A system for obtaining ore grade and a multi-source data acquisition device

By setting up a multi-source data acquisition device between the ore crushing system and the grinding graded system, combining X-ray detectors and industrial cameras to collect multi-source data, and using the feature extraction and fusion network of the data processor, X-ray imaging technology is solved The problem of noise interference in the middle improves the accuracy of ore grade recognition.

CN114689621BActive Publication Date: 2024-11-12ZHONGYE-CHANGTIAN INT ENG CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202011595949.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-12-29
Publication Date
2024-11-12
Estimated Expiration
2040-12-29

AI Technical Summary

Technical Problem

When the prior art uses X-ray imaging technology to obtain ore grade, it cannot effectively remove noise interference caused by belt transmission, resulting in a decrease in ore grade accuracy.

Method used

Design a multi-source data acquisition device, combining X-ray detectors and industrial cameras, collects X-ray transmission high and low energy grayscale images and visible light images of ores, and generates ores through a data processor using feature extraction, feature fusion and regression prediction networks. Grade prediction value.

Benefits of technology

It improves the recognition accuracy of ore grade, reduces noise interference, enhances the multi-source fusion analysis capability of the data, and generates more accurate ore grade prediction values.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114689621B_ABST
    Figure CN114689621B_ABST
Patent Text Reader

Abstract

The present application relates to the technical field of ore detection, and provides a system for obtaining ore grade and a multi-source data acquisition device. The system for obtaining ore grade comprises the multi-source data acquisition device and a data processor. The multi-source data acquisition device comprises an ore conveyor belt and a protection channel, and an opening communicating with a ray source chamber is arranged above the protection channel; an X-ray source is arranged on the top of the ray source chamber, a collimator is arranged between the X-ray source and the ore conveyor belt, a dual-energy X-ray detector is arranged directly below the ore conveyor belt, and an industrial camera is arranged on the side top of the protection channel; in actual application, X-rays are emitted to the ore on the ore conveyor belt by the X-ray source, the visible light image of the ore is collected by the industrial camera, and the X-ray transmission high and low energy grayscale images are collected by the dual-energy X-ray detector, so as to obtain multi-source data for obtaining ore grade, and the data processor generates an ore grade prediction value according to the collected multi-source data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of ore detection, and in particular to a system for obtaining ore grade and a multi-source data acquisition device. Background Art

[0002] Intelligent perception and identification of ore properties is the basis for realizing the intelligentization of mineral processing production process in the field of intelligent mining. The core development directions of intelligent mineral processing: ore property identification, intelligent flotation equipment and genetic mineral processing technology. The grade of valuable elements, particle size and crushing characteristics of ore are the three most important basic properties among the various physical and chemical properties of ore. These three basic properties determine the theoretical basis and process measures for the separation and sorting of useful components and the rational or comprehensive utilization of products. In actual production, the fluctuations of the three directly affect the stability of mineral processing production and the economic benefits, which is of great significance to the efficient and intelligent control of mineral processing production.

[0003] Among them, ore grade refers to the content of useful components or useful minerals in ore per unit volume or unit weight. Ore grade is the main criterion for determining ore and non-ore, and for identifying ore grade. At present, an X-ray generator is usually installed above the belt that transports ore, and a high- and low-energy detector is installed below the belt. The X-ray generator is used to perform X-ray irradiation detection on the lump ore on the transport belt, and the high- and low-energy transmission grayscale images of the ore are collected by the high- and low-energy detectors. Then, the high- and low-energy transmission grayscale images are processed in real time to detect the content information of the target elements in the lump ore, thereby obtaining the ore grade.

[0004] However, in actual working conditions, the method of obtaining ore grade by simply using X-ray imaging technology cannot eliminate the influence of a large amount of noise. For example, the belt used to transport ore will be transmitted by X-rays at the same time, thereby generating interference data, affecting the imaging effect of the high and low energy transmission grayscale images of the ore, and reducing the accuracy of the obtained ore grade.

[0005] In summary, in order to improve the accuracy of identifying ore grade, providing a system for obtaining ore grade and a multi-source data acquisition device is a technical problem that needs to be urgently solved by those skilled in the art. Summary of the invention

[0006] In order to obtain ore grade more accurately, the present application provides a system for obtaining ore grade and a multi-source data acquisition device.

[0007] On one hand, the present application provides a multi-source data acquisition device for obtaining ore grade, wherein the ore grade acquisition device is arranged between an ore crushing system and a grinding and grading system, and comprises an ore conveyor belt and a protective channel semi-enclosing a section of the ore conveyor belt, an opening communicating with a radiation source chamber is arranged above the protective channel, and a screed plate is arranged at the ore entrance of the protective channel;

[0008] An X-ray source is arranged on the top of the ray source chamber, a collimator is arranged between the X-ray source and the ore conveyor belt, a dual-energy X-ray detector is arranged directly below the ore conveyor belt, and an industrial camera is arranged on the side top of the protection channel;

[0009] The X-rays emitted by the X-ray source are irradiated onto the ore on the ore conveyor belt after the beam width is adjusted by the collimator; the dual-energy X-ray detector is used to collect X-ray transmission high and low energy grayscale images of the ore, and the industrial camera is used to collect visible light images of the ore, and the image acquisition area of ​​the industrial camera covers the image acquisition area of ​​the dual-energy X-ray detector, and the image acquisition frequency of the industrial camera is consistent with the image acquisition frequency of the dual-energy X-ray detector.

[0010] Optionally, a shielding body is provided inside the ray source chamber, and the shielding body is a conical structure, the narrow end of the conical structure is connected to the exit port of the X-ray source, and the collimator is arranged at the wide end of the conical structure.

[0011] Optionally, the shielding body includes an inner steel plate and an outer lead plate.

[0012] Optionally, an IP65-level industrial air conditioner is also provided inside the ray source chamber.

[0013] Optionally, a device controller is also provided inside the ray source chamber, and the controller is connected to the X-ray source, the dual-energy X-ray detector and the industrial camera.

[0014] Optionally, an X-ray source slot is provided on the top of the ray source chamber, and the outer shell of the ray source chamber is sealed by a sealing strip or a sealant.

[0015] Optionally, protective lead plates are provided on the outside of the protection channel, including a top protective lead plate, two side protective lead plates and a bottom protective lead plate, and the thickness of the lead plates are 5 mm, 4 mm and 7 mm respectively.

[0016] Optionally, the material of the dual-energy X-ray detector is GGAG scintillator.

[0017] The second aspect of the present application provides a system for obtaining ore grade, the system for obtaining ore grade comprising a multi-source data acquisition device for obtaining ore grade provided in the first aspect of the present application;

[0018] The system for obtaining ore grade also includes a data processor, which is connected to the dual-energy X-ray detector and the industrial camera, and is used to obtain the X-ray transmission high and low energy grayscale images collected by the dual-energy X-ray detector and the visible light images collected by the industrial camera; and generate an ore grade prediction value based on the X-ray transmission high and low energy grayscale images and the visible light images.

[0019] Optionally, when the data processor generates the ore grade value according to the X-ray transmission high and low energy grayscale images and the visible light image, the data processor performs the following steps:

[0020] Inputting the visible light image and the X-ray transmission high and low energy grayscale image into a pre-built grade prediction model, wherein the grade prediction model includes a feature extraction network, a feature fusion network and a regression prediction network;

[0021] Using the feature extraction network to extract a group of visible light feature maps of the visible light image, and to extract a group of high energy feature maps and a group of low energy feature maps of the X-ray transmission high and low energy grayscale image;

[0022] Using the feature fusion network, a group of visible light feature maps are internally fused to obtain a group of visible light fusion feature maps; a group of high energy feature maps are internally fused to obtain a group of high energy fusion feature maps; and a group of low energy feature maps are internally fused to obtain a group of low energy fusion feature maps;

[0023] And, using the feature fusion network to fuse the visible light fusion feature map, the high energy fusion feature map and the low energy fusion feature map to obtain a total fusion feature map;

[0024] The regression prediction network is used to perform regression prediction on the total fusion characteristic map to generate an ore grade prediction value.

[0025] It can be seen from the above technical solutions that the present application provides a system for obtaining ore grade and a multi-source data acquisition device, wherein the system for obtaining ore grade comprises a multi-source data acquisition device for obtaining ore grade and a data processor, wherein the multi-source data acquisition device for obtaining ore grade is arranged between an ore crushing system and a grinding and grading system, and comprises an ore conveyor belt and a protective channel that semi-encloses a section of the ore conveyor belt, wherein an opening connected to a ray source chamber is arranged above the protective channel, and a screed plate is arranged at the ore entrance of the protective channel; an X-ray source is arranged on the top of the ray source chamber, a collimator is arranged between the X-ray source and the ore conveyor belt, a dual-energy X-ray detector is arranged directly below the ore conveyor belt, and an industrial camera is arranged on the side top of the protective channel;

[0026] In actual application, X-rays are emitted to the ore on the ore conveyor belt by the X-ray source, the industrial camera collects visible light images of the ore, and the dual-energy X-ray detector collects X-ray transmission high and low energy grayscale images, wherein the X-ray transmission high and low energy grayscale images include X-ray transmission high energy grayscale images and X-ray transmission low energy grayscale images, the visible light image, the X-ray transmission high energy grayscale image and the X-ray transmission low energy grayscale image together constitute multi-source data for acquiring ore grade, and the data processor generates an ore grade prediction value based on the collected multi-source data. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solution of the present application, the drawings required for use in the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0028] Figure 1 A schematic diagram of the principle structure of a multi-source data acquisition device for obtaining ore grade provided in an embodiment of the present application;

[0029] Figure 2 A schematic diagram of the main structure of a multi-source data acquisition device for obtaining ore grade provided in an embodiment of the present application;

[0030] Figure 3 A schematic diagram of the side structure of a multi-source data acquisition device for obtaining ore grade provided in an embodiment of the present application;

[0031] Figure 4 A schematic diagram of the internal structure of the radiation source chamber provided in an embodiment of the present application.

[0032] In the figure, 1-ore conveyor belt, 2-protection channel, 3-ray source chamber, 31-shielding body, 32-IP65 industrial air conditioner, 33-controller, 4-screed plate, 5-X-ray source, 6-collimator, 7-dual-energy X-ray detector, 8-industrial camera. DETAILED DESCRIPTION

[0033] In order to obtain ore grade more accurately, Figure 1 The figure shows a schematic diagram of the principle structure of a multi-source data acquisition device for obtaining ore grade provided in an embodiment of the present application. A multi-source data acquisition device for obtaining ore grade provided in a first aspect of an embodiment of the present application is arranged between an ore crushing system and a grinding and grading system; it comprises an ore conveyor belt 1, one end of the ore conveyor belt 1 is connected to the ore crushing system, and the other end is connected to the grinding and grading system, and is used to convey the ore crushed by the ore crushing system to the ore crushing system. It should be noted that the purpose of the multi-source data acquisition device provided in an embodiment of the present application is to collect multi-source data for obtaining ore grade, and it is not limited to being installed between an ore crushing system and a grinding and grading system, and it can also be installed in other areas according to actual engineering requirements, or be set up separately for use.

[0034] A semi-enclosed protection channel 2 is provided in a section of the ore conveyor belt 1, and an opening connected to the radiation source chamber 3 is provided above the protection channel 2; a leveling plate 4 is provided at the ore entrance of the protection channel 2, and the leveling plate 4 is used to level the ore on the ore conveyor belt 1 to ensure that the effective light beam emitted by the X-ray source 5 arranged on the top of the radiation source chamber 3 can penetrate the thickest part of the ore layer.

[0035] A collimator 6 is provided between the X-ray source 5 and the ore conveyor belt 1 , and the collimator 6 controls the effective beam width passing through the collimation slit. The X-rays emitted by the X-ray source 5 are irradiated on the ore of the ore conveyor belt 1 after the beam width is adjusted by the collimator 6 .

[0036] A dual-energy X-ray detector 7 is arranged directly below the ore conveyor belt 1, and the dual-energy X-ray detector 7 is in the protection channel 2; an industrial camera 8 is arranged on the side top of the protection channel 2. It should be noted that it is necessary to ensure that the industrial camera 8 is outside the ray path area of ​​the X-ray source 5 to avoid interference of the industrial camera 8 to the ray. Among them, the industrial camera 8 uses a 300W pixel high-definition industrial camera, and the imaging quality is sufficient for particle size recognition of 10mm particle size grade ores.

[0037] Among them, the dual-energy X-ray detector 7 is used to collect X-ray transmission high and low energy grayscale images of the ore, and the industrial camera 8 is used to collect visible light images of the ore, and the image acquisition area of ​​the industrial camera 8 covers the image acquisition area of ​​the dual-energy X-ray detector 7. The image acquisition frequency of the industrial camera 8 is consistent with the image acquisition frequency of the dual-energy X-ray detector 7 to ensure that the industrial camera 8 and the dual-energy X-ray detector 7 can collect multi-source data for the same batch of ores.

[0038] An embodiment of the present application provides a multi-source data acquisition device for obtaining ore grade, wherein the X-ray source 5 emits X-rays to the ore on the ore conveyor belt 1, the industrial camera 8 collects visible light images of the ore, and the dual-energy X-ray detector 7 collects X-ray transmission high and low energy grayscale images, wherein the X-ray transmission high and low energy grayscale images include X-ray transmission high energy grayscale images and X-ray transmission low energy grayscale images. In actual application, the visible light image, the X-ray transmission high energy grayscale image and the X-ray transmission low energy grayscale image together constitute multi-source data for obtaining ore grade.

[0039] To ensure the safety of operators and the public at the equipment site, Figure 3 As shown in FIG. 1 , a schematic diagram of the side view structure of a multi-source data acquisition device for obtaining ore grade provided in an embodiment of the present application is shown. Figure 4 As shown, it is a schematic diagram of the internal structure of the ray source chamber provided in the embodiment of the present application. In some embodiments of the present application, the multi-source data acquisition device is provided with ray source end protection and channel protection, wherein the ray source end protection is a shielding body 31 provided inside the ray source chamber 3, the shielding body 31 is a conical structure, the narrow end of the conical structure is connected to the exit port of the X-ray source 5, the collimator 6 is provided at the wide end of the conical structure, and the tube voltage of the X-ray source 5 is 200KV and the tube current is 1.5mA. A ray channel is formed between the exit port of the X-ray source 5 and the collimator 6, and the shielding body 31 includes an inner steel plate and an outer lead plate, so as to shield the ray source end.

[0040] The channel protection is a protective lead plate arranged outside the protective channel 2, including a top protective lead plate, two side protective lead plates and a bottom protective lead plate, and the thickness of the lead plates is 5mm, 4mm and 7mm respectively. The protective channel 2 is an open structure, and there is a certain amount of scattering inside the channel. According to the scattering formula, the thickness of the top protective lead plate and the two side protective lead plates are 5mm and 4mm respectively, so as to protect the scattered radiation. For the radiation protection of the dual-energy X-ray detector 7, considering the extreme case where no ore is conveyed on the ore conveyor belt 1, at this time, the radiation at the bottom is the largest, and a bottom protective lead plate is arranged at the bottom of the protective channel 2, and the thickness of the bottom protective lead plate is 7mm.

[0041] Since the multi-source data acquisition device is arranged between the ore crushing system and the grinding and grading system, there is a large amount of ore dust at the collection site. In order to protect the working environment of the X-ray source 5, in some embodiments of the present application, an X-ray source card slot is arranged on the top of the ray source chamber 3, and the X-ray source card slot is used to install the X-ray source 5, and is designed with a shock-absorbing structure to eliminate the vibration generated by the large ore equipment at the collection site to affect the operation of the X-ray source 5, and the outer shell of the ray source chamber 3 is sealed by a sealing strip or sealant to prevent external dust from entering the ray source chamber 3 and affecting the operation of the X-ray source 5.

[0042] Further, such as Figure 2 As shown, it is a schematic diagram of the main structure of a multi-source data acquisition device for obtaining ore grade provided in an embodiment of the present application; in some embodiments of the present application, an IP65-level industrial air conditioner 32 is also provided inside the ray source chamber 3 to ensure that the X-ray source 5 can work in a good environment, ensure the accuracy of the data, and extend the working life.

[0043] Further, such as Figure 2 and Figure 3 As shown, a device controller 33 is also provided inside the ray source chamber 3, and the controller 33 is connected to the X-ray source 5, the dual-energy X-ray detector 7 and the industrial camera 8. The controller 33 is used to control the X-ray source 5 to emit X-rays, and is also used to control the dual-energy X-ray detector 7 to collect X-ray transmission high and low energy grayscale images of the ore, and is used to control the industrial camera 8 to collect visible light images of the ore, and is also used to control the start, stop and transportation speed of the ore conveyor belt 1.

[0044] It should be noted that Figure 2 This is a schematic diagram of the main structure of the multi-source data acquisition device, that is, a structural diagram parallel to the ore conveying direction. Figure 3 This is a side view structural diagram of the multi-source data acquisition device, that is, a structural diagram perpendicular to the ore conveying direction.

[0045] Further, such as Figure 3 As shown, the dual-energy X-ray detector 7 is enclosed and arranged below the ore conveyor belt 1, and the material of the dual-energy X-ray detector 7 is GGAG scintillator, and is arranged in a linear shape, for detecting the dual-energy X-ray transmission signal emitted by the X-ray source 5. Gadolinium gallium aluminum garnet (GGAG) is an ideal scintillator material with a high density (6g / cm 3 ), fast decay rate, small afterglow, high light yield, and cubic phase structure.

[0046] A second aspect of an embodiment of the present application provides a system for obtaining ore grade. The system for obtaining ore grade includes a multi-source data acquisition device for obtaining ore grade provided by the first aspect of the embodiment of the present application, and also includes a data processor. The data processor is connected to a dual-energy X-ray detector and an industrial camera, and is used to obtain X-ray transmission high and low energy grayscale images collected by the dual-energy X-ray detector, and to obtain visible light images collected by the industrial camera; and to generate an ore grade prediction value based on the X-ray transmission high and low energy grayscale images and the visible light images.

[0047] Wherein, when the data processor generates the ore grade value according to the X-ray transmission high and low energy grayscale image and the visible light image, the following steps are performed:

[0048] The visible light image and the X-ray transmission high and low energy grayscale images are input into a pre-built grade prediction model, wherein the grade prediction model includes a feature extraction network, a feature fusion network and a regression prediction network.

[0049] The feature extraction network is used to extract a group of visible light feature maps of the visible light image, and a group of high-energy feature maps and a group of low-energy feature maps of the X-ray transmission high- and low-energy grayscale image.

[0050] The feature fusion network is used to perform internal feature fusion on a group of visible light feature maps to obtain a group of visible light fusion feature maps; and a group of high energy feature maps are performed internal feature fusion to obtain a group of high energy fusion feature maps; and a group of low energy feature maps are performed internal feature fusion to obtain a group of low energy fusion feature maps.

[0051] And, the feature fusion network is used to fuse the visible light fusion feature map, the high energy fusion feature map and the low energy fusion feature map to obtain the total fusion feature map.

[0052] The regression prediction network is used to perform regression prediction on the total fusion characteristic map to generate an ore grade prediction value.

[0053] The input of the feature extraction network is a set of visible light images, a set of X-ray transmission high-energy grayscale images, and a set of X-ray transmission low-energy grayscale images, and the output is three sets of 26×26×512 feature maps. The four-layer convolution module in the feature extraction network corresponds to four convolution stages in sequence. The first-layer convolution module corresponds to the first convolution stage Conv1, including 64 convolution kernels of size 3×3, stride 2, and output feature map size of 208×208×64. The second-layer convolution module corresponds to the second convolution stage Conv2, including 128 convolution kernels of size 3×3, stride 2, and output feature map size of 104×104×128. The third-layer convolution module corresponds to the third convolution stage Conv3, including 256 convolution kernels of size 3×3, stride 2, and output feature map size of 52×52×256. The fourth-layer convolution module corresponds to the fourth convolution stage Conv4, including 512 convolution kernels of size 3×3, stride 2, and output feature map size of 26×26×512.

[0054] After multiple convolution operations of the feature extraction network, the input original image is quickly downsampled into small-size and high-dimensional feature information, discarding redundant information in the image and enriching the feature information of the image.

[0055] After the input X-ray transmission high-energy grayscale image, X-ray transmission low-energy grayscale image and visible light image are subjected to the feature extraction step, three sets of 26×26×512 feature maps are obtained. Next, the feature fusion operation is used to fuse the three sets of feature maps at the feature level. The method adopted in this embodiment is the corresponding pixel addition method, and finally a feature map of size 26×26×1 is output. The specific steps are as follows:

[0056] The first step is internal fusion of feature maps. That is, the 26×26×512 feature map generated for an input image has 512 channels with a size of 26×26. The pixel values ​​at the coordinate position (0, 0) in the 512 channels are accumulated on the 26×26 pixel matrix, so that the value is reassigned to the value at the corresponding coordinate position in a new feature map, and then calculated in the order of rows first and columns to the coordinate position (25, 25). In this way, a fused new feature map of size 26×26×1 can be obtained, which corresponds to the original input image.

[0057] The second step is to fuse the feature maps. That is, after the three input images are internally fused in the first step, a 3×(26×26×1) feature map can be obtained. The corresponding pixels of the three sets of feature maps are then added. The process is similar to the first step, and finally a fused feature map with a size of 26×26×1 is obtained.

[0058] In the above feature fusion process, the size of the three groups of fused feature maps output by the internal fusion module of the feature map and the total fused feature map output by the fusion module between the feature maps are both 26×26×1.

[0059] The feature map of size 26×26×1 after feature fusion is used for regression prediction. For prediction problems, it can be divided into regression tasks and classification tasks according to whether the type of prediction output is a continuous real value or a discrete label. Because the key indicator of sintered ore is a continuous value, the prediction model is obviously a regression task. The regression task is divided into linear regression and nonlinear regression. The prediction of the key indicator of sintered ore is not a simple linear relationship, so this embodiment will use nonlinear regression. In the standard structure of the nonlinear regression neural network, each neuron is composed of a weighted sum and a nonlinear transformation, and then multiple neurons are arranged in layers and connected to form a neural network. Therefore, this embodiment uses the method of multi-layer perceptron for numerical nonlinear regression prediction.

[0060] The predicted output value can be obtained by inputting the feature map of size 26×26×1 after feature fusion into the regression prediction network. The regression prediction network adopts a multi-layer perceptron prediction network, including four fully connected layers. The first fully connected layer is used to convert the total fusion feature map of the input into a sequence. The second fully connected layer includes 1000 hidden layer nodes. The third fully connected layer includes 26 hidden layer nodes, and the fourth fully connected layer contains only one hidden layer node.

[0061] The function of the first fully connected layer is to pull the two-dimensional feature map into a column to form a number series. The number of hidden layer nodes in the second fully connected layer is 1000, that is, the 676 input numbers are fully connected to the 1000 nodes, and each hidden layer node is connected to all the input numbers. The number of hidden layer nodes in the third fully connected layer is 26. The function of this layer is to quickly reduce the dimension and fully connect the 1000 input numbers to the 26 numbers. The fourth fully connected layer contains only one node, which is fully connected to the 26 input numbers. The value output by the fourth fully connected layer has a direct or indirect fully connected relationship with all the input data information. This value represents the abstract features of the X-ray transmission high-energy grayscale image, X-ray transmission low-energy grayscale image and visible light image that were initially input, that is, the predicted output. The value obtained after regression prediction is the final predicted output value, that is, the ore grade value.

[0062] It can be seen from the above technical scheme that the embodiment of the present application provides a system for obtaining ore grade and a multi-source data acquisition device, the system for obtaining ore grade includes a multi-source data acquisition device for obtaining ore grade and a data processor, the multi-source data acquisition device for obtaining ore grade is arranged between the ore crushing system and the grinding and grading system, including an ore conveyor belt 1 and a protective channel 2 semi-enclosed in a section of the ore conveyor belt 1, an opening connected to a ray source chamber 3 is arranged above the protective channel 2, and a screed plate 4 is arranged at the ore entrance of the protective channel 2; an X-ray source 5 is arranged on the top of the ray source chamber 3, a collimator 6 is arranged between the X-ray source 5 and the ore conveyor belt 1, a dual-energy X-ray detector 7 is arranged directly below the ore conveyor belt 1, and an industrial camera 8 is arranged on the side top of the protective channel 2.

[0063] In actual application, X-rays are emitted to the ore on the ore conveyor belt 1 by the X-ray source 5, the industrial camera 8 collects visible light images of the ore, and the dual-energy X-ray detector 7 collects X-ray transmission high and low energy grayscale images, wherein the X-ray transmission high and low energy grayscale images include X-ray transmission high energy grayscale images and X-ray transmission low energy grayscale images, the visible light image, the X-ray transmission high energy grayscale image and the X-ray transmission low energy grayscale image together constitute multi-source data for obtaining ore grade, and the data processor generates an ore grade prediction value based on the collected multi-source data.

[0064] The present application is described in detail above in conjunction with specific implementation methods and exemplary examples, but these descriptions cannot be understood as limiting the present application. Those skilled in the art understand that, without departing from the spirit and scope of the present application, a variety of equivalent replacements, modifications or improvements can be made to the technical solution of the present application and its implementation methods, all of which fall within the scope of the present application. The scope of protection of the present application shall be subject to the attached claims.

Claims

1. A multi-source data acquisition device for obtaining ore grade, wherein the ore grade acquisition device is arranged between an ore crushing system and a grinding and grading system, and is characterized in that: It comprises an ore conveyor belt (1) and a protective channel (2) which semi-encloses a section of the ore conveyor belt (1), an opening which is connected to a radiation source chamber (3) is arranged above the protective channel (2), and a screed plate (4) is arranged at the ore entrance of the protective channel (2); An X-ray source (5) is arranged on the top of the ray source chamber (3), a collimator (6) is arranged between the X-ray source (5) and the ore conveyor belt (1), a dual-energy X-ray detector (7) is arranged directly below the ore conveyor belt (1), and an industrial camera (8) is arranged on the side top of the protection channel (2); The X-rays emitted by the X-ray source (5) are irradiated onto the ore on the ore conveyor belt (1) after the beam width is adjusted by the collimator (6); the dual-energy X-ray detector (7) is used to collect X-ray transmission high- and low-energy grayscale images of the ore, and the industrial camera (8) is used to collect visible light images of the ore, and the image acquisition area of ​​the industrial camera (8) covers the image acquisition area of ​​the dual-energy X-ray detector (7), and the image acquisition frequency of the industrial camera (8) is consistent with the image acquisition frequency of the dual-energy X-ray detector (7).

2. A multi-source data acquisition device for obtaining ore grade according to claim 1, characterized in that: A shielding body (31) is arranged inside the ray source chamber (3); the shielding body (31) is a conical structure; the narrow end of the conical structure is connected to the exit port of the X-ray source (5); and the collimator (6) is arranged at the wide end of the conical structure.

3. A multi-source data acquisition device for obtaining ore grade according to claim 2, characterized in that: The shielding body (31) comprises an inner steel plate and an outer lead plate.

4. A multi-source data acquisition device for obtaining ore grade according to claim 1, characterized in that: An IP65-level industrial air conditioner (32) is also provided inside the radiation source chamber (3).

5. The multi-source data acquisition device for obtaining ore grade according to claim 1, characterized in that: A device controller (33) is also provided inside the ray source chamber (3), and the controller (33) is connected to the X-ray source (5), the dual-energy X-ray detector (7) and the industrial camera (8).

6. The multi-source data acquisition device for obtaining ore grade according to claim 1, characterized in that: An X-ray source slot is provided on the top of the ray source chamber (3), and the outer shell of the ray source chamber (3) is sealed by a sealing strip or a sealing glue.

7. The multi-source data acquisition device for obtaining ore grade according to claim 1, characterized in that: The outer side of the protection channel (2) is provided with protective lead plates, including a top protective lead plate, two side protective lead plates and a bottom protective lead plate, and the thickness of the lead plates are 5 mm, 4 mm and 7 mm respectively.

8. The multi-source data acquisition device for obtaining ore grade according to claim 1, characterized in that: The material of the dual-energy X-ray detector (7) is GGAG scintillator.

9. A system for obtaining ore grade, characterized in that: The system for obtaining ore grade comprises a multi-source data acquisition device for obtaining ore grade according to any one of claims 1 to 8; The system for obtaining ore grade also includes a data processor, which is connected to the dual-energy X-ray detector and the industrial camera, and is used to obtain the X-ray transmission high and low energy grayscale images collected by the dual-energy X-ray detector and the visible light images collected by the industrial camera; and generate an ore grade prediction value based on the X-ray transmission high and low energy grayscale images and the visible light images.

10. A system for obtaining ore grade according to claim 9, characterized in that: When the data processor generates the ore grade value according to the X-ray transmission high and low energy grayscale image and the visible light image, the data processor performs the following steps: Inputting the visible light image and the X-ray transmission high and low energy grayscale image into a pre-built grade prediction model, wherein the grade prediction model includes a feature extraction network, a feature fusion network and a regression prediction network; Using the feature extraction network to extract a group of visible light feature maps of the visible light image, and to extract a group of high energy feature maps and a group of low energy feature maps of the X-ray transmission high and low energy grayscale image; Using the feature fusion network, a group of visible light feature maps are internally fused to obtain a group of visible light fusion feature maps; and a group of high energy feature maps are internally fused to obtain a group of high energy fusion feature maps; and performing internal feature fusion on a set of low-energy feature maps to obtain a set of low-energy fused feature maps; And, using the feature fusion network to fuse the visible light fusion feature map, the high energy fusion feature map and the low energy fusion feature map to obtain a total fusion feature map; The regression prediction network is used to perform regression prediction on the total fusion characteristic map to generate an ore grade prediction value.