An intelligent in-situ coal gangue sorting machine and method in a well

Abstract

The application discloses an underground coal gangue in-situ intelligent sorting machine and method, which comprises a sorting belt, a mechanical poking hand and an upper computer. The sorting belt is provided with a queuing mechanism capable of arranging the coal gangue into single-column marching. The sorting belt is provided with a weighing sensor, and the upper portion of the sorting belt is provided with a binocular synchronous camera. The sorting belt is provided with a coal conveying belt conveyor, and the lower portion of the sorting belt is provided with a gangue conveying scraper conveyor. The conveying directions of the coal conveying belt conveyor and the gangue conveying scraper conveyor are different. The sorting belt, the mechanical poking hand, the weighing sensor and the binocular synchronous camera are all connected with the upper computer. The upper computer judges the target category of the current position by collecting the data of the weighing sensor and the binocular synchronous camera. The speed and the precision of the coal gangue sorting are improved, the invalid gangue conveying distance can be maximally shortened, the unit output energy resource consumption is reduced, an effective carbon emission control valve is formed at the production source, and the green and low-carbon intelligent mining of the coal is realized.

Description

Technical Field

[0001] This invention relates to an in-situ intelligent sorting machine and method for underground coal gangue. Background Technology

[0002] Coal seam interbedded gangue and gangue from the roof and floor of the working face are important sources of gangue underground. If the gangue associated with raw coal is sorted and backfilled in situ underground, it can not only reduce carbon dioxide emissions caused by spontaneous combustion of gangue piles on the surface, but also minimize the ineffective transportation distance of gangue, reduce energy and resource consumption per unit output, form an effective carbon emission control valve at the source of production, and realize green, low-carbon and intelligent mining of coal.

[0003] Currently, underground coal gangue sorting methods are mainly divided into two types: dry sorting and wet sorting. Wet sorting includes jigging, shallow heavy medium trough sorting, and hydrocyclone sorting; dry sorting includes manual sorting, selective crushing, image recognition, multispectral recognition, radiographic recognition (X-ray or gamma-ray), and three-dimensional laser scanning. However, each of these existing methods has certain shortcomings, such as:

[0004] The wet sorting process used in the paper "Coal Mine Gangue Underground Separation and In-situ Backfilling Mining Method" published in the Journal of Coal Science by Zhang Jixiong, Ju Feng, Li Meng, and others is complex and the equipment is large. It usually requires the excavation of large-section chambers or underground spaces to accommodate the equipment, resulting in a large amount of engineering work and high maintenance costs. Moreover, manual sorting is labor-intensive, inefficient, and detrimental to the health of workers.

[0005] The selective crushing method used in Xu Chunyun's "Research on Key Technologies of Underground Hydraulic Selective Coal and Gangue Separation Equipment" has poor safety and is prone to problems such as sparks causing coal dust or gas explosions.

[0006] The X-ray identification technology used in Feng An'an's "Research on X-ray Identification Technology of Coal Gangue in Intelligent Sorting Process" has radiation risks and is subject to strict control.

[0007] In the image recognition process used in Zhao Haodi's "Research on Coal and Gangue Sorting Technology Based on Machine Vision", the appearance of coal and gangue is extremely similar when coal powder adheres to gangue or when the gangue surface is wet, making image recognition difficult. In addition, when the conveyor belt speed is high, the object being measured will be blurred when passing through the camera. All of these factors lead to a decrease in the quality of the captured images and low image recognition accuracy. Summary of the Invention

[0008] To address the aforementioned problems, this invention provides an in-situ intelligent coal gangue sorting machine and method for underground mining, which improves the speed and accuracy of coal gangue sorting, while also minimizing the ineffective transportation distance of gangue, reducing energy and resource consumption per unit output, forming an effective carbon emission control valve at the source of production, and realizing green, low-carbon, and intelligent coal mining.

[0009] To achieve the above-mentioned technical objectives and effects, the present invention is implemented through the following technical solution:

[0010] An in-situ intelligent sorting machine for coal gangue in underground mines includes a sorting belt, a mechanical lever, and a host computer. The inlet of the sorting belt is equipped with a queuing mechanism that can arrange the coal gangue into a single column for forward movement. The sorting belt is equipped with a weighing sensor and a binocular synchronous camera is installed above the sorting belt. The outlet of the sorting belt is equipped with a coal conveying belt conveyor and a gangue scraper conveyor is installed below the sorting belt. The coal conveying belt conveyor and the gangue scraper conveyor have different conveying directions.

[0011] The sorting belt, mechanical lever, weighing sensor, and binocular synchronous camera are all connected to the host computer. The host computer determines the target category at the current location by collecting data from the weighing sensor and the binocular synchronous camera.

[0012] If the current target is determined to be gangue, when the gangue reaches the gangue distribution trough, the mechanical hand will push the gangue onto the gangue conveyor.

[0013] Preferably, there are at least two sorting belts arranged in parallel, and each sorting belt is equipped with a set of mechanical levers, queuing mechanism, weighing sensor and binocular synchronous camera;

[0014] A sorting belt baffle is provided between the sorting belts, and a cavity with a wider top and a narrower bottom is provided below the sorting belts. The scraper conveyor for transporting waste is located below the cavity.

[0015] Preferably, the conveying direction of the coal conveying belt conveyor is the same as that of the sorting belt, and the conveying direction of the gangue scraper conveyor is opposite to that of the coal conveying belt conveyor.

[0016] Preferably, each sorting belt is equipped with a supplementary light.

[0017] Preferably, the host computer determines the target category at the current location by collecting data from the weighing sensor and the binocular synchronous camera, specifically including:

[0018] (1) The binocular synchronous camera was calibrated using the Zhang Zhengyou calibration method. After calibration, the left and right images of coal gangue captured by the binocular synchronous camera were stereo corrected according to the camera parameter information obtained. The stereo corrected left and right images were converted into grayscale images, and the images were denoised using median filtering of the 3×3 neighborhood.

[0019] (2) Perform SURF feature point detection and matching to obtain a set M of matching point pairs. Let the pixel coordinates of the i-th matching point pair in the left eye image feature point be (u i ,v iThe corresponding point obtained by matching in the right eye image is (u) i ',v i If '), then iterate through the set M of matching point pairs one by one, and remove |v i '-v i For matching pairs where |>2, retain |v i '-v i For each pair of matching points |≤2, obtain the set N of matching point pairs;

[0020] (3) Sequentially select the i-th left eye image feature point (u) from set N. i ,v i ) and right eye image feature points (u i ',v i Subtracting the x-coordinates of the i-th feature points from the x-coordinates of the i-th feature points yields the disparity d. i Combining the camera baseline B obtained from camera calibration and the camera focal length f, the (X) coordinates of the i-th feature point in the world coordinate system are calculated according to the following formula. i ,Y i Z i Coordinates, where the world coordinate system coincides with the camera coordinate system:

[0021]

[0022] In the formula, B is the distance between the centers of the two lenses of the binocular synchronous camera, i.e., the baseline; f is the focal length of the camera; d i d is the disparity of the i-th feature point. i =|u i -u i '|;where Z i It is the distance of the feature point from the camera;

[0023] (4) Perform Delaunay triangulation on the central (X,Y) coordinates of the three-dimensional coordinate dataset of feature points on the coal / gangue surface, and combine the data with the i-th feature point (X... i ,Y i The corresponding Z i Coordinates, using the concept of integration to calculate the volume V of coal gangue. m The details are as follows:

[0024] If Δ i Let be the i-th triangle, and let the vertices of this triangle correspond to the 3D coordinates of the feature points as (Xi1, Yi1, Zi1), (Xi2, Yi2, Zi2), and (Xi3, Yi3, Zi3). Take the minimum value among Zi1, Zi2, and Zi3 as Δ. i For each Δ corresponding to the height of the column, calculate the values ​​for each Δ in sequence. i The volume of the coal gangue is obtained by summing the volumes of the corresponding cylinders.

[0025] In the formula, S(Δ i Let be the area of ​​the i-th triangle;

[0026] (5) Calculate the density ρ of the coal gangue based on the mass m collected by the weighing sensor:

[0027]

[0028] (6) Calculate the image grayscale feature correction parameter F ir and density feature correction parameter F bi :

[0029]

[0030]

[0031] In the formula, G is the average ash value of coal gangue in the left and right grayscale images;

[0032] (7) Calculate the criterion parameter F for coal and gangue separation according to the following formula. fuse :

[0033] F fuse =a ir F ir +(1-a ir )F bi

[0034] In the formula, a ir For grayscale features F ir The weights; when F fuse When F ≥ 0.5, the target is judged to be gangue; when F fuse If the value is less than 0.5, the target is identified as coal.

[0035] Preferably, when the target is determined to be gangue, the waiting time T for the mechanical hand action is calculated based on the distance L between the camera's image position and the gangue sorting trough:

[0036]

[0037] In the formula, ν 速 The belt speed of the sorting belt.

[0038] A method for in-situ intelligent sorting of coal gangue in underground mines includes the following steps:

[0039] S01. After the raw coal is mined, it is transported to the vibrating screen by the scraper conveyor of the working face for screening. The pulverized coal falls directly into the coal conveying belt conveyor, while the blocky coal gangue enters the queuing mechanism of the coal gangue in-situ intelligent sorting machine. The queuing mechanism arranges the coal gangue into a single column and they enter the sorting belt in sequence. The conveying direction of the sorting belt is the same as that of the coal conveying belt conveyor.

[0040] S02. The binocular synchronous camera takes pictures of the passing coal gangue targets and the weighing sensor collects the mass data, which is then transmitted to the host computer.

[0041] S03. The host computer determines the target category at the current location by collecting data from the weighing sensor and the binocular synchronous camera, specifically including:

[0042] (S0301) The binocular synchronous camera is calibrated using the Zhang Zhengyou calibration method. After calibration, the left and right images of coal gangue captured by the binocular synchronous camera are stereo corrected according to the camera parameter information obtained. The stereo corrected left and right images are converted into grayscale images, and the images are denoised using a 3×3 neighborhood median filter.

[0043] (S0302) Perform SURF feature point detection and matching on the image to obtain a set M of matching point pairs. Let the pixel coordinates of the i-th matching point pair at the feature point of the left eye image be (u i ,v i The corresponding point obtained by matching in the right eye image is (u) i ',v i If '), then iterate through the set M of matching point pairs one by one, and remove |v i '-v i For matching pairs where |>2, retain |v i '-v i For each pair of matching points |≤2, obtain the set N of matching point pairs;

[0044] (S0303) Sequentially select the i-th left eye image feature point (u) from set N. i ,v i ) and right eye image feature points (u i ',v i Subtracting the x-coordinates of the i-th feature points from the x-coordinates of the i-th feature points yields the disparity d. i Combining the camera baseline B obtained from camera calibration and the camera focal length f, the (X) coordinates of the i-th feature point in the world coordinate system are calculated according to the following formula. i ,Y i Z i Coordinates, where the world coordinate system coincides with the camera coordinate system:

[0045]

[0046] In the formula, B is the distance between the centers of the two lenses of the binocular synchronous camera, i.e., the baseline; f is the focal length of the camera; d i d is the disparity of the i-th feature point. i =|u i -u i '|;where Z iIt is the distance of the feature point from the camera; (S0304) Perform Delaunay triangulation on the central (X,Y) coordinates of the three-dimensional coordinate dataset of the feature points on the coal / gangue surface, and combine it with the i-th feature point (X i ,Y i The corresponding Z i Coordinates, using the concept of integration to calculate the volume V of coal gangue. m The details are as follows:

[0047] If Δ i Let i be the i-th triangle, and let the vertices of this triangle correspond to the three-dimensional coordinates of the feature points as (X, X) and (X, X) respectively. i1 Y i1 Z i1 ), (X i2 Y i2 Z i2 ), (X i3 Y i3 Z i3 ), take Z i1 Z i2 Z i3 The minimum value is taken as Δ i For each Δ corresponding to the height of the column, calculate the values ​​for each Δ in sequence. i The volume of the coal gangue is obtained by summing the volumes of the corresponding cylinders.

[0048] In the formula, S(Δ i Let be the area of ​​the i-th triangle;

[0049] (S0305) Calculate the density ρ of the coal gangue based on the mass m collected by the weighing sensor:

[0050]

[0051] (S0306) Calculate the image grayscale feature correction parameter F ir and density feature correction parameter F bi :

[0052]

[0053]

[0054] In the formula, G is the average ash value of coal gangue in the left and right grayscale images;

[0055] (S0307) Calculate the criterion parameter F for coal and gangue separation according to the following formula. fuse :

[0056] F fuse =a ir F ir +(1-a ir)F bi

[0057] In the formula, a ir For grayscale features F ir The weights; when F fuse When F ≥ 0.5, the target is judged to be gangue; when F fuse If the value is less than 0.5, the target is identified as coal.

[0058] S04. When the target is determined to be gangue, calculate the waiting time T for the mechanical hand action based on the distance L between the camera's shooting position and the gangue separation chute:

[0059]

[0060] In the formula, ν 速 The belt speed for sorting belts;

[0061] S05. The mechanical hand-operated device displaces the gangue into the gangue conveyor according to the action time. The gangue conveyor transports the gangue in the opposite direction to the goaf for filling. The lump coal enters the coal conveyor belt and is transported out of the working face together with the pulverized coal.

[0062] Preferably, there are 2-6 sorting belts, and each sorting belt is equipped with a set of mechanical levers, queuing mechanism, weighing sensor and binocular synchronous camera.

[0063] The beneficial effects of this invention are:

[0064] This invention utilizes an intelligent coal gangue sorting machine arranged in the transport roadway of the working face for in-situ green and intelligent sorting of coal gangue. This invention enables coal gangue to be sorted and backfilled in situ at the coal mining face without leaving the mine. This not only reduces carbon dioxide emissions caused by spontaneous combustion of gangue piles on the ground, but also minimizes the ineffective transportation distance of gangue, reduces energy and resource consumption per unit output, and forms an effective carbon emission control valve at the source of production, realizing green, low-carbon, and intelligent mining of coal.

[0065] Meanwhile, this invention utilizes a weighing sensor and a binocular synchronous camera mounted on the sorting belt to optimize and correct the coal gangue sorting algorithm using image recognition and dynamic weighing, so as to achieve accurate identification and rapid sorting of coal gangue. Attached Figure Description

[0066] Figure 1 This is a schematic diagram of the structure of an underground coal gangue in-situ intelligent sorting machine according to the present invention;

[0067] Figure 2 This is a schematic diagram of a partial arrangement of the sorting belt in this invention;

[0068] Figure 3 This is a schematic diagram of the overall layout of the present invention;

[0069] Figure 4 This is the present invention. Figure 1 Cross-sectional view along Line II;

[0070] Figure 5 This is a layout diagram of the coal sorting machine above the coal bunker in the mining area according to the present invention;

[0071] The meanings of the labels in the attached diagram are as follows:

[0072] 1. Host computer; 2. Binocular synchronous camera; 3. Coal; 4. Gangue; 5. Weighing sensor; 6. Sorting belt; 7. Mechanical lever; 8. Coal conveying belt conveyor; 9. Gangue conveying scraper conveyor; 10. Queuing mechanism; 11. Gangue chute; 12. Sorting belt baffle; 13. Outer shell; 14. Sorting machine; 15. Vibrating screen; 16. Working face transport roadway; 17. Hydraulic support; 18. Coal mining machine; 19. Working face scraper conveyor; 20. Transfer conveyor; 21. Goaf; 22. Roof; 23. Floor; 24. Guide chute; 25. Main roadway; 26. Sorting chamber; 27. Gangue chute; 28. Coal chute; 29. ​​Coal bunker; 30. Gangue bunker; 31. Supplemental lighting. Detailed Implementation

[0073] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand the present invention and implement it. However, the embodiments are not intended to limit the present invention.

[0074] like Figures 1-2 As shown, an underground coal gangue in-situ intelligent sorting machine includes a sorting belt 6, a mechanical lever 7, and a host computer 1. Coal gangue blocks mined by the coal mining machine 18 enter the sorting belt 6 for coal and gangue separation. The inlet of the sorting belt 6 (… Figure 1 The right end of the sorting conveyor belt 6 is equipped with a queuing mechanism 10 that can arrange coal gangue into a single column for queuing. For example, a camera can be used in conjunction with a mechanical claw to arrange the coal gangue into a single column, or a gate can be used to arrange the coal gangue into a single column. Near the entrance of the sorting conveyor belt 6, a dynamic weighing sensor 5 is installed below it, a binocular synchronous camera 2 is installed above it, and a mechanical lever 7 is installed at the position of the gangue trough 11.

[0075] At the outlet of the sorting belt 6 ( Figure 1 A coal conveyor belt 8 is installed at the left end of the main roadway 25, and a scraper conveyor 9 is installed below the sorting belt 6. It should be noted that both the coal conveyor belt 8 and the scraper conveyor 9 are located within the main roadway 25, but their conveying directions are different. For example, with... Figure 4 For example, the coal conveying belt conveyor 8 and the sorting belt 6 are oriented in the same direction. Figure 4 The right end of the conveyor transports coal, while the scraper conveyor for transporting gangue moves in 9 directions. Figure 4 The left end is used to transport gangue for in-situ filling.

[0076] The sorting belt 6, mechanical lever 7, weighing sensor 5, and binocular synchronous camera 2 are all connected to the host computer 1. The host computer 1 determines the target (i.e., coal gangue) category at the current position by collecting data from the weighing sensor 5 and the binocular synchronous camera 2. If the current target is determined to be gangue 4, when gangue 4 reaches the position of the gangue trough 11 (i.e., a trough opened at the gangue discharge point of the sorting belt 6), the mechanical lever 7 pushes the gangue 4 down to the gangue transport scraper conveyor 9.

[0077] Since a single sorting belt 6 can only perform sorting in one row, multiple sorting belts 6 can be set up for synchronous sorting to improve the sorting speed, based on the mining capacity of the coal mining machine 18. That is, the sorting machine has at least two sorting belts 6 arranged side by side. The host computer 1 can be set up in total, but each sorting belt 6 needs to be equipped with a set of mechanical levers 7, queuing mechanism 10, weighing sensor 5, and binocular synchronous camera 2. Depending on the actual working conditions, supplementary lights 31 can be installed above each sorting belt 6.

[0078] If a single sorting belt 6 can meet the mine's daily sorting requirements, then there is no need to deploy more sorting belts 6. If a single sorting belt 6 cannot meet the mine's daily sorting requirements, then the number of sorting belts 6 can be increased in parallel or vertically to meet the sorting requirements. When multiple sorting belts 6 are operating simultaneously, to prevent mutual interference between the sorting belts 6, it is preferable to install sorting belt baffles 12 between the sorting belts 6. This is to facilitate the collection of gangue 4. Figure 1 In the middle, a cavity with a wider top and a narrower bottom is provided below the sorting belt 6, and the scraper conveyor 9 is located below the cavity. In order to protect the sorting accuracy and service life of the sorting machine, a partial outer shell 13 can be provided on the outside of the sorting machine.

[0079] The host computer 1 determines the target category at the current location by collecting data from the weighing sensor 5 and the binocular synchronous camera 2, specifically including:

[0080] 1. The Zhang Zhengyou calibration method was used to calibrate the binocular synchronous camera 2. After calibration, the left and right images of coal gangue captured by the binocular synchronous camera 2 were stereo corrected according to the solved camera parameter information. The stereo corrected left and right images were converted into grayscale images, and the images were denoised using median filtering of 3×3 neighborhood.

[0081] 2. Perform SURF feature point detection and matching to obtain a set M of matching point pairs. Let the pixel coordinates of the i-th matching point pair at the feature point of the left eye image be (u i ,v i The corresponding point obtained by matching in the right eye image is (u) i ',v iIf '), then iterate through the set M of matching point pairs one by one, and remove |v i '-v i For matching pairs where |>2, retain |v i '-v i For each pair of matching points |≤2, obtain a set N of matching point pairs, where all matching feature points in this set satisfy row alignment.

[0082] 3. Sequentially assign the i-th left-eye image feature point (u) from set N. i ,v i ) and right eye image feature points (u i ',v i Subtracting the x-coordinates of the i-th feature points from the x-coordinates of the i-th feature points yields the disparity d. i Combining the camera baseline B obtained from camera calibration and the camera focal length f, the (X) coordinates of the i-th feature point in the world coordinate system are calculated according to the following formula. i ,Y i Z i Coordinates, where the world coordinate system coincides with the camera coordinate system:

[0083]

[0084] In the formula, B is the distance between the centers of the two lenses of the binocular synchronous camera 2, i.e., the baseline; f is the focal length of the camera; d i d is the disparity of the i-th feature point. i =|u i -u i '|, where Z i If Z is the distance from the i-th feature point to the camera, then the height of the coal / gangue at that point is equal to the distance Z' from the camera to the table minus Z. i The method for calculating Z' is as follows: Select a feature point on the desktop, calculate its disparity in the left and right images, and then use the above formula to calculate the Z' coordinate of the point, which is the height of the camera from the desktop.

[0085] 4. After the above processing, a three-dimensional coordinate dataset of feature points on the coal / gangue surface is obtained. The coordinates in the three-dimensional coordinate dataset of feature points on the coal / gangue surface are then subjected to Delaunay triangulation (this is existing technology). This is combined with the coordinates of the i-th feature point (X... i ,Y i The corresponding Z i The coordinates are used to display the triangle as a 3D surface plot, thus obtaining the 3D surface plot of the coal gangue. The volume V of the coal gangue is then calculated using the integral method. m :

[0086] If Δ i Let i be the i-th triangle, and let the vertices of this triangle correspond to the three-dimensional coordinates of the feature points as (X, X) and (X, X) respectively. i1 Yi1 Z i1 ), (X i2 Y i2 Z i2 ), (X i3 Y i3 Z i3 ), take Z i1 Z i2 Z i3 The minimum value is taken as Δ i For each Δ corresponding to the height of the column, calculate the values ​​for each Δ in sequence. i The volume of the coal gangue is obtained by summing the volumes of the corresponding cylinders.

[0087] In the formula, S(Δ i Let be the area of ​​the i-th triangle;

[0088] 5. Calculate the density ρ of the coal gangue based on the mass m collected by the weighing sensor 5:

[0089]

[0090] 6. Because the binocular camera can only capture the surface morphology of the coal gangue, and the bottom of the coal gangue is not flat with some areas protruding, other areas are suspended, forming an inverted trapezoid. The surface height of the coal gangue corresponding to these suspended areas is higher than the actual value, resulting in an overestimation of the volume and a certain error in density recognition. Similarly, image grayscale recognition is also affected by the appearance characteristics of the coal gangue, leading to certain recognition errors. Therefore, this invention optimizes the sorting algorithm and calculates the image grayscale feature correction parameter F. ir and density feature correction parameter F bi :

[0091]

[0092]

[0093] In the formula, G is the average gray value of coal gangue in the left and right grayscale images, and the value range is [0, 100].

[0094] 7. Calculate the criterion parameter F for coal and gangue separation according to the following formula. fuse :

[0095] F fuse =a ir F ir +(1-a ir )F bi

[0096] In the formula, a ir For grayscale features F ir The weight, air The value of F ranges from 0 to 1; when F fuse When F ≥ 0.5, the target is determined to be gangue 4; when F fuse If the value is less than 0.5, the target is identified as Coal 3.

[0097] When the target is determined to be gangue 4, the waiting time T for the mechanical lever 7 is calculated based on the distance L between the shooting position of the binocular synchronous camera 2 and the gangue separation trough 11:

[0098]

[0099] In the formula, ν 速 The belt speed of the sorting belt 6.

[0100] Binocular synchronous cameras can simulate the visual structure of human eyes. By utilizing the parallax of target objects in a scene within the binocular structure, depth information is obtained, enabling 3D reconstruction of the scene and targets. The application of binocular synchronous cameras for coal gangue volume recognition focuses on coal gangue images captured by binocular synchronous camera 2. A surface irregularity object volume recognition algorithm based on SURF feature point matching is employed. Row alignment constraints are used to eliminate incorrectly matched feature point pairs, and the 3D coordinate dataset is optimized. A 2D Delaunay triangulation method is used to obtain a feature point triangular network, and its surface is 3D reconstructed. The coal gangue volume is calculated using integral methods, and the density of coal gangue passing through binocular synchronous camera 2 is calculated using weighing information. The recognition algorithm is then optimized and corrected, enabling high-precision, high-speed in-situ intelligent sorting of underground coal gangue. Correspondingly, such as... Figure 3-5 As shown, an in-situ intelligent sorting method for underground coal gangue includes the following steps:

[0101] S01. After the raw coal is extracted by the coal mining machine 18, it is sequentially transported by the working face scraper conveyor 19 to the transfer conveyor 20 and vibrating screen 15 for screening. Pulverized coal falls directly into the coal conveying belt conveyor 8, while lumpy coal gangue enters the queuing mechanism 10 of the in-situ intelligent coal gangue separator 14. The separator 14 is arranged in the working face transport roadway 16 for on-site coal gangue separation. The queuing mechanism 10 arranges the coal gangue into a single-row advancing state and sequentially enters the separation belt 6. The conveying direction of the separation belt 6 is the same as that of the coal conveying belt conveyor 8. Figure 4 In the middle, when the coal mining machine 18 is located below the hydraulic support 17 for mining, a working face transport roadway 16 is formed between the roof 22 and the floor 23.

[0102] S02, the binocular synchronous camera 2 takes pictures of the passing coal gangue targets and the weighing sensor 5 collects the mass data, which is then transmitted to the host computer 1.

[0103] S03, the host computer 1 determines the target category at the current location by collecting data from the weighing sensor 5 and the binocular synchronous camera 2, specifically including:

[0104] S0301. The Zhang Zhengyou calibration method is used to calibrate the binocular synchronous camera 2. After calibration, the left and right images of coal gangue captured by the binocular synchronous camera 2 are stereo corrected according to the solved camera parameter information. The stereo corrected left and right images are converted into grayscale images, and the images are denoised using median filtering of 3×3 neighborhood.

[0105] S0302. Perform SURF feature point detection and matching on the image to obtain a set M of matching point pairs. Let the pixel coordinates of the i-th matching point pair at the feature point of the left eye image be (u i ,v i The corresponding point obtained by matching in the right eye image is (u) i ',v i If '), then iterate through the set M of matching point pairs one by one, and remove |v i '-v i For matching pairs where |>2, retain |v i '-v i For each pair of matching points |≤2, obtain the set N of matching point pairs.

[0106] S0303, sequentially extract the i-th left eye image feature point (u) from set N. i ,v i ) and right eye image feature points (u i ',v i Subtracting the x-coordinates of the i-th feature points from the x-coordinates of the i-th feature points yields the disparity d. i Combining the camera baseline B obtained from camera calibration and the camera focal length f, the (X) coordinates of the i-th feature point in the world coordinate system are calculated according to the following formula. i ,Y i Z i Coordinates, where the world coordinate system coincides with the camera coordinate system:

[0107]

[0108] In the formula, B is the distance between the centers of the two lenses of the binocular synchronous camera 2, i.e., the baseline; f is the focal length of the camera; d i d is the disparity of the i-th feature point. i =|u i -u i '|;where Z i If Z is the distance from the i-th feature point to the camera, then the height of the coal / gangue at that point is equal to the distance Z' from the camera to the table minus Z. i The method for calculating Z' is as follows: Select a feature point on the desktop, calculate its disparity in the left and right images, and then use the above formula to calculate the Z' coordinate of the point, which is the height of the camera from the desktop.

[0109] S0304. Perform Delaunay triangulation on the central (X,Y) coordinates of the three-dimensional coordinate dataset of feature points on the coal / gangue surface, and combine the data with the i-th feature point (X... i ,Y i The corresponding Z i Coordinates, using the concept of integration to calculate the volume V of coal gangue. m The details are as follows:

[0110] If Δ i Let i be the i-th triangle, and let the vertices of this triangle correspond to the three-dimensional coordinates of the feature points as (X, X) and (X, X) respectively. i1 Y i1 Z i1 ), (X i2 Y i2 Z i2 ), (X i3 Y i3 Z i3 ), take Z i1 Z i2 Z i3 The minimum value is taken as Δ i For each Δ corresponding to the height of the column, calculate the values ​​for each Δ in sequence. i The volume of the coal gangue is obtained by summing the volumes of the corresponding cylinders.

[0111] In the formula, S(Δ i Let be the area of ​​the i-th triangle.

[0112] S0305. Based on the mass m of the coal gangue collected by the weighing sensor 5, calculate the density ρ of the coal gangue:

[0113]

[0114] S0306. Calculate the image grayscale feature correction parameter F. ir and density feature correction parameter F bi :

[0115]

[0116]

[0117] In the formula, G is the average ash value of coal gangue in the left and right grayscale images.

[0118] S0307. Calculate the criterion parameter F for coal and gangue separation according to the following formula. fuse :

[0119] F fuse =a ir F ir +(1-a ir )F bi

[0120] In the formula, a ir For grayscale features F ir The weights; when F fuse When F ≥ 0.5, the target is determined to be gangue 4; when F fuse If the value is less than 0.5, the target is identified as coal.

[0121] S04. When the target is determined to be gangue 4, the waiting time T for the mechanical lever 7 is calculated based on the distance L between the shooting position of the binocular synchronous camera 2 and the gangue separation trough 11:

[0122]

[0123] In the formula, ν 速 To determine the belt speed of the sorting belt 6;

[0124] S05, the mechanical handpiece 7, according to the action time, pushes the gangue 4 into the gangue transport scraper conveyor 9. The gangue transport scraper conveyor 9 transports the gangue 4 in the opposite direction to the goaf 21 for filling, while the lump coal enters the coal transport belt conveyor 8 and is transported out of the working face together with the pulverized coal.

[0125] The sorting belts 6 consist of 2-6 belts, each equipped with a set of mechanical levers 7, queuing mechanism 10, weighing sensor 5, and binocular synchronous camera 2.

[0126] Depending on the specific conditions of different mines, a sorting machine can also be installed above the coal bunker 29 in the mining area. When the sorting system is installed above the coal bunker 29 in the mining area, it can not only serve the entire mining area, but also sort the raw coal from the coal mining face and the tunneling face together, eliminating the need for separate sorting systems for the coal mining and tunneling faces. When the raw coal is transported to the front of the coal bunker in the mining area, it is first screened by the vibrating screen 15. The pulverized coal falls into the coal conveyor belt 8 connected below the vibrating screen 15 and is then poured into the coal bunker 29. Lump coal and gangue enter the guide chute 24 from the vibrating screen, and are then diverted by the guide chute 24 to each sorting chamber 26 for sorting. The sorted gangue can be immediately returned to the goaf for filling or stored in the gangue bin 30 for later use through the gangue chute 27. The sorted lump coal flows into the coal bunker 29 through the coal chute 28 and is transported to the surface together with the pulverized coal.

[0127] This invention utilizes an intelligent coal gangue sorting machine arranged in the transport roadway of the working face for in-situ green and intelligent sorting of coal gangue. This invention enables coal gangue to be sorted and backfilled in situ at the coal mining face without leaving the mine. This not only reduces carbon dioxide emissions caused by spontaneous combustion of gangue piles on the ground, but also minimizes the ineffective transportation distance of gangue, reduces energy and resource consumption per unit output, and forms an effective carbon emission control valve at the source of production, realizing green, low-carbon, and intelligent mining of coal.

[0128] Meanwhile, this invention utilizes a weighing sensor and a binocular synchronous camera mounted on the sorting belt to optimize and correct the coal gangue sorting algorithm using image recognition and dynamic weighing, so as to achieve accurate identification and rapid sorting of coal gangue.

[0129] The above are merely preferred embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. An in-situ intelligent sorting machine for underground coal gangue, comprising a sorting belt (6), a mechanical lever (7), and a host computer (1), characterized in that, The sorting belt (6) is equipped with a queuing mechanism (10) at its inlet to arrange coal gangue into a single column for forward movement. A weighing sensor (5) is installed on the sorting belt (6), and a binocular synchronous camera (2) is installed above it. A coal conveying belt conveyor (8) is installed at the outlet of the sorting belt (6), and a gangue scraper conveyor (9) is installed below it. The coal conveying belt conveyor (8) and the gangue scraper conveyor (9) transport... The directions of delivery are different; the sorting belt (6), mechanical hand (7), weighing sensor (5) and binocular synchronous camera (2) are all connected to the host computer (1). The host computer (1) determines the target category at the current position by collecting data from the weighing sensor (5) and the binocular synchronous camera (2): if the current target is determined to be gangue (4), when the gangue (4) reaches the position of the gangue sorting trough (11), the mechanical hand (7) pushes the gangue (4) to the gangue transport scraper conveyor (9); The host computer (1) determines the target category at the current location by collecting data from the weighing sensor (5) and the binocular synchronous camera (2), specifically including: (1) The binocular synchronous camera (2) was calibrated using the Zhang Zhengyou calibration method. After calibration, the left and right images of coal gangue captured by the binocular synchronous camera (2) were stereo-corrected according to the camera parameter information obtained. The stereo-corrected left and right images were converted into grayscale images, and the images were denoised using median filtering of the 3×3 neighborhood. (2) Perform SURF feature point detection and matching to obtain a set M of matching point pairs. Let the pixel coordinates of the i-th matching point pair in the left eye image feature point be (u i ,v i The corresponding point obtained by matching in the right eye image is (u) i ',v i If '), then iterate through the set M of matching point pairs one by one, and remove |v i '-v i For matching pairs where |>2, retain |v i '-v i For each pair of matching points |≤2, obtain the set N of matching point pairs; (3) Sequentially select the i-th left eye image feature point (u) from set N. i ,v i ) and right eye image feature points (u i ',v i Subtracting the x-coordinates of the i-th feature points from the x-coordinates of the i-th feature points yields the disparity d. i Combining the camera baseline B obtained from camera calibration and the camera focal length f, the (X) coordinates of the i-th feature point in the world coordinate system are calculated according to the following formula. i ,Y i Z i Coordinates, where the world coordinate system coincides with the camera coordinate system: In the formula, B is the distance between the centers of the two lenses of the binocular synchronous camera (2), i.e., the baseline; f is the focal length of the camera; d i d is the disparity of the i-th feature point. i =|u i -u i '|;where Z i It is the distance of the feature point from the camera; (4) Perform Delaunay triangulation on the central (X,Y) coordinates of the three-dimensional coordinate dataset of feature points on the coal / gangue surface, and combine the data with the i-th feature point (X... i ,Y i The corresponding Z i Coordinates, using the concept of integration to calculate the volume V of coal gangue. m The details are as follows: If Δ i Let i be the i-th triangle, and let the vertices of this triangle correspond to the three-dimensional coordinates of the feature points as (X, X) and (X, X) respectively. i1 Y i1 Z i1 ), (X i2 Y i2 Z i2 ), (X i3 Y i3 Z i3 ), take Z i1 Z i2 Z i3 The minimum value is taken as Δ i For each Δ corresponding to the height of the column, calculate the values ​​for each Δ in sequence. i The volume of the coal gangue is obtained by summing the volumes of the corresponding cylinders. In the formula, S(Δ i Let be the area of ​​the i-th triangle; (5) Calculate the density ρ of the coal gangue based on the mass m collected by the weighing sensor (5): (6) Calculate the image grayscale feature correction parameter F ir and density feature correction parameter F bi : In the formula, G is the average ash value of coal gangue in the left and right grayscale images; (7) Calculate the criterion parameter F for coal and gangue separation according to the following formula. fuse : F fuse =a ir F ir +(1-a ir )F bi In the formula, a ir For grayscale features F ir The weights; when F fuse When F ≥ 0.5, the target is judged to be gangue (4); when F fuse If the value is less than 0.5, the target is judged to be coal (3).

2. The underground coal gangue in-situ intelligent sorting machine according to claim 1, characterized in that, The sorting belts (6) are at least two and arranged in parallel. Each sorting belt (6) is equipped with a set of mechanical levers (7), queuing mechanism (10), weighing sensor (5) and binocular synchronous camera (2). A sorting belt baffle (12) is provided between the sorting belts (6), and a cavity with a wider upper section and a narrower lower section is provided below the sorting belts (6). The scraper conveyor (9) is located below the cavity.

3. The underground coal gangue in-situ intelligent sorting machine according to claim 2, characterized in that, The conveying direction of the coal conveying belt conveyor (8) is the same as that of the sorting belt (6), and the conveying direction of the gangue scraper conveyor (9) is opposite to that of the coal conveying belt conveyor (8).

4. The underground coal gangue in-situ intelligent sorting machine according to claim 2, characterized in that, Each sorting belt (6) is equipped with a supplementary light (31).

5. The underground coal gangue in-situ intelligent sorting machine according to claim 1, characterized in that, When the target is determined to be gangue (4), the waiting time T for the mechanical lever (7) is calculated based on the distance L between the shooting position of the binocular synchronous camera (2) and the gangue sorting trough (11): In the formula, v 速 The belt speed of the sorting belt (6).

6. A method for in-situ intelligent sorting of coal gangue in underground mines, characterized in that, Includes the following steps: S01. After the raw coal is mined, it is transported to the vibrating screen (15) by the working face scraper conveyor (19) for screening. The pulverized coal falls directly into the coal conveying belt conveyor (8), while the blocky coal gangue enters the queuing mechanism (10) of the coal gangue in-situ intelligent sorting machine (14). The queuing mechanism (10) arranges the coal gangue into a single-row state and enters the sorting belt (6) in sequence. The conveying direction of the sorting belt (6) is the same as that of the coal conveying belt conveyor (8). S02, the binocular synchronous camera (2) takes pictures of the passing coal gangue targets and the weighing sensor (5) collects the mass data and transmits the data to the host computer (1); S03, The host computer (1) determines the target category at the current location by collecting data from the weighing sensor (5) and the binocular synchronous camera (2), specifically including: (S0301) The binocular synchronous camera (2) was calibrated using the Zhang Zhengyou calibration method. After calibration, the left and right images of coal gangue captured by the binocular synchronous camera (2) were stereo-corrected according to the camera parameter information obtained. The stereo-corrected left and right images were converted into grayscale images, and the images were denoised using median filtering of the 3×3 neighborhood. (S0302) Perform SURF feature point detection and matching on the image to obtain a set M of matching point pairs. Let the pixel coordinates of the i-th matching point pair at the feature point of the left eye image be (u i ,v i The corresponding point obtained by matching in the right eye image is (u) i ',v i If '), then iterate through the set M of matching point pairs one by one, and remove |v i '-v i For matching pairs where |>2, retain |v i '-v i For each pair of matching points |≤2, obtain the set N of matching point pairs; (S0303) Sequentially select the i-th left eye image feature point (u) from set N. i ,v i ) and right eye image feature points (u i ',v i Subtracting the x-coordinates of the i-th feature points from the x-coordinates of the i-th feature points yields the disparity d. i Combining the camera baseline B obtained from camera calibration and the camera focal length f, the (X) coordinates of the i-th feature point in the world coordinate system are calculated according to the following formula. i ,Y i Z i Coordinates, where the world coordinate system coincides with the camera coordinate system: In the formula, B is the distance between the centers of the two lenses of the binocular synchronous camera (2), i.e., the baseline; f is the focal length of the camera; d i d is the disparity of the i-th feature point. i =|u i -u i '|;where Z i It is the distance of the feature point from the camera; (S0304) Perform Delaunay triangulation on the central (X,Y) coordinates of the three-dimensional coordinate dataset of feature points on the coal / gangue surface, and combine the i-th feature point (X... i ,Y i The corresponding Z i Coordinates, using the concept of integration to calculate the volume V of coal gangue. m The details are as follows: If Δ i Let i be the i-th triangle, and let the vertices of this triangle correspond to the three-dimensional coordinates of the feature points as (X, X) and (X, X) respectively. i1 Y i1 Z i1 ), (X i2 Y i2 Z i2 ), (X i3 Y i3 Z i3 ), take Z i1 Z i2 Z i3 The minimum value is taken as Δ i For each Δ corresponding to the height of the column, calculate the values ​​for each Δ in sequence. i The volume of the coal gangue is obtained by summing the volumes of the corresponding cylinders. In the formula, S(Δ i Let be the area of ​​the i-th triangle; (S0305) Calculate the density ρ of the coal gangue based on the mass m collected by the weighing sensor (5): (S0306) Calculate the image grayscale feature correction parameter F ir and density feature correction parameter F bi : In the formula, G is the average ash value of coal gangue in the left and right grayscale images; (S0307) Calculate the criterion parameter F for coal and gangue separation according to the following formula. fuse : F fuse =a ir F ir +(1-a ir )F bi In the formula, a ir For grayscale features F ir The weights; when F fuse When F ≥ 0.5, the target is judged to be gangue (4); when F fuse If the value is less than 0.5, the target is identified as coal. S04. When the target is determined to be gangue (4), the waiting time T for the mechanical lever (7) is calculated based on the distance L between the shooting position of the binocular synchronous camera (2) and the gangue sorting trough (11): In the formula, v 速 The belt speed of the sorting belt (6); S05. The mechanical handpiece (7) moves the gangue (4) into the gangue conveyor (9) according to the action time. The gangue conveyor (9) transports the gangue (4) in the opposite direction to the goaf (21) for filling. The lump coal enters the coal conveyor belt (8) and is transported out of the working face together with the pulverized coal.

7. The method for in-situ intelligent sorting of coal gangue in underground mines according to claim 6, characterized in that, The sorting belts (6) consist of 2-6 belts, each of which is equipped with a mechanical lever (7), a queuing mechanism (10), a weighing sensor (5), and a binocular synchronous camera (2).

Images