A three-dimensional information recognition-assisted ore sorting decision method, system and medium

Abstract

The present invention discloses a three-dimensional information recognition-assisted ore sorting decision method, system and medium. The present invention includes obtaining three-dimensional point cloud data and X-ray transmission grayscale map of ore; projecting the three-dimensional point cloud data to the grid unit on the two-dimensional plane where the ore conveyor belt is located to calculate the ore thickness map; spatially aligning the ore thickness map and the X-ray transmission grayscale map to achieve mutual correlation between the grayscale value and thickness value of each pixel in the X-ray transmission grayscale map; dividing the pixel grayscale value of the pixel point in the X-ray transmission grayscale map by the associated thickness value to obtain the average mass absorption grayscale map of the ore and extracting the edge to obtain the pixel point set of a single ore; calculating the average grayscale value of the pixel point set of a single ore to determine whether it is waste rock. The present invention aims to decouple the ore thickness information contained in the X-ray transmission map. The entire method does not require the preparation of standard samples, and solves the problem that the ore thickness affects the grayscale value of the X-ray transmission map.

Description

Technical Field

[0001] The present invention relates to the technical field of ore sorting based on computer vision, and in particular to a three-dimensional information recognition-assisted ore sorting decision method, system and medium. Background Art

[0002] In the case of different grades of non-ferrous metal ores, it is necessary to separate the ore and waste rock through pre-waste treatment, so as to reduce the amount of ore to be processed later and improve the processing capacity and overall economic benefits of mineral processing equipment. X-ray transmission (XRT) is an effective pre-waste treatment technology currently widely used. Without destroying the object, high-energy and low-energy X-rays pass through the same part. The linear attenuation coefficients at two different energy levels are compared to obtain a characteristic value related to the type of material, that is, the dual-energy R value. Then, the grayscale image of the dual-energy R value can be used for ore identification and positioning using image processing methods. However, in the actual application of ore pre-waste treatment, since the X-ray tube produces a continuous spectrum rather than an ideal single spectrum, the dual-energy R value of the ore is not only related to its average atomic number, but also affected by the thickness of the ore. When the thickness of the ore changes greatly, it is easy to make wrong decisions on waste treatment. In order to solve the above problems, the existing technology introduces the concept of average mass absorption coefficient of multiple equivalent substrates, and solves the correction coefficient group about mass thickness after measuring the standard sample, so that the algorithm can meet the problem of sorting irregular ore and waste rock and correct the influence of mass thickness. However, this type of existing technology relies on the production of a large number of standard samples, which increases various labor and sample preparation costs. Summary of the invention

[0003] Technical problem to be solved by the present invention: In view of the above-mentioned problems of the prior art, a method, system and medium for ore sorting decision-making assisted by three-dimensional information recognition are provided. The present invention aims to decouple the ore thickness information contained in the X-ray transmission image. The entire method does not require the preparation of standard samples, thereby solving the problem that the ore thickness affects the grayscale value of the X-ray transmission image.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0005] A three-dimensional information recognition-assisted ore sorting decision method comprises the following steps:

[0006] S1, obtaining the three-dimensional point cloud data and X-ray transmission grayscale image of the ore;

[0007] S2, projecting the three-dimensional point cloud data into the grid cells on the two-dimensional plane where the ore conveyor belt is located, calculating the thickness value of each grid cell and using it as the thickness value of the pixel point, thereby obtaining the ore thickness map;

[0008] S3, aligning the space of the ore thickness map and the X-ray transmission grayscale map to achieve the association between the grayscale value of each pixel in the X-ray transmission grayscale map and the thickness value;

[0009] S4, decoupling the pixel grayscale values ​​and the associated thickness values ​​of the pixel points in the X-ray transmission grayscale image using a curve fitting formula, to obtain an average mass absorption grayscale image consisting of average mass absorption coefficients of the reaction material types of the ore;

[0010] S5, performing edge extraction on the average mass absorption grayscale image of the ore to obtain a pixel point set of a single ore;

[0011] S6, calculate the average gray value of a single ore pixel set, and compare the average gray value with a preset threshold. If the average gray value is greater than the preset threshold, the ore is determined to be valid ore, otherwise the ore is determined to be waste rock.

[0012] Optionally, step S2 includes:

[0013] S2.1, divide the xy two-dimensional plane where the ore conveyor belt is located into The size of each grid cell is ,in and are the width and height of the grid cell respectively, and any index is The coordinate range of the grid cells is:

[0014] ,

[0015] ,

[0016] In the above formula, are the coordinates in the grid cell, and Respectively represent ore point cloud data All coordinate points are Axis and The minimum value on the axis;

[0017] S2.2, for each point in the 3D point cloud data of the ore , find the index of the grid cell it falls on according to the following formula :

[0018] , ;

[0019] S2.3, for any index respectively The thickness of the grid cell is calculated according to the following formula:

[0020] ,

[0021] In the above formula, The index is The thickness of the grid cell, For the index The points in the grid cell The number of For the Points Axis coordinates;

[0022] S2.4, taking the grid unit as the pixel point and the thickness value of the grid unit as the pixel value of the pixel point, the size is obtained Ore thickness map .

[0023] Optionally, step S3 includes:

[0024] S3.1, the ore thickness map and X-ray transmission grayscale map are respectively extracted using the Harris corner point detection method to obtain the Harris response map :

[0025] ,

[0026] In the above formula, Harris response graph Central coordinates The pixel value of is the image gradient matrix, is the image gradient matrix The determinant of is the image gradient matrix , and there are:

[0027] , , ,

[0028] In the above formula, and X-ray transmission grayscale images exist and The gradient of the direction, is an empirical constant; if the pixel value If it is greater than the preset threshold, the coordinates are determined The pixel point is the key feature point;

[0029] S3.2, construct the scale space of each key feature point according to the following formula:

[0030] ,

[0031] In the above formula, is the scale space of key feature points, is the pixel value of the key feature point, are the coordinates of the key feature points, is the scale factor; the gradient size and direction in the neighborhood of each key feature point are calculated according to the scale space of the key feature point, and a direction histogram containing multiple intervals to cover 360 degrees is created, and the highest peak in the direction histogram is selected as the main direction of the key feature point;

[0032] S3.3, extract a pixel area of ​​a specified size in the main direction of the key feature point and rotate it according to the main direction, divide the pixel area of ​​the specified size into multiple grid-like small blocks, calculate the gradient histograms of the eight directions around each small block, and splice all the gradient histograms to generate a 128-dimensional feature point descriptor, thereby obtaining the feature point descriptors of the ore thickness map and the X-ray transmission grayscale image respectively;

[0033] S3.4, taking the ore thickness map as the target image and the X-ray transmission grayscale map as the source image, calculate the distance between the feature point descriptors in the source image and the target image, and determine the feature point descriptors in the source image according to the distance between the feature point descriptors. Feature points ~ And its matching in the target image Feature points ~ ;

[0034] S3.5, the source image Feature points ~ And its matching in the target image Feature points ~ Substitute the following formula to solve the affine transformation matrix :

[0035] , ,

[0036] In the above formula, , , , , and is the affine transformation matrix The transformation coefficients in ;

[0037] S3.6, aligning the ore thickness map and the X-ray transmission grayscale map in space according to the following formula to achieve the association between the grayscale value and thickness value of each pixel in the X-ray transmission grayscale map;

[0038] ,

[0039] In the above formula, The ore thickness map after spatial alignment Pixels The pixel coordinates of .

[0040] Optionally, in step S4, the pixel grayscale value and the associated thickness value of the pixel point in the X-ray transmission grayscale image are decoupled by using a curve fitting formula, and the function expression of the average mass absorption grayscale image composed of the average mass absorption coefficient of the reaction material type of the ore is obtained as follows:

[0041] ,

[0042] In the above formula, is the average mass absorption pixel in the grayscale image The transmission gray value, is the pixel point in the X-ray transmission grayscale image The pixel gray value, is the associated thickness value in the spatially aligned ore thickness map, , , , and is the fitting coefficient.

[0043] Optionally, step S5 includes:

[0044] S5.1, using the image edge extraction algorithm Canny operator to calculate the gradient amplitude image of the average mass absorption grayscale image of the ore according to the following formula:

[0045] ,

[0046] ,

[0047] ,

[0048] In the above formula, is the extracted gradient magnitude image, and They are gradient magnitude images of Quantity and Quantity, is the average mass absorption grayscale image of the ore, Represents the convolution operation;

[0049] S5.2, for gradient magnitude image The non-maximum suppression algorithm NMS is used to remove non-edge points, and only the local gradient maximum points are retained as edge points; then the edge points in the gradient amplitude image are classified into strong edge points, weak edge points, and non-edge points. By connecting the strong edge points and the weak edge points, these edge pixels are used as seed points to obtain the pixel point set of a single ore using the regional growing method. .

[0050] Optionally, the function expression for calculating the average grayscale value of a single ore pixel set in step S6 is:

[0051] ,

[0052] In the above formula, is the average gray value of a single ore pixel set of the ore, The number of points in the 3D point cloud data of the ore, A pixel set of a single ore Middle indivual The pixel value of a single ore Middle indivual The pixel value is equal to the average mass absorption grayscale image of the ore Middle Pixels Pixel value .

[0053] Optionally, step S6 also includes calculating the centroid position of a single ore according to the following formula to provide positioning information for subsequent blowing separation actions:

[0054] ,

[0055] ,

[0056] In the above formula, The centroid position Coordinates and coordinate, They are the first Points Coordinates and coordinate, The number of points in the 3D point cloud data of the ore.

[0057] In addition, the present invention also provides a three-dimensional information recognition-assisted ore sorting decision system, comprising a microprocessor and a memory connected to each other, wherein the microprocessor is programmed or configured to execute the three-dimensional information recognition-assisted ore sorting decision method.

[0058] In addition, the present invention also provides a computer-readable storage medium, in which a computer program or instruction is stored, and the computer program or instruction is programmed or configured to execute the three-dimensional information recognition-assisted ore sorting decision method through a processor.

[0059] In addition, the present invention also provides a computer program product, including a computer program or instructions, which are programmed or configured to execute the three-dimensional information recognition-assisted ore sorting decision method through a processor.

[0060] Compared with the prior art, the present invention has the following advantages: the three-dimensional information recognition-assisted ore sorting decision method of the present invention obtains the thickness data of the ore by scanning and measuring the ore through a multi-line laser radar device, and decouples the ore thickness information contained in the X-ray transmission image with the help of space-time registration technology. The entire method does not require the preparation of standard samples, thus solving the problem that the ore thickness affects the grayscale value of the X-ray transmission image. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 Schematic diagram of the basic flow of the method of the embodiment of the present invention.

[0062] Figure 2 It is a schematic diagram of the structure of an ore separator in an embodiment of the present invention.

[0063] Figure 3 This is a diagram of ore thickness obtained in an embodiment of the present invention.

[0064] Figure 4 : is an X-ray transmission grayscale image obtained in an embodiment of the present invention.

[0065] Figure 5 This is an X-ray transmission grayscale image of the unit thickness of the ore obtained in the embodiment of the present invention.

[0066] Figure 6 It is a schematic diagram comparing the ore separation results of the method according to the embodiment of the present invention and the existing method. DETAILED DESCRIPTION

[0067] The embodiments of the present invention will be described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention. In the description of the present invention, if there is a description of the first, the second, etc., it is only for the purpose of distinguishing the technical features, and cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features. If it involves the description of the orientation, such as the orientation or position relationship indicated by the upper and lower indications, the above-mentioned orientation description is 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. Unless otherwise clearly defined, the terms such as setting, installing, connecting, etc. in the description of the present invention should be understood in a broad sense, and the technicians in the relevant technical field can reasonably determine the specific meaning of the above terms in the present invention in combination with the specific content of the technical solution.

[0068] like Figure 1 As shown in FIG. 1 , the three-dimensional information recognition-assisted ore sorting decision-making method of this embodiment includes the following steps:

[0069] S1, obtaining the three-dimensional point cloud data and X-ray transmission grayscale image of the ore;

[0070] S2, projecting the three-dimensional point cloud data into the grid cells on the two-dimensional plane where the ore conveyor belt is located, calculating the thickness value of each grid cell and using it as the thickness value of the pixel point, thereby obtaining the ore thickness map;

[0071] S3, aligning the space of the ore thickness map and the X-ray transmission grayscale map to achieve the association between the grayscale value of each pixel in the X-ray transmission grayscale map and the thickness value;

[0072] S4, decoupling the pixel grayscale values ​​and the associated thickness values ​​of the pixel points in the X-ray transmission grayscale image using a curve fitting formula, to obtain an average mass absorption grayscale image consisting of average mass absorption coefficients of the reaction material types of the ore;

[0073] S5, performing edge extraction on the average mass absorption grayscale image of the ore to obtain a pixel point set of a single ore;

[0074] S6, calculate the average gray value of a single ore pixel set, and compare the average gray value with a preset threshold. If the average gray value is greater than the preset threshold, the ore is determined to be valid ore, otherwise the ore is determined to be waste rock.

[0075] Figure 2Schematic diagram of the structure of the ore sorting machine in this embodiment. In step S1 of this embodiment, ore is collected and the ore three-dimensional point cloud data and X-ray transmission grayscale image are respectively collected by a high-precision linear array laser radar and an X-ray detector.

[0076] Assume that the ore point cloud data is ,in is the index of the point. The goal of the ore thickness map is to map these three-dimensional points to the two-dimensional plane (xy plane) where the belt is located, and record the z-axis height of each position. Step S2 is used to project the three-dimensional point cloud data obtained by the multi-line laser radar scanning, with the projection direction perpendicular to the detector receiving plane, calculate the thickness value of each pixel point on the projection plane, and obtain the ore thickness map; specifically, step S2 in this embodiment includes:

[0077] S2.1, divide the xy two-dimensional plane where the ore conveyor belt is located into The size of each grid cell is ,in and are the width and height of the grid cell respectively, and any index is The coordinate range of the grid cells is:

[0078] ,

[0079] ,

[0080] In the above formula, are the coordinates in the grid cell, and Respectively represent ore point cloud data All coordinate points are Axis and The minimum value on the axis;

[0081] S2.2, for each point in the 3D point cloud data of the ore , find the index of the grid cell it falls on according to the following formula :

[0082] , ;

[0083] S2.3, for any index respectively The thickness of the grid cell is calculated according to the following formula:

[0084] ,

[0085] In the above formula, The index is The thickness of the grid cell, For the index The points in the grid cell The number of For the Points Axis coordinates;

[0086] S2.4, taking the grid unit as the pixel point and the thickness value of the grid unit as the pixel value of the pixel point, the size is obtained Ore thickness map ,Right now:

[0087] ,

[0088] In the above formula, is a grid representation of the ore thickness map. The ore thickness map obtained in this embodiment is as follows: Figure 3 shown.

[0089] In this embodiment, step S3 includes:

[0090] S3.1, the ore thickness map and X-ray transmission grayscale map are respectively extracted using the Harris corner point detection method to obtain the Harris response map :

[0091] ,

[0092] In the above formula, Harris response graph Central coordinates The pixel value of is the image gradient matrix, is the image gradient matrix The determinant of is the image gradient matrix , and there are:

[0093] , , ,

[0094] In the above formula, and X-ray transmission grayscale images exist and The gradient of the direction, is an empirical constant (usually 0.04 to 0.06); if the pixel value If it is greater than the preset threshold, the coordinates are determined The pixel point is the key feature point;

[0095] S3.2, construct the scale space of each key feature point according to the following formula (to ensure that the descriptor is invariant to scale changes):

[0096] ,

[0097] In the above formula, is the scale space of key feature points, is the pixel value of the key feature point, are the coordinates of the key feature points, is the scale factor; the gradient size and direction in the neighborhood of each key feature point are calculated according to the scale space of the key feature point, and a direction histogram containing multiple intervals to cover 360 degrees is created. The highest peak in the direction histogram is selected as the main direction of the key feature point, which can be expressed as:

[0098] ,

[0099] In this embodiment, a direction histogram containing 36 intervals is specifically created to cover 360 degrees. Assuming that the gradient direction of a point is 18.7 degrees, it will fall into the interval range of 10-19 degrees, and the increase in this interval is proportional to the gradient size of the point;

[0100] S3.3, extract a pixel area of ​​a specified size in the main direction of the key feature point and rotate it in the main direction, and divide the pixel area of ​​the specified size into multiple grid-like small blocks. In this embodiment, specifically extract a 16×16 pixel area in the main direction of the feature point and rotate it in the main direction. Divide the area into 4×4 small blocks; calculate the gradient histograms in 8 directions around each small block, and splice all the gradient histograms to generate a 128-dimensional feature point descriptor, thereby obtaining the feature point descriptors of the ore thickness map and the X-ray transmission grayscale map respectively;

[0101] S3.4, taking the ore thickness map as the target image and the X-ray transmission grayscale image as the source image, calculate the distance between the feature point descriptors in the source image and the target image. The commonly used distance metric is the Euclidean distance :

[0102] ,

[0103] in, and are the feature point descriptors of the source image and the target image respectively. The distance between the feature point descriptors is used to complete the matching between feature points, and all feature point pairs are recorded. and In this embodiment, the distances between the feature point descriptors are used to determine the Feature points ~ And its matching in the target image Feature points ~ ;

[0104] S3.5, the source image Feature points ~ And its matching in the target image Feature points ~ Substitute the following formula to solve the affine transformation matrix :

[0105] , ,

[0106] In the above formula, , , , , and is the affine transformation matrix The transformation coefficients in; the affine transformation matrix that satisfies the least squares method can be obtained through the overdetermined equations , if a series of matching point pairs of the source image and the target image are recorded as and ; and and , just put it into the above formula, and we can get the affine transformation 6 transformation coefficients of;

[0107] S3.6, aligning the ore thickness map and the X-ray transmission grayscale map in space according to the following formula to achieve the association between the grayscale value and thickness value of each pixel in the X-ray transmission grayscale map;

[0108] ,

[0109] In the above formula, The ore thickness map after spatial alignment Pixels Then the ore thickness map and the X-ray transmission grayscale image can be spatially aligned according to the affine transformation matrix, that is, for any pixel coordinate of the X-ray transmission grayscale image You can find the pixel coordinates associated with it in the ore thickness map .

[0110] The X-ray transmission grayscale image and the ore thickness image are recorded as and , any pixel X-ray transmission grayscale image per unit thickness ,in and The correlation relationship between them can be calculated by the above affine transformation matrix. In step S4 of this embodiment, the pixel grayscale value and the associated thickness value of the pixel point in the X-ray transmission grayscale image are decoupled by using a curve fitting formula, and the function expression of the average mass absorption grayscale image composed of the average mass absorption coefficient of the reaction material type of the ore is obtained:

[0111] ,

[0112] In the above formula, is the average mass absorption pixel in the grayscale image The transmission gray value, is the pixel point in the X-ray transmission grayscale image The pixel gray value, is the associated thickness value in the spatially aligned ore thickness map, , , , and is the fitting coefficient. The derivation process of the function expression of the average mass absorption grayscale diagram composed of the average mass absorption coefficient of the reaction material type of the ore includes: preparing a mineral standard sample and determining the average mass absorption coefficient. Preparing natural, ultra-high-grade, standard test samples of different thicknesses and determining the first three main mineral components The absorption coefficient . The sample preparation process is as follows: weigh a certain amount of ultra-high-grade target ore powder particles respectively, grind them in an agate mortar, and add ethanol solution during the process to improve the grinding efficiency; place the powder suspension in a tungsten carbide grinding jar, add tungsten carbide balls, and grind in a planetary ball mill; after the ball milling, load the evenly ground metal ore powder into a customized graphite mold with different depths and pre-press it; place the mold filled with the mixed powder in a spark plasma sintering furnace for sintering, and use the customized graphite mold to prepare the standard test samples by sintering the powder samples of the same grade and different thickness after the above grinding. Fit the function of the average mass absorption coefficient and thickness signal intensity. Energy deposition of X-ray transmission grayscale image under continuous spectrum. The calculation function expression is:

[0113] ,

[0114] Where: is the maximum energy of incident X-ray photons; is the number of incident X-ray photons; When the X-ray energy is The detection efficiency of the detector; is the energy of the incident X-ray photon; The energy of the substance being measured is The average mass absorption coefficient at this time is proportional to the average atomic number and directly reflects the type of substance; is the thickness of the object. From this formula, we can see that the X-ray transmission grayscale image The value of is related to the thickness and average mass absorption coefficient of the object being measured, that is, ,in other words It is the X-ray transmission grayscale image and the thickness of the object The function can be obtained by curve fitting method to obtain the specific function expression. The specific function expression obtained by curve fitting method includes:

[0115] Using high-precision linear array laser radar and X-ray detectors to measure ore samples of different thicknesses, the thickness of the ore can be obtained. 、X-ray transmission grayscale images at high and low energy As well as the average mass absorption coefficients of the three main mineral components at high and low energies, the data are shown in Table 1 (taking copper oxide minerals with a single mineral component as an example).

[0116] Table 1: Average mass absorption coefficients of three main mineral components

[0117]

[0118] According to the data in Table 1, the thickness signal intensity ( ) and mineral components Average mass absorption coefficient The fitting curve of the graph is shown in Figure 2, where the average mass absorption coefficient is used as the dependent variable and the thickness signal intensity ( ) as the independent variable.

[0119] The main mineral components were analyzed by least square method (high-order polynomial). The average mass absorption coefficient and thickness signal intensity ( ) is fitted and corrected to obtain:

[0120] .

[0121] This expression has five unknown parameters, so at least five sets of ore measurement data with different thicknesses are needed to substitute into the expression to solve the parameters. The more data there is, the better the fitting effect will be.

[0122] Solve the equation to get the main mineral composition The average mass absorption coefficient and thickness signal intensity ( ) After the specific function expression is obtained, for any ore to be tested, after obtaining the X-ray transmission grayscale image And thickness map On this basis, the average mass absorption coefficient of the reaction substance type is directly calculated, and the function expression of the average mass absorption grayscale image composed of the average mass absorption coefficient of the reaction substance type of the ore is obtained. On this basis, the calculated average mass absorption coefficient grayscale image can be used as the input of various ore sorting algorithms for subsequent identification and positioning.

[0123] In this embodiment, step S5 includes:

[0124] S5.1, using the image edge extraction algorithm Canny operator to calculate the gradient amplitude image of the average mass absorption grayscale image of the ore according to the following formula:

[0125] ,

[0126] ,

[0127] ,

[0128] In the above formula, is the extracted gradient magnitude image, and They are gradient magnitude images of Quantity and Quantity, is the average mass absorption grayscale image of the ore, Represents the convolution operation;

[0129] S5.2, for gradient magnitude image The non-maximum suppression algorithm NMS (an existing well-known method) is used to remove non-edge points, and only the local gradient maximum points are retained as edge points; then the size relationship between the pixel gradient amplitude and the pre-set threshold is compared, and the edge points in the gradient amplitude image are classified into strong edge points, weak edge points, and non-edge points. By connecting the strong edge points and the weak edge points, these edge pixels are used as seed points, and finally the pixel point set of a single ore is obtained using the existing region growing method. .

[0130] The function expression for calculating the average grayscale value of a single ore pixel set in step S6 of this embodiment is:

[0131] ,

[0132] In the above formula, is the average gray value of a single ore pixel set of the ore, The number of points in the 3D point cloud data of the ore, A pixel set of a single ore Middle indivual The pixel value of a single ore Middle indivual The pixel value is equal to the average mass absorption grayscale image of the ore Middle Pixels Pixel value . Figure 4 is the X-ray transmission grayscale image obtained in this embodiment, Figure 5 This is the X-ray transmission grayscale image of the ore per unit thickness obtained in this example.

[0133] As an optional implementation, step S6 of this embodiment further includes calculating the centroid position of a single ore according to the following formula to provide positioning information for subsequent blowing separation actions:

[0134] ,

[0135] ,

[0136] In the above formula, The centroid position Coordinates and coordinate, They are the first Points Coordinates and coordinate, The number of points in the 3D point cloud data of the ore.

[0137] Figure 6 Schematic diagram for comparing the ore separation results of the method of this embodiment (this method) and the existing method (dual energy R value). Figure 6 It can be seen that the method of this embodiment can obtain better ore separation results than the existing method (dual-energy R value), and can effectively solve the problem that the dual-energy R value algorithm is affected by the ore thickness.

[0138] In summary, the three-dimensional information recognition-assisted ore sorting decision method of this embodiment obtains the thickness data of the ore by scanning and measuring the ore through multi-line laser radar equipment, and decouples the ore thickness information contained in the X-ray transmission image with the help of spatiotemporal registration technology. The entire method does not require the preparation of standard samples, and solves the problem that the ore thickness affects the grayscale value of the X-ray transmission image. It mainly has the following advantages: 1. This embodiment proposes to convert the three-dimensional point cloud data of the ore into the form of occupied voxel grids, adjust the spatial resolution so that the occupied voxel grid is infinitely close to the three-dimensional structure of the ore, and then calculate the number of voxel grids in the vertical direction of each pixel point on the X-ray detector imaging plane, so as to obtain the thickness map of the ore, which solves the problem that the multi-line laser radar scans the ore to obtain the three-dimensional point cloud, and cannot directly obtain the ore thickness distribution in the direction perpendicular to the X-ray detector imaging plane. 2. This embodiment proposes to use various feature detection operators such as SURF matching to extract feature points of ore thickness map and X-ray transmission grayscale map, solve the affine transformation matrix to complete spatial alignment, and realize that each pixel in the X-ray transmission grayscale map can be accurately associated with the ore thickness value at the pixel position, thereby completing the problem of accurate decoupling of pixel-level thickness information and grayscale information. 3. The dual-energy X-ray ore image recognition method based on thickness recognition provided in this embodiment can effectively solve the problem that the dual-energy R value algorithm is affected by the ore thickness, so that the algorithm can meet the problem of sorting irregular shaped ores and waste rocks.

[0139] In addition, this embodiment also provides a three-dimensional information recognition-assisted ore sorting decision system, including a microprocessor and a memory connected to each other, and the microprocessor is programmed or configured to execute the three-dimensional information recognition-assisted ore sorting decision method.

[0140] In addition, this embodiment also provides a computer-readable storage medium, in which a computer program or instruction is stored, and the computer program or instruction is programmed or configured to execute the three-dimensional information recognition-assisted ore sorting decision method through a processor.

[0141] In addition, this embodiment also provides a computer program product, including a computer program or instructions, which are programmed or configured to execute the three-dimensional information recognition-assisted ore sorting decision method through a processor.

[0142] Those skilled in the art should understand that the technical solutions provided by the embodiments of the present invention may be in the form of methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that instructions executed by the processor of a computer or other programmable data processing device generate instructions for implementing the functions in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the functions specified in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0143] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A three-dimensional information recognition-assisted ore sorting decision method, characterized in that: The steps include: S1, obtaining the three-dimensional point cloud data and X-ray transmission grayscale image of the ore; S2, projecting the three-dimensional point cloud data into the grid cells on the two-dimensional plane where the ore conveyor belt is located, calculating the thickness value of each grid cell and using it as the thickness value of the pixel point, thereby obtaining the ore thickness map; S3, aligning the space of the ore thickness map and the X-ray transmission grayscale map to achieve the association between the grayscale value of each pixel in the X-ray transmission grayscale map and the thickness value; S4, decoupling the pixel grayscale values ​​and the associated thickness values ​​of the pixel points in the X-ray transmission grayscale image using a curve fitting formula, to obtain an average mass absorption grayscale image consisting of average mass absorption coefficients of the reaction material types of the ore; S5, performing edge extraction on the average mass absorption grayscale image of the ore to obtain a pixel point set of a single ore; S6, calculate the average gray value of a single ore pixel set, and compare the average gray value with a preset threshold. If the average gray value is greater than the preset threshold, the ore is determined to be valid ore, otherwise the ore is determined to be waste rock.

2. The three-dimensional information recognition-assisted ore sorting decision method according to claim 1 is characterized in that: Step S2 includes: S2.1, divide the xy two-dimensional plane where the ore conveyor belt is located into The size of each grid cell is ,in and are the width and height of the grid unit respectively, and any index is The coordinate range of the grid cells is: ， ， In the above formula, are the coordinates in the grid cell, and Respectively represent ore point cloud data The minimum value of all coordinate points on the x-axis and y-axis; S2.2, for each point in the 3D point cloud data of the ore , find the index of the grid cell it falls on according to the following formula : ， ； S2.3, for any index respectively The thickness of the grid cell is calculated according to the following formula: ， In the above formula, The index is The thickness of the grid cell, For the index The points in the grid cell The number of For the Points Axis coordinates; S2.4, taking the grid unit as the pixel point and the thickness value of the grid unit as the pixel value of the pixel point, the size is obtained Ore thickness map .

3. The three-dimensional information recognition-assisted ore sorting decision method according to claim 1 is characterized in that: Step S3 includes: S3.1, the ore thickness map and X-ray transmission grayscale map are respectively extracted using the Harris corner point detection method to obtain the Harris response map : ， In the above formula, Harris response graph Central coordinates The pixel value of is the image gradient matrix, is the image gradient matrix The determinant of is the image gradient matrix , and there are: ， ， ， In the above formula, and X-ray transmission grayscale images exist and The gradient of the direction, is an empirical constant; if the pixel value If it is greater than the preset threshold, the coordinates are determined The pixel point is the key feature point; S3.2, construct the scale space of each key feature point according to the following formula: ， In the above formula, is the scale space of key feature points, is the pixel value of the key feature point, are the coordinates of the key feature points, is the scale factor; the gradient size and direction in the neighborhood of each key feature point are calculated according to the scale space of the key feature point, and a direction histogram containing multiple intervals to cover 360 degrees is created, and the highest peak in the direction histogram is selected as the main direction of the key feature point; S3.3, extract a pixel area of ​​a specified size in the main direction of the key feature point and rotate it according to the main direction, divide the pixel area of ​​the specified size into multiple grid-like small blocks, calculate the gradient histograms of the eight directions around each small block, and splice all the gradient histograms to generate a 128-dimensional feature point descriptor, thereby obtaining the feature point descriptors of the ore thickness map and the X-ray transmission grayscale image respectively; S3.4, taking the ore thickness map as the target image and the X-ray transmission grayscale map as the source image, calculate the distance between the feature point descriptors in the source image and the target image, and determine the feature point descriptors in the source image according to the distance between the feature point descriptors. Feature points ~ And its matching in the target image Feature points ~ ; S3.5, the source image Feature points ~ and its n feature points matched in the target image ~ Substitute the following formula to solve the affine transformation matrix : ， ， In the above formula, , , , , and is the affine transformation matrix The transformation coefficients in ; S3.6, aligning the ore thickness map and the X-ray transmission grayscale map in space according to the following formula to achieve the association between the grayscale value and thickness value of each pixel in the X-ray transmission grayscale map; ， In the above formula, The ore thickness map after spatial alignment Pixels The pixel coordinates of .

4. The three-dimensional information recognition-assisted ore sorting decision method according to claim 1 is characterized in that: In step S4, the pixel grayscale value and the associated thickness value of the pixel point in the X-ray transmission grayscale image are decoupled by using a curve fitting formula, and the function expression of the average mass absorption grayscale image composed of the average mass absorption coefficient of the reaction material type of the ore is obtained as follows: ， In the above formula, is the average mass absorption pixel in the grayscale image The transmission gray value, is the pixel point in the X-ray transmission grayscale image The pixel gray value, is the associated thickness value in the spatially aligned ore thickness map, , , , and is the fitting coefficient.

5. The three-dimensional information recognition-assisted ore sorting decision method according to claim 1 is characterized in that: Step S5 includes: S5.1, using the image edge extraction algorithm Canny operator to calculate the gradient amplitude image of the average mass absorption grayscale image of the ore according to the following formula: ， ， ， In the above formula, is the extracted gradient magnitude image, and They are gradient magnitude images of Quantity and Quantity, is the average mass absorption grayscale image of the ore, Represents the convolution operation; S5.2, for gradient magnitude image The non-maximum suppression algorithm NMS is used to remove non-edge points, and only the local gradient maximum points are retained as edge points; then the edge points in the gradient amplitude image are classified into strong edge points, weak edge points, and non-edge points. By connecting the strong edge points and the weak edge points, these edge pixels are used as seed points to obtain the pixel point set of a single ore using the regional growing method. .

6. The three-dimensional information recognition-assisted ore sorting decision method according to claim 1 is characterized in that: The function expression for calculating the average grayscale value of a single ore pixel set in step S6 is: ， In the above formula, is the average gray value of a single ore pixel set of the ore, The number of points in the 3D point cloud data of the ore, A pixel set of a single ore Middle indivual The pixel value of a single ore The i-th The pixel value is equal to the average mass absorption grayscale image of the ore Middle Pixels Pixel value .

7. The three-dimensional information recognition-assisted ore sorting decision method according to claim 1 is characterized in that: Step S6 also includes calculating the centroid position of a single ore according to the following formula to provide positioning information for subsequent blowing and sorting actions: ， ， In the above formula, The centroid position Coordinates and coordinate, They are the first Points Coordinates and coordinate, The number of points in the 3D point cloud data of the ore.

8. A three-dimensional information recognition-assisted ore sorting decision system, comprising a microprocessor and a memory connected to each other, characterized in that: The microprocessor is programmed or configured to execute the three-dimensional information recognition-assisted ore sorting decision-making method as described in any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program or instruction stored therein, characterized in that: The computer program or instruction is programmed or configured to execute the three-dimensional information recognition-assisted ore sorting decision-making method as described in any one of claims 1 to 7 through a processor.

10. A computer program product comprising a computer program or instructions, characterized in that The computer program or instruction is programmed or configured to execute the three-dimensional information recognition-assisted ore sorting decision-making method as described in any one of claims 1 to 7 through a processor.

Images