Copper ore pre-selection method based on generative adversarial network and low-rank linear pooling

Abstract

The present invention discloses a copper ore preselection method based on a generative adversarial network and low-rank linear pooling, comprising the following steps: obtaining a high-energy X-ray transmission image and a low-energy X-ray transmission image of the copper ore; fusing the high-energy and low-energy X-ray transmission images using an improved generative adversarial network to obtain a dual-energy X-ray fusion image; constructing a copper ore preselection model; inputting the high-energy, low-energy and dual-energy X-ray fusion images of the copper ore into the model for feature extraction, and fusing the extracted feature vectors using low-rank trilinear pooling and then classifying them to obtain a copper ore preselection result; in the copper ore preselection model designed by the present invention, by improving low-rank bilinear pooling and using a low-rank trilinear pooling feature fusion strategy, a joint representation solution of three image features is realized, which can adapt to the requirements of multimodal data input and further improve the classification performance of copper ore target ore and waste rock.

Description

Technical Field

[0001] The present invention relates to the technical field of copper ore preselection, and in particular to a copper ore preselection method based on a generative adversarial network and low-rank linear pooling. Background Art

[0002] Copper metal, with its electrical and thermal conductivity, strong ductility, and ease of alloying, has extensive applications in modern industry and is a key strategic resource in my country. However, my country's copper ore reserves are primarily poor, with relatively few rich ores. During the mining process, large amounts of surrounding rock and waste rock are mixed in, increasing the difficulty and cost of subsequent flotation and refining processes. Therefore, there is an urgent need to research ore sorting methods that improve the efficiency of copper ore waste separation. Industry often incorporates an ore pre-selection step between ore mining and post-processing. Pre-selection through ore sorting methods can improve the grade of ore entering subsequent processing, reduce the amount of waste rock entering, achieve ore pre-enrichment, effectively improve ore resource utilization, and reduce production costs. Common ore sorting methods suffer from a narrow application range, damage to ore morphology, and significant associated pollution. In recent years, intelligent photoelectric beneficiation methods based on artificial intelligence have been widely used in ore pre-selection. Intelligent photoelectric beneficiation methods utilize image color sorting and X-ray transmission technology to non-destructively collect ore information. They apply artificial intelligence algorithms, such as deep learning models, to quickly and accurately separate target ore from waste rock. The performance of ore classification algorithms directly determines classification efficiency. Therefore, developing deep learning-based copper ore pre-selection models has become essential for the efficient utilization of copper resources.

[0003] In recent years, deep learning-based image recognition methods have demonstrated strong performance in classification tasks. Dual-energy X-ray transmission technology uses the different attenuation effects of X-rays on different materials to identify material types by analyzing the intensity of the transmitted radiation. This technology does not destroy the material's morphology and is unaffected by differences in appearance, size, or magnetic properties. It has been widely used in photoelectric mineral processing. Using dual-energy X-ray imaging equipment to obtain high- and low-energy X-ray transmission images of copper ore, and processing these ore images with deep learning models to distinguish target ore from waste rock, offers a new solution for improving ore sorting efficiency. Deep learning-based ore sorting models, leveraging the automatic feature learning capabilities of deep neural networks, have achieved strong recognition results on both high- and low-energy X-ray transmission image datasets of copper ore. However, deep neural network image feature processing is often viewed as a black-box model, lacking interpretability. Current work applying deep learning models to ore sorting focuses on simple approaches to simultaneously learning features from both high- and low-energy image data, lacking in-depth research. Consequently, there is no guarantee that the models can effectively learn and utilize features from both high- and low-energy image data. Summary of the Invention

[0004] In response to the shortcomings of the existing technology, the present invention provides a copper ore pre-selection method based on generative adversarial networks and low-rank linear pooling, which aims to solve the problems in the background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a copper ore preselection method based on generative adversarial networks and low-rank linear pooling, comprising the following steps:

[0006] Step S1: obtaining a high-energy X-ray transmission image and a low-energy X-ray transmission image of the copper ore using a dual-energy X-ray imaging device;

[0007] Step S2: using an improved generative adversarial network to perform modal fusion on the high-energy X-ray transmission image and the low-energy X-ray transmission image of the copper ore to obtain a dual-energy X-ray fusion image of the copper ore;

[0008] The improved generative adversarial network adds a discriminator on the basis of the original generative adversarial network;

[0009] Step S3: constructing a copper ore pre-selection model, which includes a feature extraction coding layer, a feature fusion coding layer, and a classification decoding layer;

[0010] Step S4: inputting the high-energy X-ray transmission image, the low-energy X-ray transmission image and the dual-energy X-ray fusion image of the copper ore into a feature extraction coding layer to obtain high-energy, low-energy and dual-energy feature vectors;

[0011] Step S5: inputting the high-energy, low-energy and dual-energy feature vectors into the feature fusion coding layer, which fuses the high-energy, low-energy and dual-energy feature vectors using low-rank trilinear pooling to obtain a fused feature vector;

[0012] Step S6: Input the fused feature vector into the classification decoding layer for classification to obtain the pre-selection result of the copper ore.

[0013] Furthermore, the specific process of step S5 is as follows:

[0014] Assume that the high energy, low energy and dual energy eigenvectors are 、 and , solved by low-rank trilinear pooling 、 and The joint feature representation of is:

[0015] (1);

[0016] Where, is the first elements; is the weight parameter for trilinear pooling, controlling 、 、 The interaction of the corresponding elements in for The elements; for The elements; for The elements; For the Bias terms; For the A weight tensor; 、 、 They are 、 、 Dimensions; , , , , is the set of real numbers;

[0017] The weight tensor Decompose into the product of three 2D matrices: , Indicates that From the dimension Reduce to dimension Matrix of express The transpose of Indicates that From the dimension Reduce to dimension Matrix of express The transpose of Indicates that From the dimension Reduce to dimension Matrix of

[0018] The decomposed weight tensor Substitute into formula (1) to simplify formula (1):

[0019] (2);

[0020] Where, express The transpose of express The transpose of express The transpose of represents the Hadamard product; express The transpose of represents a vector of all 1s;

[0021] Reuse Replace the formula (2) ,definition 、 and , obtain the joint feature representation, that is, the fusion feature vector :

[0022] (3);

[0023] Where, 、 and Respectively 、 and The transpose of 、 and Respectively 、 、 The dimension reduction matrix of Represents the bias vector, dimension is ; express The transpose of Represents the weight matrix used for linear transformation.

[0024] Furthermore, the improved generative adversarial network consists of a discriminator , Discriminator and generator G; first, the high-energy X-ray transmission image and low-energy X-ray transmission image channel dimensions of the copper ore are superimposed together as the input of the generator G, and the generator outputs the fused image after processing. and the discriminator The high-energy X-ray transmission image and the low-energy X-ray transmission image of the copper ore are received respectively, and the high-energy X-ray transmission image and the low-energy X-ray transmission image of the copper ore are discriminated with the fusion image output by the generator. and the discriminator The discrimination result is input into the generator G, which adjusts the fused image and finally outputs the dual-energy X-ray fusion image of copper ore.

[0025] Furthermore, the generator G is composed of a five-layer convolutional neural network, and the first, second, third, fourth and fifth layers of the generator G are all convolutional layers; the first, second, third and fourth convolutional layers of the generator G are successively connected to a batch normalization layer and a Leaky ReLU activation layer, and the fifth convolutional layer is connected to a Tanh activation layer; wherein, the first and second convolutional layers of the generator G use a 5*5 convolution kernel, the third and fourth convolutional layers use a 3*3 convolution kernel, and the fifth convolutional layer uses a 1*1 convolution kernel; the convolutional layer step size of each layer in the generator G is set to 1, and no padding operation is performed in the convolution.

[0026] Furthermore, the discriminator and the discriminator The structure is the same, also composed of a five-layer convolutional neural network, the discriminator and the discriminator The first, second, third and fourth layers are convolutional layers, and the fifth layer is a linear layer; the discriminator and the discriminator The first, second, third, and fourth convolutional layers are followed by a batch normalization layer and a Leaky ReLU activation layer; the discriminator and the discriminator The first, second, third and fourth convolutional layers use 3*3 convolution kernels and the stride is set to 2.

[0027] Furthermore, the feature extraction encoding layer adopts three parallel and identically structured feature extraction encoders DenseNet to extract features from the high-energy X-ray transmission image, low-energy X-ray transmission image and dual-energy X-ray fusion image of copper ore, respectively, and obtain high-energy, low-energy and dual-energy feature vectors.

[0028] Further, improve the generator in the generative adversarial network The loss function Expressed as:

[0029] (4);

[0030] Where, It means fighting against loss; Indicates content loss; Represents a hyperparameter used to adjust the content loss in The weight in

[0031] Expressed as:

[0032] (5);

[0033] Where, Indicates expected value; Represents dual-energy X-ray fusion image.

[0034] Further, improve the discriminator in the generative adversarial network The loss function and the discriminator The loss function Expressed as:

[0035] (6);

[0036] (7);

[0037] Where, Represents high-energy X-ray transmission image; Represents a low-energy X-ray transmission image.

[0038] Compared with the existing technology, the present invention has the following beneficial effects:

[0039] (1) The present invention improves the collaborative design of generative adversarial networks and low-rank trilinear pooling, adopts multimodal adaptive adversarial training to target the differences in physical properties between high-energy and low-energy X-ray images, balances multi-source information in pixel-level fusion, and applies low-rank trilinear pooling to efficiently establish high-order interactions of the three image features of high-energy, low-energy, and fused images through tensor decomposition, thereby enhancing the joint characterization capability. This combined strategy further improves the model classification accuracy compared to a single method and significantly enhances the model sorting performance.

[0040] (2) The copper ore preselection model designed by the present invention improves low-rank bilinear pooling and adopts a low-rank trilinear pooling feature fusion strategy to realize the joint representation solution of the three image features of high-energy and low-energy X-ray transmission images and dual-energy X-ray fusion images extracted by the encoder. It can adapt to the requirements of multimodal data input and further improve the classification performance of copper ore target ore and waste rock. The improved generative adversarial network designed by the present invention adapts to the input of high-energy and low-energy transmission images of copper ore by adopting two discriminators, so that the generator learns the probability distribution of the two image data at the same time, realizes image fusion at the pixel level, and obtains a fusion image with stronger representation ability. The input data of the classification model is optimized in the early fusion stage, and the performance of the copper ore classification model is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 Flow chart of the method of the present invention.

[0042] Figure 2 This is a schematic diagram of the improved generative adversarial network structure of the present invention.

[0043] Figure 3 This is a schematic diagram of the structure of the copper ore pre-selection model of the present invention. DETAILED DESCRIPTION

[0044] like Figure 1 As shown, the present invention provides a technical solution: a copper ore pre-selection method based on a generative adversarial network and low-rank linear pooling, comprising the following steps:

[0045] Step S1: obtaining a high-energy X-ray transmission image and a low-energy X-ray transmission image of the copper ore based on a dual-energy X-ray imaging device.

[0046] Step S2: Using an improved generative adversarial network to perform modal fusion on the high-energy X-ray transmission image and the low-energy X-ray transmission image of the copper ore, a dual-energy X-ray fusion image of the copper ore is obtained.

[0047] like Figure 2 As shown in the figure, the improved generation adversarial network adds a discriminator on the basis of the original generation adversarial network (GAN); the improved generation adversarial network consists of the discriminator , Discriminator and generator G; first, the high-energy X-ray transmission image and low-energy X-ray transmission image channel dimensions of the copper ore are superimposed together as the input of the generator G, and the generator outputs the fused image after processing. and the discriminator The high-energy X-ray transmission image and the low-energy X-ray transmission image of the copper ore are received respectively, and the high-energy X-ray transmission image and the low-energy X-ray transmission image of the copper ore are discriminated with the fusion image output by the generator. and the discriminator The discrimination result is input into the generator G, which adjusts the fused image and finally outputs the dual-energy X-ray fusion image of copper ore.

[0048] Among them, the generator G consists of a five-layer convolutional neural network, and the first, second, third, fourth and fifth layers of the generator G are all convolutional layers; the first, second, third and fourth convolutional layers of the generator G are connected in sequence with a batch normalization layer and a Leaky ReLU activation layer, and the fifth convolutional layer is connected with a Tanh activation layer; among them, the first and second convolutional layers of the generator G use a 5*5 convolution kernel, the third and fourth convolutional layers use a 3*3 convolution kernel, and the fifth convolutional layer uses a 1*1 convolution kernel; the convolutional layer step size of each layer in the generator G is set to 1, and no padding operation is performed in the convolution.

[0049] Among them, the discriminator and the discriminator The structure is the same, also composed of a five-layer convolutional neural network, the discriminator and the discriminator The first, second, third and fourth layers are convolutional layers. At the same time, since the role of the discriminator is to determine whether the generated image is a real image, the fifth layer is a linear layer for classification; the discriminator and the discriminator The first, second, third, and fourth convolutional layers are followed by a batch normalization layer and a Leaky ReLU activation layer; the discriminator and the discriminator The first, second, third and fourth convolutional layers use 3*3 convolution kernels and the stride is set to 2.

[0050] Step S3: Construct a copper ore pre-selection model (LRTP-PTENet), such as Figure 3 As shown in FIG, the copper ore preselection model includes a feature extraction coding layer, a feature fusion coding layer and a classification decoding layer.

[0051] Step S4: inputting the high-energy X-ray transmission image, the low-energy X-ray transmission image and the dual-energy X-ray fusion image of the copper ore into the feature extraction coding layer to obtain high-energy, low-energy and dual-energy feature vectors.

[0052] Among them, the feature extraction coding layer uses three parallel and identically structured feature extraction encoders (DenseNet) to extract features from the high-energy X-ray transmission image, low-energy X-ray transmission image and dual-energy X-ray fusion image of copper ore, respectively, and obtain high-energy, low-energy and dual-energy feature vectors.

[0053] Step S5: The high-energy, low-energy and dual-energy feature vectors are input into the feature fusion coding layer. The feature fusion coding layer uses low-rank trilinear pooling to fuse the high-energy, low-energy and dual-energy feature vectors to obtain a fused feature vector.

[0054] The fully bilinear pooling method will produce high-dimensional features that are expanded twice, which is prone to the problem of too high a fusion feature dimension. This will consume a lot of computing resources and become a bottleneck limiting model performance.

[0055] Low-Rank Bilinear pooling (LRBP) decomposes a three-dimensional weight tensor required by full bilinear pooling into three two-dimensional weight matrices, thereby reducing the rank of the weight tensor and effectively reducing the number of computational parameters.

[0056] In view of the multimodal fusion of copper ore, the low-rank bilinear pooling method is extended, and the high-energy, low-energy and dual-energy feature vectors are set as 、 and , solved by low-rank trilinear pooling 、 and The joint feature representation of is:

[0057] (1);

[0058] Where, is the first elements; is the weight parameter for trilinear pooling, controlling 、 、 The interaction of the corresponding elements in for The elements; for The elements; for The elements; For the A bias term used to adjust the joint feature representation; For the A weight tensor, used for fusion of feature vectors; 、 、 They are 、 、 Dimensions; , , , , is a set of real numbers; the dimension of the fused feature vector can be set to ,but 、 .

[0059] Since the weight tensor is directly calculated is extremely difficult and time-consuming, so the weight tensor Decompose into the product of three 2D matrices: , Indicates that From the dimension Reduce to dimension Matrix of express The transpose of Indicates that From the dimension Reduce to dimension Matrix of express The transpose of Indicates that From the dimension Reduce to dimension The matrix of .

[0060] The decomposed weight tensor Substitute into formula (1) to simplify formula (1):

[0061] (2);

[0062] Where, express The transpose of express The transpose of express The transpose of represents the Hadamard product (element-wise product); express The transpose of represents a vector of all 1s.

[0063] Reuse Replace the formula (2) ,definition 、 and , obtain the joint feature representation, that is, the fusion feature vector :

[0064] (3);

[0065] Where, 、 and Respectively 、 and The transpose of 、 and Respectively 、 、 The dimension reduction matrix of Represents the bias vector, with dimension ; express The transpose of Represents the weight matrix used for linear transformation.

[0066] Step S6: Input the fused feature vector into the classification decoding layer for classification to obtain the pre-selection result of the copper ore.

[0067] Among them, the generator in the improved generative adversarial network The loss function Expressed as:

[0068] (4);

[0069] Where, Denotes adversarial loss, which is used to measure the difference between the samples generated by the generator and the real samples in the eyes of the discriminator; adversarial loss encourages the samples generated by the generator to deceive the discriminator, making it difficult to distinguish between generated samples and real samples; Represents content loss, which is used to measure the similarity between the generated sample and the target sample in terms of content. Content loss ensures that the generated sample is not only visually similar to the real sample, but also maintains consistency in content; Represents a hyperparameter used to adjust the content loss in The weights in , which can help strike a balance between generation quality and diversity.

[0070] Expressed as:

[0071] (5);

[0072] Where, Indicates expected value; Represents dual-energy X-ray fusion image.

[0073] Improving the Discriminator in Generative Adversarial Networks The loss function and the discriminator The loss function Expressed as:

[0074] (6);

[0075] (7);

[0076] Where, Represents high-energy X-ray transmission image; Represents a low-energy X-ray transmission image.

[0077] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A copper ore preselection method based on generative adversarial networks and low-rank linear pooling, characterized in that: The steps include: Step S1: obtaining a high-energy X-ray transmission image and a low-energy X-ray transmission image of the copper ore using a dual-energy X-ray imaging device; Step S2: using an improved generative adversarial network to perform modal fusion on the high-energy X-ray transmission image and the low-energy X-ray transmission image of the copper ore to obtain a dual-energy X-ray fusion image of the copper ore; The improved generative adversarial network adds a discriminator on the basis of the original generative adversarial network; Step S3: constructing a copper ore pre-selection model, which includes a feature extraction coding layer, a feature fusion coding layer, and a classification decoding layer; Step S4: inputting the high-energy X-ray transmission image, the low-energy X-ray transmission image and the dual-energy X-ray fusion image of the copper ore into a feature extraction coding layer to obtain high-energy, low-energy and dual-energy feature vectors; Step S5: inputting the high-energy, low-energy and dual-energy feature vectors into the feature fusion coding layer, which fuses the high-energy, low-energy and dual-energy feature vectors using low-rank trilinear pooling to obtain a fused feature vector; Step S6: Input the fused feature vector into the classification decoding layer for classification to obtain the pre-selection result of the copper ore.

2. The copper ore preselection method based on generative adversarial network and low-rank linear pooling according to claim 1 is characterized in that: The specific process of step S5 is: Assume that the high energy, low energy and dual energy eigenvectors are 、 and , solved by low-rank trilinear pooling 、 and The joint feature representation of is: （1）； Where, is the first elements; is the weight parameter for trilinear pooling, controlling 、 、 The interaction of the corresponding elements in for The elements; for The elements; for The elements; For the Bias terms; For the A weight tensor; 、 、 They are 、 、 Dimensions; , , , , is the set of real numbers; The weight tensor Decompose into the product of three 2D matrices: , Indicates that From the dimension Reduce to dimension Matrix of express The transpose of Indicates that From the dimension Reduce to dimension Matrix of express The transpose of Indicates that From the dimension Reduce to dimension Matrix of The decomposed weight tensor Substitute into formula (1) to simplify formula (1): （2）； Where, express The transpose of express The transpose of express The transpose of represents the Hadamard product; express The transpose of represents a vector of all 1s; Reuse Replace the formula (2) ,definition 、 and , obtain the joint feature representation, that is, the fusion feature vector : （3）； Where, 、 and Respectively 、 and The transpose of 、 and Respectively 、 、 The dimension reduction matrix of Represents the bias vector, dimension is ; express The transpose of Represents the weight matrix used for linear transformation.

3. The copper ore preselection method based on generative adversarial networks and low-rank linear pooling according to claim 2, characterized in that: Improved Generative Adversarial Network by Discriminator , Discriminator and generator G; first, the high-energy X-ray transmission image and low-energy X-ray transmission image channel dimensions of the copper ore are superimposed together as the input of the generator G, and the generator outputs the fused image after processing. and the discriminator The high-energy X-ray transmission image and the low-energy X-ray transmission image of the copper ore are received respectively, and the high-energy X-ray transmission image and the low-energy X-ray transmission image of the copper ore are discriminated with the fusion image output by the generator. and the discriminator The discrimination result is input into the generator G, which adjusts the fused image and finally outputs the dual-energy X-ray fusion image of copper ore.

4. The copper ore preselection method based on generative adversarial networks and low-rank linear pooling according to claim 3 is characterized in that: The generator G is composed of a five-layer convolutional neural network, and the first, second, third, fourth and fifth layers of the generator G are all convolutional layers; the first, second, third and fourth convolutional layers of the generator G are successively connected to a batch normalization layer and a Leaky ReLU activation layer, and the fifth convolutional layer is connected to a Tanh activation layer; wherein, the first and second convolutional layers of the generator G use a 5*5 convolution kernel, the third and fourth convolutional layers use a 3*3 convolution kernel, and the fifth convolutional layer uses a 1*1 convolution kernel; the convolutional layer step size of each layer in the generator G is set to 1, and no padding operation is performed in the convolution.

5. The copper ore preselection method based on generative adversarial networks and low-rank linear pooling according to claim 4, characterized in that: The discriminator and the discriminator The structure is the same, also composed of a five-layer convolutional neural network, the discriminator and the discriminator The first, second, third and fourth layers are convolutional layers, and the fifth layer is a linear layer; the discriminator and the discriminator The first, second, third, and fourth convolutional layers are followed by a batch normalization layer and a Leaky ReLU activation layer; the discriminator and the discriminator The first, second, third and fourth convolutional layers use 3*3 convolution kernels and the stride is set to 2.

6. The copper ore pre-selection method based on generative adversarial networks and low-rank linear pooling according to claim 5, characterized in that: The feature extraction encoding layer uses three parallel and identically structured feature extraction encoders DenseNet to extract features from the high-energy X-ray transmission image, low-energy X-ray transmission image and dual-energy X-ray fusion image of copper ore, respectively, to obtain high-energy, low-energy and dual-energy feature vectors.

7. The copper ore pre-selection method based on generative adversarial networks and low-rank linear pooling according to claim 1, characterized in that: Improving the Generator in Generative Adversarial Networks The loss function Expressed as: （4）； Where, It means fighting against loss; Indicates content loss; Represents a hyperparameter used to adjust the content loss in The weight in Expressed as: （5）； Where, Indicates expected value; Represents dual-energy X-ray fusion image.

8. The copper ore pre-selection method based on generative adversarial networks and low-rank linear pooling according to claim 7, characterized in that: Improving the Discriminator in Generative Adversarial Networks The loss function and the discriminator The loss function Expressed as: （6）； （7）； Where, Represents high-energy X-ray transmission image; Represents a low-energy X-ray transmission image.

Images