Transmission type intelligent dry separator and identification method

Abstract

The invention relates to the technical field of coal mine separation, in particular to a transmission type intelligent dry separator and an identification method. The transmission type intelligent dry separator comprises a conveying belt, a camera, a linear light source and a recognition system, the conveying belt is a light-transmitting belt and used for conveying to-be-separated minerals, the camera is located on the upper side of the conveying belt, the position, corresponding to the camera, of the conveying belt is a recognition area, and the linear light source is located in the recognition area. The camera is used for shooting an image of the identification area. According to the mineral sorting device, the camera is arranged above the conveying belt, the first linear light source is located below the recognition area and is opposite to the camera, and the conveying belt is a light-transmitting belt, so that light of the first linear light source penetrates through the conveying belt to light minerals to be sorted in the recognition area on the conveying belt; the camera shoots the to-be-sorted minerals in the recognition area, the recognition system receives the images shot by the camera and obtains the light transmittance of each to-be-sorted mineral from the images, and the distinguishing degree is higher.

Description

Technical Field

[0001] The present invention relates to the technical field of coal mine sorting, and in particular to a transmission-type intelligent dry sorting machine and an identification method. Background Art

[0002] In mineral sorting, it is generally necessary to crush the minerals to less than 7mm for sufficient dissociation before sorting. Some mixed ores with translucency have low color differentiation. Generally, ores with higher purity have better translucency. Such ores are concentrates, otherwise they are tailings.

[0003] At present, intelligent dry sorting machines are often used to sort minerals. Intelligent dry sorting machines using image recognition include a feeding system, a material distribution system, a conveyor belt, a recognition system, and an execution system. The recognition system includes a recognition system, a camera, and a linear light source. Since it is sorted by identifying the color and texture of the minerals to be sorted, the color of the minerals to be sorted is identified by using a linear light source to emit light, and the light is irradiated on the minerals, and is realized according to the reflection characteristics of the minerals. Therefore, the camera and the linear light source are both located above the conveyor belt, and the distance from the camera to the conveyor belt is greater than the distance from the linear light source to the conveyor belt. The position on the conveyor belt corresponding to the camera shooting is the recognition area, and the linear light source is located on one side of the recognition area. The mineral image of the recognition area is captured by the camera, and the color and texture of each mineral to be sorted obtained in the image are compared and sorted by the execution system to achieve the classification of the minerals.

[0004] The smaller the size, the more difficult it is to identify the texture, and the closer the colors are, the lower the recognition efficiency and accuracy. Therefore, the recognition method similar to that used in the above-mentioned intelligent dry sorting machine has low differentiation and recognition accuracy when sorting minerals with small color differences and small sizes. Summary of the invention

[0005] (I) The problem to be solved by the present invention is: how to improve the distinction and recognition accuracy of small-sized minerals with similar colors.

[0006] (II) Technical solution

[0007] The present invention provides a transmission-type intelligent dry sorting machine, comprising a conveyor belt, a camera, a linear light source and a recognition system;

[0008] The conveyor belt is a light-transmitting belt, and the conveyor belt is used to convey the minerals to be separated;

[0009] The camera is located on the upper side of the conveyor belt, and the position on the conveyor belt corresponding to the camera is an identification area, and the camera is used to capture an image of the identification area;

[0010] The linear light source comprises at least a first linear light source, and the first linear light source is located below the recognition area and opposite to the camera;

[0011] The recognition system is connected to the camera signal, and is used to receive the image taken by the camera, and obtain the light transmittance of each mineral to be separated in the image according to the image.

[0012] According to an embodiment of the present invention, the conveyor belt is ring-shaped, a receiving space is formed inside the conveyor belt, and the first linear light source is located in the receiving space.

[0013] The linear light source further includes a second linear light source, and the second linear light source is located on an upper side of the identification area.

[0014] According to one embodiment of the present invention, the transmission-type intelligent dry sorting machine further includes a cleaning device, and the cleaning device is used to clean dirt on the bearing surface of the conveyor belt.

[0015] According to one embodiment of the present invention, the transmission-type intelligent dry sorting machine further comprises a frame, and a plurality of the cameras are provided, and the plurality of the cameras are sequentially spaced apart along the width direction of the conveyor belt;

[0016] The linear light source, the conveyor belt and the plurality of cameras are all arranged on the frame;.

[0017] The cleaning device is located at the lower side of the conveyor belt.

[0018] A transmission type intelligent dry sorting machine identification method comprises the following steps:

[0019] S1: Determine the types of concentrates and tailings according to the minerals to be separated, and obtain the average grayscale value Gj of the concentrate image and the average grayscale value Gw of the tailings image;

[0020] S2: Obtain a gray value bj that appears most frequently in the concentrate image and the number of times the gray value appears dj, and obtain a gray value bw that appears most frequently in the tailings image and the number of times the gray value appears dw;

[0021] S3: Establish a classification function y=kb+c for distinguishing concentrates and tailings, where y is the number, k and c are known parameters, and b is the gray value that appears most frequently in a certain mineral image;

[0022] S4: The minerals to be sorted are placed on a conveyor belt for transportation, and images of each mineral to be sorted are collected, and the average gray value Gray of each mineral image to be sorted, the gray value bx that appears most frequently in each mineral image to be sorted, and the number of times the gray value appears dx are obtained;

[0023] S5: compare the average gray value Gray of each to-be-separated mineral image with half the sum of the average gray value Gj of the concentrate image and the average gray value Gw of the tailings image;

[0024] and comparing the number dx corresponding to the gray value that appears most frequently in each image of the mineral to be sorted with the reference number y corresponding to the gray value bx that appears most frequently;

[0025] If the average gray value Gray of the image of the mineral to be separated is greater than (Gj+Gw) / 2, and dx is greater than kx+c, then the mineral to be separated is a concentrate;

[0026] If the average gray value Gray of the image of the mineral to be separated is greater than (Gj+Gw) / 2, and dx is less than kb+c, then the mineral to be separated is a homogeneous medium ore;

[0027] If the average gray value Gray of the image of the mineral to be separated is less than (Gj+Gw) / 2, and dx>kb+c, then the mineral to be separated is an undissociated medium ore;

[0028] If the average gray value of the image of the mineral to be separated is Gray<(Gj+Gw) / 2, and dx<kb+c, then the mineral to be separated is tailings.

[0029] According to one embodiment of the present invention, in the step S1, the following steps are specifically included:

[0030] S11: multiple concentrates of the same type are placed on the conveyor belt, and the camera captures images of all the concentrates in the identification area until all the concentrates are captured, and the position and outline of each concentrate in each image are found, and a small image of only one concentrate of a fixed size is generated based on the ore circumscribed rectangle;

[0031] Put multiple tailings of the same type onto the conveyor belt. The camera takes images of all tailings in the identification area until all tailings are photographed. The camera then finds the location and outline of each tailing in each image. Based on the ore circumscribed rectangle, a small image of only one tailing with a fixed size is generated. The size of the small image generated by the tailings is the same as that of the concentrate.

[0032] S12: Count the grayscale value of each pixel of the concentrate small image in the concentrate data set in step S1, and calculate the average grayscale value of multiple pixels in each small image

[0033] Count the grayscale value of each pixel in each tailings small image in the tailings dataset in step S1, and calculate the average grayscale value of multiple pixels in each small image

[0034] S13: The amount of concentrate is m, represents the average gray value of the mth concentrate, the number of tailings is n, Represents the average gray value of the nth tailings, and counts the average gray value Gj of m concentrates, and the average gray value Gw of n tailings;

[0035]

[0036] According to one embodiment of the present invention, in the step S2, the following steps are specifically included:

[0037] S21: A grayscale histogram is correspondingly established for each small image of the concentrate and tailings, wherein the horizontal axis of the grayscale histogram is the grayscale value b of 0-255, and the vertical axis is the number of occurrences d of the grayscale value corresponding to the pixel;

[0038] S22: According to the grayscale histogram of each concentrate small image, the coordinate points of the maximum peak of each concentrate small image in its corresponding grayscale histogram are counted (bj1, dj1), (bj2, dj2)...(bjm, djm), and according to the grayscale histogram of each tailings small image, the coordinate points of the maximum peak of each tailings small image in its corresponding grayscale histogram are counted (bw1, dw1), (bw2, dw2)...(bwn, dwn).

[0039] According to one embodiment of the present invention, in the step S3, the following steps are specifically included:

[0040] S31: Establish a rectangular coordinate system, where the horizontal axis of the rectangular coordinate system is the gray value b, and the vertical axis is the number of occurrences d of the gray value corresponding to the pixel;

[0041] The coordinate points (bj1, dj1), (bj2, dj2) ... (bjm, djm) of the maximum wave peak in each small image of the concentrate are placed in the rectangular coordinate system, and the coordinate points (bw1, dw1), (bw2, dw2) ... (bwn, dwn) of the maximum wave peak in each small image of the tailings are placed in the rectangular coordinate system;

[0042] S32: Based on the distribution area of ​​the maximum wave crest coordinate points of multiple concentrates and the distribution area of ​​the maximum wave crest coordinate points of multiple tailings in the rectangular coordinate system, classification fitting is performed through the support vector machine machine learning method to determine the parameters k and c in the classification function y=kb+c.

[0043] According to one embodiment of the present invention, the step S1 further includes:

[0044] The camera captures the background image when the conveyor belt is not transporting minerals, and the recognition system obtains the average gray value G of the background image;

[0045] Calculate the average gray value of multiple pixels in each small concentrate image Previously, the background pixels between the bounding rectangle and the actual contour of the particle were first removed, and only the average gray value of the remaining multiple pixels in each small concentrate image was calculated;

[0046] Calculate the average gray value of multiple pixels in each tailings image Before, the background pixels between the bounding rectangle and the actual outline of the particle were first removed, and only the average gray value of the remaining multiple pixels in each tailings small image was calculated.

[0047] According to one embodiment of the present invention, in the step S4, the following steps are specifically included:

[0048] S41: The camera takes images of all the minerals to be separated in the identification area, finds the location and outline of each mineral in each image, and generates a small image of a fixed size with only one mineral to be separated;

[0049] Collect the gray value corresponding to each pixel in each small image of the mineral to be separated, remove the gray value of the pixel that is the same as the average gray value G of the background image, and calculate the average gray value Gray of the remaining multiple pixels in each small image;

[0050] A grayscale histogram is established for each small image of the mineral to be sorted, and the coordinate points (bx1, dx1), (bx2, dx2)...(bxt, dxt) of the maximum peak in the grayscale histogram corresponding to each small image of the mineral to be sorted are obtained, and (bxt, dxt) is the coordinate point of the maximum peak of the tth mineral to be sorted.

[0051] S42: Substitute bx in the coordinates (bx, dx) of the maximum wave peak of each mineral to be sorted into the classification function y=kb+c to determine the type of mineral.

[0052] According to an embodiment of the present invention, in the step S1, the maximum inter-class variance method is used to determine the appropriate segmentation, and the Canny edge detection algorithm is used to determine the position and contour of each concentrate and each tailing.

[0053] According to one embodiment of the present invention, between the steps S21 and S22, the method further includes:

[0054] The grayscale histogram of each concentrate small image and the grayscale histogram of each tailings small image are smoothed, and the coordinate points of the maximum peak in each concentrate small image and the coordinate points of the maximum peak in each tailings small image are counted using a second-order difference recognition algorithm.

[0055] Beneficial effects of the present invention:

[0056] The camera is arranged above the conveyor belt, and the first linear light source is located at the lower side of the identification area and is arranged opposite to the camera up and down. Since the conveyor belt is a light-transmitting belt, the light of the first linear light source can pass through the conveyor belt to illuminate the minerals to be sorted in the identification area on the conveyor belt. The camera photographs the minerals to be sorted in the identification area, and the identification system receives the image taken by the camera and obtains the transmittance of each mineral to be sorted from the image, with higher distinction, which is different from the traditional distinction method by color recognition, and has higher recognition efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0058] Figure 1 A front view schematic diagram of a transmission-type intelligent dry sorting machine provided by an embodiment of the present invention;

[0059] Figure 2 A front view of a removal execution system, a feeding device and a material distribution device of a transmission-type intelligent dry sorting machine provided in an embodiment of the present invention;

[0060] Figure 3 A three-dimensional diagram of a removal execution system, a feeding device and a material distribution device of a transmission-type intelligent dry sorting machine provided in an embodiment of the present invention;

[0061] Figure 4 A cross-sectional view of a removal execution system, a feeding device and a material distribution device of a transmission-type intelligent dry sorting machine provided in an embodiment of the present invention;

[0062] Figure 5 The embodiment of the present invention provides Figure 4 A magnified view of part A;

[0063] Figure 6 The embodiment of the present invention provides Figure 4 A magnified view of part B;

[0064] Figure 7 A schematic diagram of the positional relationship of a camera, a first linear light source, a cleaning device, a sweeping device, and a conveyor belt of a transmission-type intelligent dry sorting machine provided in an embodiment of the present invention;

[0065] Figure 8 This is an identification flow chart of the transmission-type intelligent dry sorting machine provided in an embodiment of the present invention.

[0066] Icons: 1. Conveyor belt; 2. Camera; 3. Linear light source; 4. Cleaning device; 401. Brush; 402. High-pressure air duct; 403. Blowing head; 5. Air nozzle; 6. Second driving assembly; 601. Throwing roller; 602. Active roller; 603. Driven roller; 604. Support roller; 605. Second motor 605; 7. Cleaning device; 701. Water supply pipe; 702. Nozzle; 8. Frame; 9. Feeding device; 10. Fabric device. DETAILED DESCRIPTION

[0067] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] Embodiment 1:

[0069] like Figure 1-Figure 7 As shown, an embodiment of the present invention provides a transmission-type intelligent dry sorting machine, including a conveyor belt 1, a camera 2, a linear light source and a recognition system;

[0070] The conveyor belt 1 is a light-transmitting belt, and is used to convey the minerals to be separated;

[0071] The camera 2 is located on the upper side of the conveyor belt 1. The position on the conveyor belt 1 corresponding to the camera 2 is the identification area. The camera 2 is used to capture images of the identification area.

[0072] The linear light source at least includes a first linear light source 3, and the first linear light source 3 is located below the recognition area and opposite to the camera 2;

[0073] The recognition system is connected to the camera 2 signal, and is used to receive the image taken by the camera 2, and obtain the transmittance of each mineral to be separated in the image according to the image.

[0074] The camera 2 is arranged above the conveyor belt 1, and the first linear light source 3 is located below the identification area and opposite to the camera 2. Since the conveyor belt 1 is a light-transmitting belt, the light of the first linear light source 3 can pass through the conveyor belt 1 to illuminate the minerals to be sorted in the identification area on the conveyor belt 1. The camera 2 photographs the minerals to be sorted in the identification area. The recognition system receives the image taken by the camera 2 and obtains the transmittance of each mineral to be sorted from the image, with higher distinction. Different from the traditional distinction method by color recognition, the recognition efficiency and accuracy are higher.

[0075] The concentrate is transparent due to its high purity, such as silica, and the tailings are white or other colors, such as potassium feldspar. When the tailings are similar in color to the concentrate, potassium feldspar is light white and slightly yellow, while silica is relatively transparent. The linear light source illuminates potassium feldspar and silica through the conveyor belt 1, and the camera 2 captures the corresponding image. The silica will appear light in color with high transmittance, while the opaque mineral particles of potassium feldspar will appear black, and the degree of differentiation is much higher than that of the traditional intelligent dry sorting machine that classifies by color.

[0076] The traditional color sorter recognition is performed after the mineral to be sorted is thrown out of the belt. The posture of the mineral to be sorted is unstable and it is very easy to rotate. During the rotation process, the transmittance of the mineral to be sorted obtained by the color sorter when projected with a thinner surface and a thicker surface is quite different, and the recognition accuracy is low. In this embodiment, since the recognition is performed during the transportation of the conveyor belt 1, each mineral to be sorted has a stable posture and passes smoothly through the recognition area in the conveyor belt 1, so the recognition accuracy is higher.

[0077] Preferably, the conveyor belt 1 is a colorless light-transmitting belt, and optionally, the conveyor belt 1 is a light-colored light-transmitting belt. The conveyor belt 1 is preferably made of polyurethane material for easy cleaning.

[0078] According to an embodiment of the present invention, the conveyor belt 1 is annular, and a receiving space is formed inside the conveyor belt 1, and the first linear light source 3 is located in the receiving space. The first linear light source 3 is located in the receiving space, which can make the lighting effect of the first linear light source 3 better.

[0079] According to an embodiment of the present invention, the linear light source further includes a second linear light source, and the second linear light source is located on the upper side of the recognition area.

[0080] Preferably, Figure 4 As shown, along the length direction of the conveyor belt 1, at least two cameras 2 are provided, and the first linear light source 3 corresponds to one of the cameras 2, so that the recognition system can obtain the transmittance of the mineral to be sorted, and the second linear light source corresponds to the other camera 2, so that the recognition system can obtain the color of the mineral to be sorted. By obtaining the color and transmittance of the mineral to be sorted, the distinction is higher and the recognition accuracy is better.

[0081] Optionally, the first linear light source 3, the second linear light source and the camera 2 correspond one to one and can be turned on at the same time, so that the recognition system can simultaneously obtain the color and transmittance of the mineral to be sorted in the image taken by the camera 2; or, according to the actual sorted minerals, only the first linear light source 3 or the second linear light source is turned on, and the recognition system obtains the transmittance or color of the mineral to be sorted in the image taken by the camera 2.

[0082] According to one embodiment of the present invention, Figure 4 and Figure 5As shown, the transmission type intelligent dry separator further includes a cleaning device 4 , which is used to clean dirt on the bearing surface of the conveyor belt 1 .

[0083] By means of the cleaning device 4, the bearing surface of the conveyor belt 1 can be cleaned to prevent the conveyor belt 1 from being too dirty and affecting the actual light transmittance of the minerals to be separated.

[0084] According to one embodiment of the present invention, the cleaning device 4 is located at the lower side of the conveyor belt 1 .

[0085] Of course, in this embodiment, the cleaning device 4 can also be located at other positions, for example, the cleaning device 4 is located on one side of the feeding end of the conveyor belt 1, as long as it can play a cleaning role without affecting the transportation and sorting of minerals.

[0086] According to one embodiment of the present invention, the transmission type intelligent dry sorting machine further comprises a frame 8, a plurality of cameras 2 are provided, and the plurality of cameras 2 are sequentially arranged at intervals along the width direction of the conveyor belt 1;

[0087] The linear light source, the conveyor belt 1 and the plurality of cameras 2 are all arranged on a frame 8 .

[0088] The frame 8 includes a base frame, a first connecting frame and a second connecting frame. The conveyor belt 1 is located on the base frame, the camera 2 is fixed on the first connecting frame, and the linear light source is fixed on the second connecting frame.

[0089] It should be noted that in order to transport and sort more minerals to be sorted and improve work efficiency, the width of the conveyor belt 1 is relatively wide. A plurality of cameras 2 are arranged in sequence and at intervals along the width direction of the conveyor belt 1, so that the clarity of the captured image is not affected and the minerals to be sorted that pass through the identification area of ​​the conveyor belt 1 can be photographed each time, without missing any minerals, thereby improving the sorting effect.

[0090] The frame 8 is provided with a second driving assembly 6, which includes a second motor 605, a throwing roller 601, a driving roller 602, a driven roller 603 and a supporting roller 604;

[0091] The active roller 602, the throwing roller 601 and the driven roller 603 are all in contact with the supporting surface of the conveyor belt 1, and are used to support the conveyor belt 1. The supporting roller 604 is located on the lower side of the conveyor belt 1, and the side of the supporting roller 604 is in contact with the bearing surface of the conveyor belt 1. The active roller 602, the throwing roller 601 and the driven roller 603 are all connected by the conveyor belt 1. The distance from the driven roller 603 to the ground is smaller than the distance from the active roller 602 to the ground, so that the accommodating space can have enough space to accommodate the first linear light source 3.

[0092] The driven roller 603 is installed on the frame 8 through a mounting frame, and a plurality of first mounting holes are provided on the mounting frame, and a plurality of second mounting holes are provided on the frame 8. The locking bolt passes through one of the first mounting holes and one of the second mounting holes in sequence and is screwed with the locking nut, so that the position of the driven roller 603 can be adjusted, the tightness of the conveyor belt 1 can be adjusted, and the size of the accommodation space can be adjusted at the same time, so as to facilitate the placement of the first linear light source 3

[0093] Since the smaller the diameter of the throwing roller 601 is, the greater the centrifugal force is, therefore, the diameter of the throwing roller 601 is not greater than 80 mm, so that the minerals to be separated can be thrown out more smoothly.

[0094] By setting up multiple supporting rollers 604, the conveyor belt 1 from the nozzle 702 to the active roller 602 is inclined relative to the horizontal plane, and one end of the conveyor belt 1 close to the active roller 602 is lower than the other end. The water sprayed by the nozzle 702 is not easy to flow to the position of the brush 401, reducing the probability of interference with the cleaning of the brush 401.

[0095] The output end of the second motor 605 is fixedly connected to the active roller 602, and the fixed end of the second motor 605 is fixedly connected to the frame 8. The second motor 605 drives the active roller 602 fixedly connected to its output end to rotate, thereby enabling the conveyor belt 1 to move and convey the minerals to be sorted.

[0096] According to one embodiment of the present invention, Figure 4 As shown, the cleaning device 4 includes a brush 401 and a first driving assembly;

[0097] The brush 401 contacts the bearing surface of the conveyor belt 1 , and the first driving assembly is used to drive the brush 401 to rotate along its axial direction.

[0098] The first driving assembly includes a first motor, a first connecting shaft and two first bearing seats. The fixed end of the first motor is fixedly connected to the frame 8. The output end of the first motor is fixedly connected to the first connecting shaft. The brush 401 is fixedly arranged on the first connecting shaft. The first motor rotates, driving the first connecting shaft to which its output end is fixedly connected to rotate, thereby driving the brush 401 to rotate, and cleaning the bearing surface of the conveyor belt 1.

[0099] The provision of the first bearing seat can improve the stability of the brush 401.

[0100] Preferably, the brush 401 is cylindrical, the brush 401 and the first connecting shaft are coaxially arranged, the length direction of the brush 401 is the same as the width direction of the conveyor belt 1, and the length of the brush 401 is not less than the width of the conveyor belt 1, so that the brush 401 can clean all positions of the conveyor belt 1, and the cleaning effect is better.

[0101] According to one embodiment of the present invention, the cleaning device 4 includes a high-pressure air duct 402 and a blowing head 403 that are connected;

[0102] The air outlet of the blowing head 403 is arranged toward the bearing surface of the conveyor belt 1 .

[0103] Furthermore, a plurality of blowing heads 403 are provided, and are arranged in sequence and at intervals along the length direction of the high-pressure air duct 402 , and the blowing areas of the plurality of blowing heads 403 can completely cover the bandwidth of the conveyor belt 1 .

[0104] The high-pressure air duct 402 is externally connected to an air supply system for supplying air to the high-pressure air duct 402 .

[0105] After the brush 401 cleans, the high-pressure air duct 402 delivers high-pressure air to the blowing head 403 through the blowing head 403 and blows it toward the bearing surface of the conveyor belt 1, thereby cleaning the remaining dirt on the conveyor belt 1, making the bearing surface of the conveyor belt 1 cleaner and not affecting the subsequent sorting and identification of the minerals to be sorted.

[0106] It should be noted that the rotation direction of the brush 401 is opposite to the conveying direction of the conveyor belt 1, so that the cleaning effect is better.

[0107] According to one embodiment of the present invention, the transmission type intelligent dry sorting machine further includes a cleaning device 7, and the cleaning device 7 includes a water supply pipe 701 and a nozzle 702 connected to each other;

[0108] The water spraying port of the nozzle 702 is arranged toward the bearing surface of the conveyor belt 1. The water supply pipe 701 is used to transport external water source to the nozzle 702.

[0109] It should be noted that the nozzle 702 is arranged between the brush 401 and the blowing head 403 .

[0110] There are multiple nozzles 702, and the multiple nozzles 702 are arranged in sequence and spaced apart along the length direction of the water supply pipe 701. The length direction of the water supply pipe 701 is the same as the width direction of the conveyor belt 1, and the water supply pipe 701 is arranged in sequence and spaced apart in the length direction, and the spraying areas of each two adjacent water supply pipes 701 to the bearing surface of the conveyor belt 1 partially overlap, so that the conveyor belt 1 can be cleaned, and the cleaning effect is better.

[0111] The transmission type intelligent sorting machine also includes a feeding device 9, a material distribution device 10 and an execution system. The material distribution device 10 is located between the feeding device 9 and the conveyor belt 1. The feeding device 9 is used for feeding materials, and the material distribution device 10 is used for distributing materials on the conveyor belt 1. The feeding device 9 and the material distribution device 10 are both existing technologies and will not be described in detail here.

[0112] The execution system includes a plurality of air nozzles 5, which are arranged in sequence and spaced apart along the width direction of the conveyor belt 1, and the spraying direction of the plurality of air nozzles 5 is perpendicular to the horizontal plane. The air supply system is connected to the air nozzles 5 to provide the air nozzles 5 with high-pressure air.

[0113] The distance between the projection of the air nozzle 5 on the horizontal plane and the projection of the throwing roller 601 on the horizontal plane is the first distance, which is between 100-200 mm, preferably 150 mm. The minerals to be sorted are thrown under the conveyor belt 1, and the concentrate is thrown normally. The air nozzle 5 in the execution system sprays the tailings according to the sorting results. In order to prevent the tailings from hitting the conveyor belt 1, the first distance is not less than 100 mm. And because the farther the minerals to be sorted are thrown, the greater the divergence distance, in order to improve the spraying accuracy of the air nozzle 5, the first distance is not greater than 200 mm.

[0114] Of course, the ejected minerals can also be concentrates, and the tailings can be thrown out normally.

[0115] In this embodiment, a plurality of partitions are fixedly connected to the conveyor belt 1 along the length direction of the conveyor belt 1, dividing the conveyor belt 1 into four equal parts, each part corresponds to a feeding device 9 and a distribution device 10, and each feeding device 9 and distribution device 10 can be started individually, thereby improving production efficiency. Even if one group of the feeding devices 9 and the distribution devices 10 fails, it will not affect the operation of the feeding devices 9 and the distribution devices 10 at other positions.

[0116] Specifically, the minerals to be sorted are evenly spread from the distribution device 10 onto the conveyor belt 1 for transportation, the linear light source is turned on to illuminate the identification area, the camera 2 is turned on, and when the minerals to be sorted pass through the identification area on the conveyor belt 1, the camera 2 photographs the minerals to be sorted in the identification area, and the identification system obtains the transmittance of each mineral to be sorted in the image based on the photographed image. At this time, the brush 401 on the lower side of the conveyor belt 1 rotates driven by the first driving component to clean the bearing surface of the conveyor belt 1, the water supply pipe 701 supplies water at regular intervals, and sprays water from the nozzle 702 to clean the conveyor belt 1, and the high-pressure air duct 402 at the bottom of the conveyor belt 1 continuously supplies air, and blows out from the blowing head 403 to further clean the bearing surface of the conveyor belt 1 without affecting the lighting of the linear light source.

[0117] Embodiment 2:

[0118] like Figure 8 As shown, a transmission type intelligent dry sorting machine identification method includes the following steps:

[0119] S1: Determine the types of concentrates and tailings according to the minerals to be separated, and obtain the average grayscale value Gj of the concentrate image and the average grayscale value Gw of the tailings image;

[0120] According to one embodiment of the present invention, step S1 specifically includes the following steps:

[0121] S11: multiple concentrates of the same type are placed on the conveyor belt 1, and the camera 2 captures images of all the concentrates in the identification area until all the concentrates are captured, and the position and outline of each concentrate in each image are found, and a small image of only one concentrate of a fixed size is generated based on the ore circumscribed rectangle;

[0122] Put multiple tailings of the same type onto the conveyor belt 1. Camera 2 takes images of all tailings in the identification area until all tailings are photographed, and finds the location and outline of each tailing in each image. Based on the circumscribed rectangle of the ore, a small image of only one tailing with a fixed size is generated. The size of the small image generated by the tailings is the same as that of the small image generated by the concentrate.

[0123] According to an embodiment of the present invention, in the step S1, the maximum inter-class variance method is used to determine the appropriate segmentation, and the Canny edge detection algorithm is used to determine the position and contour of each concentrate and each tailing.

[0124] S12: Count the grayscale value of each pixel of the concentrate small image in the concentrate data set in step S1, and calculate the average grayscale value of multiple pixels in each small image

[0125] Count the grayscale value of each pixel in each tailings small image in the tailings dataset in step S1, and calculate the average grayscale value of multiple pixels in each small image

[0126] S13: The amount of concentrate is m, represents the average gray value of the mth concentrate, the number of tailings is n, Represents the average gray value of the nth tailings, and counts the average gray value Gj of m concentrates, and the average gray value Gw of n tailings;

[0127]

[0128] According to one embodiment of the present invention, the step S1 further includes:

[0129] Camera 2 captures the background image when the conveyor belt is not transporting minerals, and the recognition system obtains the average gray value G of the background image;

[0130] Calculate the average gray value of multiple pixels in each small concentrate image Previously, the background pixels between the bounding rectangle and the actual contour of the particle were first removed, and only the average gray value of the remaining multiple pixels in each small concentrate image was calculated;

[0131] Calculate the average gray value of multiple pixels in each tailings image Before, the background pixels between the bounding rectangle and the actual outline of the particle were first removed, and only the average gray value of the remaining multiple pixels in each tailings small image was calculated.

[0132] It should be noted that the concentrates and tailings are manually sorted and confirmed according to the types of the minerals to be separated and placed in the feeding device 9 , and transported to the conveyor belt 1 through the distribution device 10 .

[0133] m and n are values ​​not less than 1. The larger the values ​​of m and n are, the more accurate the average gray value of the concentrate and the average gray value Gj of the tailings are.

[0134] S2: Obtain a gray value bj that appears most frequently in the concentrate image and the number of times the gray value appears dj, and obtain a gray value bw that appears most frequently in the tailings image and the number of times the gray value appears dw;

[0135] According to one embodiment of the present invention, in step S2, the following steps are specifically included:

[0136] S21: A grayscale histogram is correspondingly established for each small image of the concentrate and tailings, wherein the horizontal axis of the grayscale histogram is the grayscale value b of 0-255, and the vertical axis is the number of occurrences d of the grayscale value corresponding to the pixel;

[0137] According to one embodiment of the present invention, before step S22, the grayscale histogram of each concentrate small image and the grayscale histogram of each tailings small image are smoothed, and the coordinate points of the maximum peak in each concentrate small image and the coordinate points of the maximum peak in each tailings small image are counted using a second-order difference recognition algorithm.

[0138] S22: According to the grayscale histogram of each concentrate small image, the coordinate points of the maximum peak of each concentrate small image in its corresponding grayscale histogram are counted (bj1, dj1), (bj2, dj2)...(bjm, djm), and according to the grayscale histogram of each tailings small image, the coordinate points of the maximum peak of each tailings small image in its corresponding grayscale histogram are counted (bw1, dw1), (bw2, dw2)...(bwn, dwn).

[0139] S3: Establish a classification function y=kb+c for distinguishing concentrates and tailings, where y is the number, k and c are known parameters, and b is the gray value that appears most frequently in a certain mineral image;

[0140] According to one embodiment of the present invention, in step S3, the following steps are specifically included:

[0141] S31: Establish a rectangular coordinate system, where the horizontal axis of the rectangular coordinate system is the gray value b, and the vertical axis is the number of occurrences d of the gray value corresponding to the pixel;

[0142] The coordinate points (bj1, dj1), (bj2, dj2) ... (bjm, djm) of the maximum wave peak in each small image of the concentrate are placed in the rectangular coordinate system, and the coordinate points (bw1, dw1), (bw2, dw2) ... (bwn, dwn) of the maximum wave peak in each small image of the tailings are placed in the rectangular coordinate system;

[0143] S32: Based on the distribution area of ​​the maximum wave crest coordinate points of multiple concentrates and the distribution area of ​​the maximum wave crest coordinate points of multiple tailings in the rectangular coordinate system, classification fitting is performed through the support vector machine machine learning method to determine the parameters k and c in the classification function y=kb+c.

[0144] S4: Put the minerals to be sorted on the conveyor belt 1 for transportation, collect the image of each mineral to be sorted, and obtain the average gray value Gray of each mineral image to be sorted, the gray value bx that appears most frequently in each mineral image to be sorted, and the number of times the gray value appears dx;

[0145] S5: compare the average gray value Gray of each to-be-separated mineral image with half the sum of the average gray value Gj of the concentrate image and the average gray value Gw of the tailings image;

[0146] and comparing the number dx corresponding to the gray value that appears most frequently in each image of the mineral to be sorted with the reference number y corresponding to the gray value bx that appears most frequently;

[0147] According to one embodiment of the present invention, the step S4 includes the following specific steps:

[0148] S41: Camera 2 captures images of all the minerals to be separated in the identification area, finds the location and outline of each mineral in each image, and generates a small image of a fixed size with only one mineral to be separated;

[0149] Collect the gray value corresponding to each pixel in each small image of the mineral to be separated, remove the gray value of the pixel that is the same as the average gray value G of the background image, and calculate the average gray value Gray of the remaining multiple pixels in each small image;

[0150] A grayscale histogram is established for each small image of the mineral to be sorted, and the coordinate points (bx1, dx1), (bx2, dx2)...(bxt, dxt) of the maximum peak in the grayscale histogram corresponding to each small image of the mineral to be sorted are obtained, and (bxt, dxt) is the coordinate point of the maximum peak of the tth mineral to be sorted.

[0151] If the average gray value Gray of the image of the mineral to be separated is greater than (Gj+Gw) / 2, and dx is greater than kx+c, then the mineral to be separated is a concentrate;

[0152] If the average gray value Gray of the image of the mineral to be separated is greater than (Gj+Gw) / 2, and dx is less than kb+c, then the mineral to be separated is a homogeneous medium ore;

[0153] If the average gray value Gray of the image of the mineral to be separated is less than (Gj+Gw) / 2, and dx>kb+c, then the mineral to be separated is an undissociated medium ore;

[0154] If the average gray value of the image of the mineral to be separated is Gray<(Gj+Gw) / 2, and dx<kb+c, then the mineral to be separated is tailings.

[0155] When the type of mineral is determined to be an undissociated mineral, it is taken out and crushed into smaller pieces of minerals, and the crushed smaller pieces of minerals are placed in a feeding device, and transported to the conveyor belt 1 through a distribution device, and steps 4 and 5 are repeated until each small piece of mineral is identified as one of concentrate, tailings and homogeneous medium ore.

[0156] It should be noted that Gray>(Gj+Gw) / 2, indicating that the overall light transmittance of the mineral is higher than the average level;

[0157] dx>kbx+c means that the local transmittance of the mineral is higher than the average level. Therefore, when the transmittance of the entire mineral is lower than the average level, but the local transmittance is higher than the average level, it means that the ore may be partly concentrate and partly tailings, and can be further crushed and dissociated.

[0158] Whether it is a concentrate, tailings or a small image of a mineral to be separated, it is preferred to use 224*224 pixels. Further, an example is given to illustrate the average gray value of a small image of a concentrate. Calculation method of

[0159] Determine the grayscale value of each pixel, recorded as Gj1, Gj2, ...Gj 50176 , remove the pixels whose pixel positions are between the particle circumscribed rectangle and the actual contour, such as Gj 200 -Gj 299 , will soon 200 -Gj 299 After removing all the pixels, 50166 pixels remain. Add the pixel values ​​of the remaining 50166 pixels and divide them by 50166, which represents the average grayscale value of the pixels in the small concentrate image. Then the average gray value of all concentrate images is obtained.

[0160] Similarly, the average gray value Gw of the tailings and the average gray value Gray of the minerals to be separated can be obtained.

[0161] Furthermore, before establishing a corresponding grayscale histogram for each small image, it is also necessary to remove pixels with the same grayscale value as the background grayscale value G, which can improve the recognition accuracy.

[0162] It should be noted that the transmittance is proportional to the grayscale value, and the greater the transmittance, the greater the grayscale value.

[0163] In the description of the present invention, it should be noted that the terms "upper", "lower", etc., indicating the orientation or positional relationship are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0164] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be a connection between the inside of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. In addition, in the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0165] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A transmission type intelligent dry sorting machine, characterized in that: It comprises a conveyor belt (1), a camera (2), a linear light source and a recognition system; The conveyor belt (1) is a light-transmitting belt, and the conveyor belt (1) is used to convey the minerals to be separated; The camera (2) is located on the upper side of the conveyor belt (1); the position on the conveyor belt (1) corresponding to the camera (2) is an identification area; and the camera (2) is used to capture an image of the identification area; The linear light source comprises at least a first linear light source (3), and the first linear light source (3) is located below the recognition area and opposite to the camera (2); The recognition system is connected to the camera (2) by signal, and is used to receive the image taken by the camera (2), and obtain the light transmittance of each mineral to be separated in the image based on the image.

2. A transmission type intelligent dry separator according to claim 1, characterized in that: The conveyor belt (1) is ring-shaped, a receiving space is formed inside the conveyor belt (1), and the first linear light source (3) is located in the receiving space.

3. A transmission type intelligent dry separator according to claim 1, characterized in that: The linear light source further includes a second linear light source, and the second linear light source is located on an upper side of the identification area.

4. A transmission type intelligent dry separator according to claim 1 or 2, characterized in that: The transmission-type intelligent dry sorting machine further comprises a cleaning device (4), and the cleaning device (4) is used to clean dirt on the bearing surface of the conveyor belt (1).

5. A transmission type intelligent dry separator according to claim 4, characterized in that: The transmission-type intelligent dry sorting machine also includes a frame (8), a plurality of the cameras (2) are provided, and the plurality of the cameras (2) are sequentially spaced apart along the width direction of the conveyor belt (1); The linear light source, the conveyor belt (1) and the plurality of cameras (2) are all arranged on the frame (8); The cleaning device (4) is located at the lower side of the conveyor belt (1).

6. A transmission type intelligent dry sorting machine identification method, characterized in that: The following steps are involved: S1: Determine the types of concentrates and tailings according to the minerals to be separated, and obtain the average grayscale value Gj of the concentrate image and the average grayscale value Gw of the tailings image; S2: Obtain a gray value bj that appears most frequently in the concentrate image and the number of times the gray value appears dj, and obtain a gray value bw that appears most frequently in the tailings image and the number of times the gray value appears dw; S3: Establish a classification function y=kb+c for distinguishing concentrates and tailings, where y is the number, k and c are known parameters, and b is the gray value that appears most frequently in a certain mineral image; S4: the minerals to be separated are placed on the conveyor belt (1) for transportation, and the images of each mineral to be separated are collected, and the average gray value Gray of each mineral image to be separated and the gray value bx that appears most frequently in each mineral image to be separated and the number of times the gray value appears dx are obtained; S5: compare the average gray value Gray of each to-be-separated mineral image with half the sum of the average gray value Gj of the concentrate image and the average gray value Gw of the tailings image; and comparing the number dx corresponding to the gray value that appears most frequently in each image of the mineral to be sorted with the reference number y corresponding to the gray value bx that appears most frequently; If the average gray value Gray of the image of the mineral to be separated is greater than (Gj+Gw) / 2, and dx is greater than kx+c, then the mineral to be separated is a concentrate; If the average gray value Gray of the image of the mineral to be separated is greater than (Gj+Gw) / 2, and dx is less than kb+c, then the mineral to be separated is a homogeneous medium ore; If the average gray value Gray of the image of the mineral to be separated is less than (Gj+Gw) / 2, and dx>kb+c, then the mineral to be separated is an undissociated medium ore; If the average gray value of the image of the mineral to be separated is Gray<(Gj+Gw) / 2, and dx<kb+c, then the mineral to be separated is tailings.

7. A transmission type intelligent dry sorting machine identification method according to claim 6, characterized in that: In the step S1, the following steps are specifically included: S11: multiple concentrates of the same type are placed on the conveyor belt (1), and the camera (2) captures images of all the concentrates in the identification area until all the concentrates are captured, and the position and outline of each concentrate in each image are found, and a small image of only one concentrate of a fixed size is generated based on the ore circumscribed rectangle; A plurality of tailings of the same type are placed on a conveyor belt (1), and a camera (2) takes images of all tailings in an identification area until all tailings are photographed, and finds the location and outline of each tailing in each image, and generates a small image of only one tailing of a fixed size based on the ore circumscribed rectangle, wherein the size of the small image generated by the tailings is the same as the size of the small image generated by the concentrate; S12: Count the grayscale value of each pixel of the concentrate small image in the concentrate data set in step S1, and calculate the average grayscale value of multiple pixels in each small image Count the grayscale value of each pixel in each tailings small image in the tailings dataset in step S1, and calculate the average grayscale value of multiple pixels in each small image S13: The amount of concentrate is m, represents the average gray value of the mth concentrate, the number of tailings is n, Represents the average gray value of the nth tailings, and counts the average gray value Gj of m concentrates, and the average gray value Gw of n tailings; 8. A transmission type intelligent dry sorting machine identification method according to claim 7, characterized in that: In the step S2, the following steps are specifically included: S21: A grayscale histogram is correspondingly established for each small image of the concentrate and tailings, wherein the horizontal axis of the grayscale histogram is the grayscale value b of 0-255, and the vertical axis is the number of occurrences d of the grayscale value corresponding to the pixel; S22: According to the grayscale histogram of each concentrate small image, the coordinate points of the maximum peak of each concentrate small image in its corresponding grayscale histogram are counted (bj1, dj1), (bj2, dj2)...(bjm, djm), and according to the grayscale histogram of each tailings small image, the coordinate points of the maximum peak of each tailings small image in its corresponding grayscale histogram are counted (bw1, dw1), (bw2, dw2)...(bwn, dwn).

9. A transmission type intelligent dry sorting machine identification method according to claim 8, characterized in that: In the step S3, the following steps are specifically included: S31: Establish a rectangular coordinate system, where the horizontal axis of the rectangular coordinate system is the gray value b, and the vertical axis is the number of occurrences d of the gray value corresponding to the pixel; The coordinate points (bj1, dj1), (bj2, dj2) ... (bjm, djm) of the maximum wave peak in each small image of the concentrate are placed in the rectangular coordinate system, and the coordinate points (bw1, dw1), (bw2, dw2) ... (bwn, dwn) of the maximum wave peak in each small image of the tailings are placed in the rectangular coordinate system; S32: Based on the distribution area of ​​the maximum wave crest coordinate points of multiple concentrates and the distribution area of ​​the maximum wave crest coordinate points of multiple tailings in the rectangular coordinate system, classification fitting is performed through the support vector machine machine learning method to determine the parameters k and c in the classification function y=kb+c.

10. A transmission type intelligent dry sorting machine identification method according to claim 9, characterized in that: The step S1 also includes: The camera (2) captures a background image of the conveyor belt (1) when it is not transporting minerals, and the recognition system obtains an average grayscale value G of the background image; Calculate the average gray value of multiple pixels in each small concentrate image Previously, the background pixels between the bounding rectangle and the actual contour of the particle were first removed, and only the average gray value of the remaining multiple pixels in each small concentrate image was calculated; Calculate the average gray value of multiple pixels in each tailings image Before, the background pixels between the bounding rectangle and the actual outline of the particle were first removed, and only the average gray value of the remaining multiple pixels in each tailings small image was calculated.

11. A transmission type intelligent dry sorting machine identification method according to claim 10, characterized in that: In the step S4, the following steps are specifically included: S41: The camera (2) takes images of all the minerals to be separated in the identification area, finds the location and outline of each mineral in each image, and generates a small image of a fixed size with only one mineral to be separated; Collect the gray value corresponding to each pixel in each small image of the mineral to be separated, remove the gray value of the pixel that is the same as the average gray value G of the background image, and calculate the average gray value Gray of the remaining multiple pixels in each small image; A grayscale histogram is established for each small image of the mineral to be sorted, and the coordinate points (bx1, dx1), (bx2, dx2)...(bxt, dxt) of the maximum peak in the grayscale histogram corresponding to each small image of the mineral to be sorted are obtained, and (bxt, dxt) is the coordinate point of the maximum peak of the tth mineral to be sorted.

12. A transmission type intelligent dry sorting machine identification method according to claim 7, characterized in that: In the step S1, the maximum inter-class variance method is used to determine the appropriate segmentation, and the Canny edge detection algorithm is used to determine the location and contour of each concentrate and each tailing.

13. The identification method of a transmission type intelligent dry sorting machine according to claim 8, characterized in that: Between the steps S21 and S22, the method further includes: The grayscale histogram of each concentrate small image and the grayscale histogram of each tailings small image are smoothed, and the coordinate points of the maximum peak in each concentrate small image and the coordinate points of the maximum peak in each tailings small image are counted using a second-order difference recognition algorithm.

Images