the coal flow is clearly described. The calculated volumes are 11.06L and 10.17L, and the relative errors are 10.6% and 1.7%, respectively, both within 15%, which are within the acceptable range. Therefore, the optical machine under high brightness is conducive to improving the scanning accuracy;

[0278] In addition, by comparing the scanning results obtained in different environments, under low brightness, the point cloud obtained by scanning in a dark environment has fewer points than that obtained by scanning in a bright environment; under high brightness, the local features obtained by scanning in a bright environment are clearer than those in a dark environment. Therefore, a bright environment is more conducive to the work of the detector, and the use of a high-brightness optical machine to add an external auxiliary light source for supplementary lighting is the most effective means of coal flow monitoring.

[0279] From the above experiments, we can conclude that:

[0280] The coal discharge monitoring system designed based on binocular camera shows good stability and measurement accuracy during scanning and is highly feasible. However, changes in the placement of coal samples will have a certain impact on the scanning results.

[0281] The experimental results show that the accumulation state of coal flow has little effect on the scanning accuracy of the system. However, under the same scraper conveyor length, the volume of the scattered coal flow is the smallest, so the accumulation state of the coal flow can be inferred by measuring the volume.

[0282] The size of coal flow has a great influence on the measurement results of the system, especially in the case of large pieces, the relative measurement error is large, while in the case of coal powder, the relative measurement error is the smallest;

[0283] The brightness of the optical machine has the greatest impact on the measurement results. Using a high-brightness optical machine and adding an external auxiliary light source for supplementary lighting is the most effective means of coal flow monitoring.

[0284] To summarize, the present invention calibrates the binocular camera to make a detector, and reasonably arranges a bracket equipped with the detector. When coal is released in the comprehensive caving working face, the coal flow information above the scraper conveyor can be obtained in real time. At the same time, the coal flow information is reconstructed in three dimensions to obtain three-dimensional point cloud data of the coal flow. After obtaining the coal flow scanning volume, the coal flow rate per unit time can be calculated, so that the staff knows the specific coal flow rate, thereby accurately monitoring the coal flow rate of the comprehensive caving working face and avoiding under-release or over-release. The invention ensures that the coal flow on the scraper conveyor is stable, which is beneficial to production.

[0285] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A coal flow monitoring method based on binocular vision, characterized in that: The following steps are involved: Step 1: Select two binocular cameras for calibration to make a detector; Step 2: Reasonably arrange the detector bracket equipped with the detector; Step 3: Based on the detector, the coal flow information above the scraper conveyor is obtained and the coal flow Q of the fully mechanized caving working face is calculated. The specific steps are as follows: 3.

1. The coal flow information above the scraper conveyor obtained by the detector is reconstructed in three dimensions to obtain three-dimensional point cloud data of the coal flow; 3.

2. Starting from t=0, with Δt as the time interval, the 3D point cloud data is segmented according to the scanning time, and a total of t / Δt segment data is obtained; 3.

3. Take any period of time Δt i Analyze the three-dimensional point cloud data in the section, and let the coal flow scanning volume corresponding to this section be V; 3.

4. The coal flow scanning volume V is calculated according to its area length z s Divided into multiple small units, the calculation formula of the coal flow scanning volume V is as follows: Among them, V i The volume of each small unit; Δz is the length of each small unit; 3.

5. Let the cross-sectional area of ​​the coal flow obtained by the detector scanning the coal flow be S i , using the differential method to calculate the cross-sectional area of ​​the coal flow is S i , the specific steps are as follows: 3.5.

1. Assume that the bottom is the scraper conveyor, the top is the cross-sectional profile of the coal flow, and establish a rectangular coordinate system with the optical center of the detector as the origin; 3.5.

2. Use point cloud data to divide the cross-sectional area of ​​the coal flow into n regions, and obtain the cross-sectional area S of the coal flow based on the starting and ending coordinates of the interval. i , the specific formula is as follows: Among them, x is the coal flow profile coordinate, y is the coordinate of the bottom contour of the scraper conveyor; 3.

6. According to the calculated coal flow cross-sectional area S i , the volume of all small units V i Calculate the sum and get the coal flow scanning volume V. The specific formula is as follows: 3.

7. Combining formulas (1), (2) and (3), we can get the formula for the coal flow scanning volume V, which is as follows: 3.

8. Calculate the additional coal flow volume V obtained by the detector in a single time x , and the calculation formula is as follows: V x =VV c in, V c The repeated partial volumes of the coal flow volume database acquired for the detector; 3.

9. The additional coal flow volume V obtained by a single pass of all detectors x The total volume of coal released is obtained by summing up, and the coal flow size per unit time, that is, the coal flow rate, is calculated as follows, and the calculation formula is as follows:

2. A method for monitoring coal flow based on binocular vision according to claim 1, characterized in that: In step 1, the binocular camera is selected as the MV-SUA630C-M industrial-grade camera, and the lens of the MV-SUA630C-M industrial-grade camera is selected as the M2808-1K-4 lens, and an active light source is added, and the active light source is selected as SDI 3005.

3. A method for monitoring coal flow based on binocular vision according to claim 2, characterized in that: In step 1, the binocular camera is calibrated, and the specific steps of making the detector are as follows: 1.

1. Prepare a calibration plate with a black and white chessboard paper inside; 1.

2. Place the calibration plate at the scanning position and adjust the two binocular cameras so that the calibration plate appears within the center of the field of view of both cameras at the same time. At this time, fix the binocular cameras. Among them, the fixed positions of the binocular cameras are one on the left and one on the right on the same straight line; 1.

3. Adjust the position of the calibration plate to change its image position in the binocular camera, thereby obtaining multiple sets of chessboard images at different positions and angles, and storing them in the disk; 1.

4. The binocular camera on the left is calibrated separately. The chessboard image acquired by the binocular camera on the left is imported into the MATLAB program. The specifications of the calibration plate are input. All vertices on the calibration plate are extracted using the MATLAB program. Then the parameters of the binocular camera on the left are calibrated. The three-dimensional position simulation diagram and parameters of the calibration plate are obtained, and then the calibration results of the binocular camera on the left are obtained. The specific results are as follows; The internal parameter matrix is: The radial distortion is: [-0.060.47] The correction matrix is: 1.

5. Import the chessboard image acquired by the right binocular camera into the MATLAB program, input the calibration plate specifications, use the MATLAB program to extract all vertices on the calibration plate, and then calibrate the parameters of the right binocular camera. The three-dimensional position simulation diagram and parameters of the calibration plate are obtained, and then the calibration results of the right binocular camera are obtained. The specific results are as follows; The internal parameter matrix is: The radial distortion is: [-0.1040.917] The correction matrix is: 1.

6. Input the chessboard images acquired by the two binocular cameras into the MATLAB program at the same time. After completing the vertex acquisition, the three-dimensional position simulation diagram of the calibration plate is obtained. At this time, the rotation matrix and translation matrix of the two binocular cameras are calculated to complete the calibration of the two binocular cameras. The specific results are as follows; the rotation matrix R of the two binocular cameras is: The translation matrix of the two binocular cameras is: T = [210.8023 12.9958 10.1175]; 1.

7. After the two binocular cameras have been calibrated, they can be directly installed side by side to make a detector.

4. A method for monitoring coal flow based on binocular vision according to claim 1, characterized in that: in step 2, the specific steps of rationally arranging the detector bracket equipped with the detector are: 2.

1. Calculate the maximum distance between the detector bracket installation position and the coal bracket. The specific calculation formula is as follows: Where, v is the running speed of the scraper conveyor, and v = 1.5m / s, Δt s It is the time interval between the top coal leaving the coal discharge port and the monitoring system scanning the coal flow. N a is the serial number on the target detector bracket, where the detector bracket at the fully mechanized caving working face is numbered 1, and increases outwards in sequence. N f It is the serial number on the coal caving support. The serial number of the detector support at the fully mechanized caving working face is 1, and it increases outwards in sequence. W y is the support width of the hydraulic support; 2.

2. Analyze the dust concentration of coal flow and get |N f -N a |The relationship is: 4≤|N f -N a |; 2.

3. Analyze the distance of coal flow movement and get |N f -N a Another relational expression for | is: 1≤|N f -N a |; 2.

4. Based on the analysis of the dust concentration of the coal flow and the distance of the coal flow movement, the specific location for installing the detector bracket is determined.

5. A method for monitoring coal flow based on binocular vision according to claim 4, characterized in that: in step 2.2, the specific steps of analyzing the dust concentration are: 2.2.

1. Determine the hydraulic support that needs to be caving coal at the caving working face. Assume that the hydraulic support is support No. 80, define the serial number of the hydraulic support as 80, and set dust concentration detection points from the 15th support on the left to the 15th support on the right with the hydraulic support as the center; 2.2.

2. After coal discharge begins, record the changes in wind direction and dust concentration at the working face during the coal discharge process.

6. The method for monitoring coal flow based on binocular vision according to claim 4, characterized in that: In step 2.3, the specific process of analyzing the distance of coal flow movement is as follows: The process of top coal discharge is divided into the first stage in which the top coal is broken and moves downward from the rear of the tail beam under the influence of gravity, the second stage in which the opening and closing of the coal discharge port is controlled by controlling the contraction of the plug plate, and the third stage in which the top coal is broken and contacts the upper surface of the rear scraper conveyor. And in the third stage: let v x is the lateral velocity of the top coal when it leaves the coal discharge port, v y is the lateral velocity of the top coal when it leaves the coal discharge port, v z is the movement speed of the top coal, v is the running speed of the scraper conveyor, μ is the surface friction coefficient of the scraper conveyor, and its specific value is 0.

07. The relationship between the distance d of coal flow movement and the distance d is as follows: Among them, v0 is the initial lateral velocity when the top coal contacts the scraper conveyor, which is 0m / s. g is the acceleration due to gravity, which is 9.8 m / s 2 , Substituting into the above formula we obtain: d = 1.64m.

7. The method for monitoring coal flow based on binocular vision according to claim 1, characterized in that: In step 3: Assume the scanning frequency of the binocular camera is f, and f = 18FPS, The time consumed by a single scan is Δt, and Δt=0.5s, The width of a single identification area is w, and the length of a single identification area on a scraper conveyor is z s , When the detector is installed at a height of 90 cm, the width of the detector's single identification area is w = 150 cm, and the length of the single identification area on the scraper conveyor is z s =90cm.

Images