CN114535063B - Underground coal gangue sorting integrated device based on artificial intelligent image recognition - Google Patents

Abstract

The invention discloses an underground coal gangue sorting integrated device based on artificial intelligent image recognition. The device mainly comprises a conveying device, a particle size sorting device and a coal gangue identification and separation device; the conveying device mainly comprises a conveying belt and is responsible for transferring coal materials in the device; the particle size sorting device mainly comprises screens with different apertures and is used for sorting coal materials with different particle sizes; the gangue identification and separation device mainly comprises a camera, a weighing device, an embedded computer and a controllable rotating baffle plate, and can accurately identify coal and gangue by image identification and weighing based on artificial intelligence, and then can knock down the gangue by the controllable rotating baffle plate, thereby achieving the aim of gangue separation. The device has simple structure, large processible grain diameter range and high sorting accuracy, and reduces labor cost; gangue is separated before the coal is discharged from the well, so that the well discharge amount of the gangue is effectively reduced.

Description

An integrated device for underground coal and gangue sorting based on artificial intelligence image recognition

technical field

The invention relates to an integrated underground coal gangue sorting device based on artificial intelligence image recognition, belonging to the technical field of coal mining.

Background technique

For a long time, in the practice of coal mine production, the removal of gangue from the mined raw coal is mainly realized by washing the gangue entrained in the raw coal to the ground and then washing it in the coal preparation plant. However, with the continuous production, the sorted waste rocks are piled up on the ground, which is called gangue hills, which not only brings extra transportation and processing costs for waste rocks, but also causes land occupation by waste dumps and ground environmental pollution. question. It has become a key scientific problem urgently to be solved to make coal mining gangue "not up the well", save energy and reduce emissions, and recycle underground gangue. Therefore, promoting the development and improvement of underground separation technology is the only way for the future development of the coal industry.

Contents of the invention

The present invention aims to provide an integrated underground coal and gangue separation device based on artificial intelligence image recognition with simple structure, convenient operation, strong reliability and good sorting effect.

The invention provides an integrated device for underground coal and gangue sorting based on artificial intelligence image recognition, which includes an iron shell, a transmission device, a particle size sorting device, and a coal and gangue identification and separation device; Reprinted in the device; the particle size sorting device includes screens with different apertures for sorting coal materials of different particle sizes; the coal gangue identification and separation device includes a camera, a weighing device, an embedded computer and a controllable rotary gear Through artificial intelligence-based image recognition and weighing, the coal and gangue can be accurately identified, and then the gangue can be shot down through the controllable rotating baffle, so as to achieve the purpose of coal gangue sorting.

Further, the conveying device is composed of a low conveyor belt, a lifting conveyor belt, a high conveyor belt, an extra large lump coal conveyor belt, a large lump coal conveyor belt, a medium lump coal conveyor belt, a coal ash conveyor belt, a gangue conveyor belt, an iron chute and a fixed baffle; The low conveyor belt in the conveying device mentioned above is connected with the high conveyor belt after being connected with the lifting conveyor belt;

The particle size sorting device is composed of a shovel plate, a V-shaped groove, a large-aperture screen, a medium-aperture screen, a small-aperture screen, a ceramic chute and a screen support; The upper surface slides; the V-shaped groove is located on the edge of the lower side of the screen;

The large-aperture screen is located under the high-position conveyor belt, the medium-aperture screen is located under the large-aperture screen, and the small-aperture screen is located under the medium-aperture screen, and the screens are all obliquely fixed on the screen support , the vibration is controlled by the screen support; the large-aperture screen and a V-shaped groove are connected with the extra-large lump coal conveyor belt through a ceramic chute, and the medium-aperture screen and the small-aperture screen are respectively connected with the large lump coal in the same way. The conveyor belt is connected with the medium lump coal conveyor belt; the screen is inclined to the direction parallel to the rolling of the coal block and the direction perpendicular to the rolling of the coal block at the same time, and the two directions are inclined 10°-20°; the V-shaped groove is placed obliquely with the screen , so that coal blocks larger than the aperture of the screen can flow from the screen to the V-shaped groove and quickly roll down, thereby speeding up the sorting speed.

The coal and gangue sorting and identification device is composed of a camera, a controllable rotating baffle, a fill light, a weighing device and an embedded computer; the embedded computer communicates with the camera, the fill light, the weighing device and the The controllable rotating baffle is connected;

A video camera and a supplementary light are fixed above the extra large lump coal conveyor belt, large lump coal conveyor belt and medium lump coal conveyor belt, and the supplementary light is close to the camera; the super large lump coal conveyor belt, large lump coal conveyor belt and There is a controllable rotating baffle on the medium lump coal conveyor belt; there is a weighing device under the extra large lump coal conveyor belt, the large lump coal conveyor belt and the medium lump coal conveyor belt;

The coal ash conveyor belt is located under the small-diameter screen; the gangue conveyor belt is located under the iron chute, and a fixed baffle is set on the gangue conveyor belt corresponding to the position of the iron chute; the cameras all have self-cleaning functions ;

Except that a part of the low conveyor belt, the coal ash conveyor belt and the gangue conveyor belt are located outside the iron shell, other parts are all located inside the iron shell.

Furthermore, the lifting conveyor belt uses a belt with a large inclination angle, so that the device can lift coal materials in a small space;

Large-diameter screens, medium-diameter screens, and small-diameter screens preliminarily sort coal materials with different particle sizes to improve the accuracy of image recognition and reduce the amount of subsequent sorting work; and the screens are placed obliquely and Equipped with a screen support with a vibrating device, and the matching shovel plate can remove the coal stuck on the screen to speed up the separation speed of coal with different particle sizes; the screen support is composed of iron rods and vibration motors, and the vibration motor The vibration of the iron rod is controlled to make the screen vibrate.

The aperture sizes of the large-aperture screen, medium-aperture screen and small-aperture screen are 100mm, 50mm, and 25mm respectively, so that the coal on the extra-large coal conveyor belt, the large-lump coal conveyor belt, and the medium-lump coal conveyor belt are respectively extra-large pieces. Coal, large lump coal and medium lump coal; coal on the coal ash conveyor belt is small lump coal, granular coal and pulverized coal.

Furthermore, the embedded computer can distinguish coal and gangue through the artificial intelligence training model through the texture features of coal and gangue in the image obtained by the camera, and can estimate the volume according to the size of coal and gangue in the image;

The embedded computer can adjust the rotation speed of the extra-large lump coal conveyor belt, the large lump coal conveyor belt and the medium lump coal conveyor belt according to the density of coal and gangue blocks in the image obtained by the camera through the artificial intelligence training model.

The present invention provides a kind of adopting above-mentioned device to carry out underground coal gangue sorting method, comprises the following steps:

a. Lift the coal material to the top of the large-aperture screen through the low-level conveyor belt, lifting conveyor belt, and high-level conveyor belt;

b. Use large-diameter screens, medium-diameter screens and small-diameter screens for screening according to the particle size of the coal material;

c. According to the particle size of the coal material, the sieved coal material enters the extra large lump coal conveyor belt, the large lump coal conveyor belt, the medium lump coal conveyor belt and the coal ash conveyor belt through the ceramic chute for subsequent sorting;

d. Judging coal and gangue by weighting the results of texture feature recognition based on artificial intelligence images and the results of weighing and calculating density:

The density of coal and gangue is calculated according to the weighing device and the estimated volume, and the difference between coal and gangue can also be distinguished according to the difference in density. Then the results of weighted image recognition texture feature distinction and density distinction are used as the criterion for distinguishing coal and gangue;

The embedded computer is manually intervened in its identification results at the initial stage of use, and the model is trained while working, so that the accuracy of identification is getting higher and higher;

e. Through the controllable rotating baffle, the gangue is knocked down to the lowest gangue conveyor belt and transported out;

f. Extra large lump coal, large lump coal and medium lump coal are transported out through the extra large lump coal conveyor belt, large lump coal conveyor belt and medium lump coal conveyor belt, and small lump coal, granular coal and pulverized coal are transported out through the coal ash conveyor belt.

Beneficial effects of the present invention:

(1) The present invention uses a multi-stage screen structure to make up for the defect in artificial intelligence image recognition that the recognition accuracy rate is reduced due to the large difference in particle size;

(2) The present invention uses the structure of multi-stage screens to classify coal blocks according to the market, and sorts out extra large lump coal, large lump coal, medium lump coal and coal materials smaller than medium lump coal, reducing the subsequent classification. selection;

(3) The present invention uses artificial intelligence image recognition to simultaneously calculate the texture features and volumes of coal and gangue, and the density of coal and gangue can be obtained by integrating the data of the weighing device. The results obtained by weighting density and image texture features are more accurate than the existing single recognition image texture features.

Description of drawings

Figure 1 is a schematic diagram of an integrated underground coal and gangue sorting device based on artificial intelligence image recognition.

In the figure: 1a—low conveyor belt; 1b—lifting conveyor belt; 1c—high conveyor belt; 1d—extra large lump coal conveyor belt; 1e—large lump coal conveyor belt; 1f—medium lump coal conveyor belt; 1g—coal ash conveyor belt; 2—iron shell; 3—shovel plate; 4—V-shaped groove; 5a—large aperture screen; 5b—medium aperture screen; 5c—small aperture screen; 6—ceramic chute; 7—camera; 8 —Controllable rotating baffle; 9—iron chute; 10—fixed baffle; 11—screen support; 12—fill light; 13—weighing device; 14—embedded computer.

Detailed ways

An embodiment of the present invention will be further described below in conjunction with the accompanying drawings, but not limited to the following embodiment.

As shown in Fig. 1, the integrated underground coal and gangue sorting device based on artificial intelligence image recognition of the present invention includes an iron casing, a transmission system, a particle size sorting system and a coal and gangue identification and separation system.

As shown in Fig. 1, the integrated underground coal and gangue sorting device based on artificial intelligence image recognition of the present invention includes an iron casing, a transmission system, a particle size sorting system and a coal and gangue identification and separation system.

The conveying system consists of a low conveyor belt 1a, a lifting conveyor belt 1b, a high conveyor belt 1c, an extra large lump coal conveyor belt 1d, a large lump coal conveyor belt 1e, a medium lump coal conveyor belt 1f, a coal ash conveyor belt 1g, gangue conveyor belt 1h, iron chutes 9 and Fixed baffle 10 is formed;

The particle size sorting system is composed of a shovel plate 3, a V-shaped groove 4, a large-aperture screen 5a, a medium-aperture screen 5b, a small-aperture screen 5c, a ceramic chute 6 and a screen support 11;

The gangue sorting and identification system is composed of a camera 7, a controllable rotating baffle 8, a fill light 12, a weighing device 13 and an embedded computer 14;

The low-level conveyor belt 1a in the described transmission system connects with the lifting conveyor belt 1b and then connects with the high-level conveyor belt 1c;

The large-aperture screen 5a is located below the high-position conveyor belt 1c, the medium-aperture screen 5b is located below the large-aperture screen 5a, and the small-aperture screen 5c is located below the medium-aperture screen 5b, and the screens are all inclined Fixed on the screen support 11, the vibration is controlled by the screen support 11;

The shovel plate 3 can slide on the upper surface of the screen cloth 5a-5c; the V-shaped groove 4 is located at the edge of the lower side of the screen cloth 5; It is inclined at the same time as the rolling direction of the coal block, and both directions are inclined 10°-20°; the V-shaped groove is placed obliquely with the screen, so that the coal block larger than the aperture of the screen can flow from the screen to the V-shaped groove Roll off quickly, thus speeding up sorting.

The large-aperture screen 5a and a V-shaped groove 4 are connected with the extra-large coal conveyor belt 1d through the ceramic chute 6, and the medium-aperture screen 5b and the small-aperture screen 5c are respectively connected with the large coal conveyor belt in the same way. 1e links to each other with medium lump coal conveyor belt 1f;

A video camera 7 and a supplementary light 12 are fixed above the extra large lump coal conveyor belt 1d, large lump coal conveyor belt 1e and medium lump coal conveyor belt 1f, and the supplementary light 12 is close to the camera 7;

There is a controllable rotating baffle plate 8 on the extra large lump coal conveyor belt 1d, large lump coal conveyor belt 1e and medium lump coal conveyor belt 1f;

There is a weighing device 13 under the extra large lump coal conveyor belt 1d, the large lump coal conveyor belt 1e and the medium lump coal conveyor belt 1f;

The embedded computer 14 is connected to the video camera 7, the fill light 12, the weighing device 13 and the controllable rotating baffle 8 through cables;

The coal ash conveyor belt 1g is located below the small-diameter screen 5c; the gangue conveyor belt 1h is located below the iron chute 9 and is equipped with a fixed baffle 10; the cameras 7 have self-cleaning functions;

Said casing 2 covers the whole device except part of the conveying system. Except that a part of the low conveyor belt 1a, the coal ash conveyor belt 1g and the gangue conveyor belt 1h are located outside the iron shell, other parts are all located in the iron shell 2 .

Further, the lifting conveyor belt 1b uses a belt with a large inclination angle, so that the device can lift coal material in a small space.

The large-aperture screen 5a, medium-aperture screen 5b, and small-aperture screen 5c preliminarily sort coal materials with different particle sizes to improve the accuracy of image recognition and reduce the amount of subsequent sorting work; and the screens are all inclined It is placed and equipped with a screen support with a vibrating device, and the matching shovel plate 3 can remove the coal stuck on the screen to speed up the separation speed of coal with different particle sizes; the screen support is composed of iron rods and vibration motors Composition, the vibration motor controls the vibration of the iron rod to make the screen vibrate.

The apertures of the large-aperture screen 5a, the medium-aperture screen 5b and the small-aperture screen 5c are 100mm, 50mm, and 25mm respectively, so that the extra large lump coal conveyor belt 1d, the large lump coal conveyor belt 1e and the medium lump coal conveyor belt 1f The coals on the coal ash conveyor belt are small lumps, large lumps and medium lumps respectively; the coals on the 1g coal ash conveyor belt are small lumps, granulated coals and pulverized coal.

Described embedded computer 14 can distinguish coal and gangue through the texture features of coal and gangue in the image obtained by camera 7 through the artificial intelligence training model, and can estimate the volume according to the size of coal and gangue in the image;

The embedded computer 14 can adjust the rotation speeds of the extra large lump coal conveyor belt 1d, the large lump coal conveyor belt 1e and the medium lump coal conveyor belt 1f according to the density of coal and gangue blocks in the image obtained by the camera 7 through the artificial intelligence training model.

As shown in the figure, the residual coal re-mining three-dimensional similar simulation experiment method of the present invention comprises the following steps:

a. The coal material is lifted to the top of the large-aperture screen 5a through the low-level conveyor belt 1a, the lifting conveyor belt 1b, and the high-level conveyor belt 1c;

b. According to the size of the coal particle size, use the large-aperture screen 5a, the medium-aperture screen 5b and the small-aperture screen 5c to sieve;

c. According to the size of the coal material particle size, the coal material sieved enters the extra large lump coal conveyor belt 1d, the large lump coal conveyor belt 1e, the medium lump coal conveyor belt 1f and the coal ash conveyor belt 1g through the ceramic chute 6 for subsequent sorting;

d. Judging coal and gangue based on the results of artificial intelligence image recognition texture features and the results of weighing and calculating density: calculate the density of coal and gangue based on the weighing device and estimated volume, and can also be distinguished according to the difference in density The difference between coal and gangue; the result of weighted image recognition texture feature distinction and the result of density distinction are used as the criterion for distinguishing coal and gangue; the embedded computer is manually intervened in its identification results at the initial stage of use, and it is trained while working Model, so that the accuracy of recognition is getting higher and higher;

Let the weight formula be a = k×s + (1-k)×m;

s is the result of image recognition texture features, its value is 0 to 100;

m is the result of weighing and calculating the density, and its value is 0 to 100;

k is the weight of image recognition, ranging from 0 to 1;

a is the criterion for separation of coal and gangue;

When a >=50, the block is considered to be coal;

When a < 50, the block is considered as gangue.

e. Through the controllable rotating baffle plate 8, the gangue is knocked down to the bottom gangue conveyor belt for 1h and transported out;

f. Extra large lump coal, large lump coal and medium lump coal are transported out through extra large lump coal conveyor belt 1d, large lump coal conveyor belt 1e and medium lump coal conveyor belt 1f, and small lump coal, granular coal and pulverized coal are transported through coal ash conveyor belt 1g out.

Claims (7)

1. An integrated device for underground coal and gangue sorting based on artificial intelligence image recognition, characterized in that: it includes an iron shell, a conveying device, a particle size sorting device and a coal and gangue identification and separation device; the conveying device includes a conveyor belt, which is responsible for coal The reprinting of materials in the device; the particle size sorting device includes screens with different apertures for sorting out coal materials of different particle sizes; the coal gangue identification and separation device includes cameras, weighing devices, embedded computers and controllable Rotate the baffle, accurately identify coal and gangue through image recognition and weighing based on artificial intelligence, and then shoot down the gangue through the controllable rotating baffle, so as to achieve the purpose of coal and gangue separation; The conveying device consists of a low conveyor belt (1a), a lifting conveyor belt (1b), a high conveyor belt (1c), an extra large lump coal conveyor belt (1d), a large lump coal conveyor belt (1e), a medium lump coal conveyor belt (1f), and a coal ash conveyor belt. (1g), gangue conveyor belt (1h), iron chute (9) and fixed baffle plate (10); The particle size separation device consists of a shovel plate (3), a V-shaped groove (4), a large-aperture screen (5a), a medium-aperture screen (5b), a small-aperture screen (5c), a ceramic chute (6) and the screen support (11); the shovel plate (3) can slide on the upper surface of the screen; the V-shaped groove (4) is located at the edge of the lower side of the screen (5) ; The large-aperture screen (5a) is located under the high-position conveyor belt (1c), the medium-aperture screen (5b) is located below the large-aperture screen (5a), and the small-aperture screen (5c) is located under the medium-aperture screen (5b) ), and the described screens are obliquely fixed on the screen support (11), and the vibration is controlled by the screen support (11); the large-aperture screen (5a) and a V-shaped groove (4 ) through the ceramic chute (6) is connected with the extra large lump coal conveyor belt (1d), and the medium aperture screen (5b) and the small aperture screen (5c) are respectively connected with the large lump coal conveyor belt (1e) and the medium lump coal conveyor belt (1e) in the same way The conveyor belt (1f) is connected; The gangue sorting and identification device is composed of a camera (7), a controllable rotating baffle (8), a fill light (12), a weighing device (13) and an embedded computer (14); Formula computer (14) links to each other with video camera (7), fill light (12), weighing device (13) and controllable rotating baffle (8) by cable; The low-position conveyor belt (1a) in the conveying device is connected with the lifting conveyor belt (1b) and then connected with the high-position conveyor belt (1c); A camera (7) and a supplementary light (12) are fixed above the coal conveyor belt (1f), and the supplementary light (12) is close to the camera (7); A controllable rotating baffle plate (8) is all arranged on the coal conveyor belt (1e) and the medium lump coal conveyor belt (1f); 1f) has a weighing device (13) under it; The coal ash conveyor belt (1g) is located under the small-diameter screen (5c); the gangue conveyor belt (1h) is located under the iron chute (9), and a fixed The baffle (10); the camera (7) has a self-cleaning function; Except that part of the low-level conveyor belt (1a), coal ash conveyor belt (1g) and gangue conveyor belt (1h) are located outside the iron casing, other parts are located inside the iron casing (2); Judging coal and gangue based on the results of artificial intelligence image recognition texture features and the results of weighing and calculating density: calculate the density of coal and gangue according to the weighing device and estimated volume, and distinguish coal and gangue according to the difference in density difference; the result of weighted image recognition texture feature distinction and the result of density distinction are used as the criterion for distinguishing coal and gangue; the embedded computer is manually intervened in its recognition results at the initial stage of use, and the model is trained while working, so that The accuracy of recognition is getting higher and higher; Let the weight formula be a=k×s+(1-k)×m; s is the result of image recognition texture features, its value is 0 to 100; m is the result of weighing and calculating the density, and its value is 0 to 100; k is the weight of image recognition, ranging from 0 to 1; a is the criterion for separation of coal and gangue; When a>=50, the block is considered to be coal; When a<50, the block is considered as gangue. 2. The integrated underground coal and gangue sorting device based on artificial intelligence image recognition according to claim 1, characterized in that: the lifting conveyor belt (1b) uses a belt with a large inclination angle, so that the device can be used in a small space Raise coal. 3. The underground coal and gangue sorting integrated device based on artificial intelligence image recognition according to claim 1, characterized in that: large-aperture screen (5a), medium-aperture screen (5b), small-aperture screen (5c) ) to preliminarily sort coal materials with different particle sizes to improve the accuracy of image recognition and reduce the amount of follow-up sorting work; and the screens are placed obliquely and equipped with screen supports with vibrating devices, and the matching The shovel plate (3) can remove the coal stuck on the screen to speed up the separation of coal with different particle sizes; The aperture size of described large-aperture screen cloth (5a), medium-aperture screen cloth (5b) and small-aperture screen cloth (5c) is respectively 100mm, 50mm, 25mm, so that extra large lump coal conveyor belt (1d), large lump coal conveyor belt (1e) and the coal on the medium lump coal conveyor belt (1f) are extra large lump coal, large lump coal and medium lump coal; the coal on the coal ash conveyor belt (1g) is small lump coal, granular coal and pulverized coal. 4. The integrated underground coal and gangue sorting device based on artificial intelligence image recognition according to claim 3, characterized in that: the screen is inclined simultaneously to the direction parallel to the rolling of the coal block and the direction perpendicular to the rolling of the coal block, and the two All directions are inclined 10°-20°; the V-shaped groove is placed obliquely with the screen, so that coal blocks larger than the aperture of the screen can flow from the screen to the V-shaped groove and quickly roll off, thereby speeding up the sorting speed. 5. The integrated underground coal and gangue sorting device based on artificial intelligence image recognition according to claim 3, characterized in that: the screen support is composed of iron rods and vibration motors, and the vibration motor controls the vibration of the iron rods so that The screen vibrates. 6. The underground coal and gangue sorting integrated device based on artificial intelligence image recognition according to claim 1, characterized in that: The embedded computer (14) can distinguish coal and gangue through the texture features of coal and gangue in the image obtained by the camera (7) through the artificial intelligence training model, and can estimate the volume according to the size of coal and gangue in the image; The embedded computer (14) can adjust the extra large lump coal conveyor belt (1d), the large lump coal conveyor belt (1e) and the middle lump coal conveyor belt (1d) according to the density of coal and gangue blocks in the image obtained by the camera (7) through the artificial intelligence training model. The speed of the coal conveyor belt (1f). 7. An underground coal and gangue sorting method based on artificial intelligence image recognition, adopting the integrated device for underground coal and gangue sorting based on artificial intelligence image recognition according to any one of claims 1 to 6, characterized in that it comprises the following steps : a. The coal material is lifted to the top of the large-aperture screen (5a) through the low-level conveyor belt (1a), the lifting conveyor belt (1b), and the high-level conveyor belt (1c); b. Use a large-aperture screen (5a), a medium-aperture screen (5b) and a small-aperture screen (5c) to sieve according to the particle size of the coal material; c. The sieved coal according to the particle size of the coal passes through the ceramic chute (6) and enters the extra large lump coal conveyor belt (1d), the large lump coal conveyor belt (1e), the medium lump coal conveyor belt (1f) and the coal ash Conveyor belt (1g) carries out follow-up sorting; d. Judging coal and gangue by weighting the results of texture feature recognition based on artificial intelligence images and the results of weighing and calculating density: Calculate the density of coal and gangue according to the weighing device (13) and the estimated volume, and distinguish the difference between coal and gangue according to the difference in density; Criterion; Described embedded computer (14) is by artificial intervention its recognition result at the early stage of use, while training model while working, makes the accuracy rate of recognition higher and higher; e. Through the controllable rotating baffle plate (8), the gangue is knocked down to the bottom gangue conveyor belt (1h) and transported out; f. Extra large lump coal, large lump coal and medium lump coal are transported out through extra large lump coal conveyor belt (1d), large lump coal conveyor belt (1e) and medium lump coal conveyor belt (1f), and small lump coal, granular coal and pulverized coal Transported by coal ash conveyor belt (1g).

Images