Multi-quality parameter cooperative detection method and system based on near infrared spectrum

Abstract

The invention relates to a multi-quality parameter cooperative detection method and system based on near infrared spectrum. The method comprises the following steps: providing to-be-detected near infrared spectrum data of a to-be-detected substance, loading the to-be-detected near infrared spectrum data into a constructed quality parameter collaborative detection model, carrying out collaborative detection processing by utilizing the quality parameter collaborative detection model, and generating multi-quality parameter information of the to-be-detected substance, during cooperative detection processing, at least performing feature dimension expansion processing, feature extraction processing and feature fusion prediction processing on the near infrared spectrum data to be detected, and generating multi-quality parameter information of the substance to be detected after the feature fusion prediction processing. According to the invention, cooperative detection of multiple quality parameters can be effectively realized, the precision and reliability of multi-quality parameter detection are improved, and the cost and complexity of multi-quality parameter cooperative detection are reduced.

Description

Technical Field

[0001] The present invention relates to a collaborative detection method and system, in particular to a multi-quality parameter collaborative detection method and system based on near-infrared spectroscopy. Background Art

[0002] Mineral products (such as coal, petroleum, bauxite) are important resources and play an indispensable role in the global economy and industry. The quality of mineral products directly affects the efficiency of various industrial productions and the quality of products. Therefore, accurate detection of the quality of mineral products is of great significance for improving resource utilization rate, optimizing technological processes, and reducing energy consumption. However, when using traditional methods to detect the quality of mineral products, there are often problems such as high cost, long time consumption, high requirements for instruments, and strict requirements for the skills of operators, which limit the wide application of the detection methods in the field of mineral resources.

[0003] Modern detection technologies have shown high efficiency and accuracy in the quality detection of mineral products and have great application potential, such as atomic absorption spectrometry, X-ray fluorescence spectroscopy, and laser-induced breakdown spectroscopy. Although these methods can provide high accuracy, they all have defects such as complex sample preparation and high instrument costs. In addition, most existing detection technologies focus on the detection of single quality parameters, ignoring the internal correlation between different quality parameters and making it difficult to comprehensively and accurately evaluate the overall quality of mineral products.

[0004] To solve these problems, near-infrared spectroscopy technology has gradually become an emerging technology for the quality detection of mineral products. Compared with traditional methods, near-infrared spectroscopy technology has the advantages of low cost, no need for complex sample preparation, fast detection speed, etc., and can obtain the quality information of mineral products in real time without damaging the samples. However, how to effectively achieve the collaborative detection of multi-quality parameters is still a technical problem that needs to be solved urgently at present. Summary of the Invention

[0005] The object of the present invention is to overcome the deficiencies in the prior art and provide a multi-quality parameter collaborative detection method and system based on near-infrared spectroscopy, which can effectively achieve the collaborative detection of multi-quality parameters, improve the accuracy and reliability of multi-quality parameter detection, and reduce the cost and complexity of multi-quality parameter collaborative detection.

[0006] According to the technical solution provided by the present invention, a multi-quality parameter collaborative detection method based on near-infrared spectroscopy, the method includes: Providing the near-infrared spectrum data of the substance to be detected, and loading the near-infrared spectrum data of the substance to be detected into the constructed quality parameter collaborative detection model, so as to perform collaborative detection processing by using the quality parameter collaborative detection model and generate the multi-quality parameter information of the substance to be detected, wherein, When performing collaborative detection processing, at least feature dimension expansion processing, feature extraction processing, and feature fusion prediction processing are performed on the near-infrared spectral data to be detected, and multi-quality parameter information of the substance to be detected is generated after the feature fusion prediction processing.

[0007] The quality parameter collaborative detection model includes an embedding layer, a feature extraction module, and a customized gated network connected in sequence, where The embedding layer performs feature dimension expansion processing on the near-infrared spectral data to be detected and generates the spectral data after dimension expansion to be detected; The feature extraction module performs feature extraction processing on the spectral data after dimension expansion to be detected and generates multi-quality shared features to be detected; The customized gated network performs feature fusion prediction processing on the multi-quality shared features to be detected and the near-infrared spectral data to be detected to generate multi-quality parameter information, where When performing feature fusion prediction processing, first perform feature linear activation processing on the multi-quality shared features to be detected to generate a shared expert feature to be detected and several specific expert features to be detected. Thereafter, each specific expert feature to be detected is respectively subjected to gated weighted fusion with the shared expert feature to be detected and the near-infrared spectral data to be detected, and after regression prediction, the corresponding quality parameter prediction value is generated, where the number of specific expert features to be detected is consistent with the number of quality parameter prediction values in the multi-quality parameter information; Based on all the quality parameter prediction values, multi-quality parameter information of the substance to be detected is formed.

[0008] The feature extraction module includes several feature extraction sub-modules connected in sequence, where For any feature extraction sub-module, it includes a depthwise separable convolutional network and a spatial dimension feature weighting module connected in sequence, and configures the depthwise separable convolutional network and the spatial dimension feature weighting module to form a residual connection; When performing feature extraction processing, for any feature extraction sub-module, first use the depthwise separable convolutional network to perform single-channel long-distance feature extraction processing and channel fusion processing on the basic data for feature extraction to be detected in sequence, and generate the feature after extraction and fusion to be detected; Use the spatial dimension feature weighting module to perform spatial dimension feature weighting processing on the feature after extraction and fusion to be detected to generate the spatially dimensionally weighted feature to be detected; thereafter, the spatially dimensionally weighted feature to be detected and the basic data for feature extraction to be detected are processed through a residual connection, and the sub-module data feature to be detected is generated.

[0009] The depthwise separable convolutional network includes a depth convolutional layer using a large-size convolutional kernel and a batch normalization layer, a first point convolutional layer, a GeLU activation function, and a second point convolutional layer connected to the depth convolutional layer in sequence, where The depth convolution layer with a large-size convolution kernel performs single-channel long-distance feature extraction processing on the basic data to be detected for feature extraction; The inverted bottleneck structure is formed by the first point convolution layer and the second point convolution layer, and the formed inverted bottleneck structure is used to perform channel fusion processing. During the channel fusion processing, the channel dimension is first expanded by r times through the first point convolution layer, and then, the expanded channel dimension is restored through the second point convolution layer.

[0010] The spatial dimension feature weighting module includes a spatial dimension convolution block, a spatial dimension Sigmoid layer, and a spatial dimension multiplier, where When performing spatial dimension feature weighting processing, for the feature to be detected after extraction and fusion, first, a convolution operation is performed using the spatial dimension convolution block to generate a single-channel feature sequence to be weighted. Then, the single-channel feature sequence to be weighted is converted into a probability distribution using the spatial dimension Sigmoid layer to form single-channel probability distribution information, where the feature to be detected after extraction and fusion is generated by the depth point convolution network within the same feature extraction sub-module; The single-channel probability distribution information is multiplied by the feature to be detected after extraction and fusion using the spatial dimension multiplier to generate the feature to be detected with spatial dimension weighting.

[0011] The customized gating network includes a customized gating module and a tower network adaptively connected to the customized gating module, where The customized gating module includes an expert unit and a gating unit, where the expert unit includes a shared expert module and several specific expert modules, The gating unit includes several gating modules. The number of specific expert modules is the same as the number of gating modules, and the number of gating modules is not less than the number of quality parameter prediction values within the multi-quality parameter information; The tower network includes several tower modules for regression prediction, and the tower modules are connected in a one-to-one correspondence with the gating modules; During the feature fusion prediction processing, the expert unit performs feature linear activation processing on the multi-quality shared feature to be detected, generates the shared expert feature to be detected through the shared expert module, and generates the corresponding specific expert feature to be detected through a specific expert module; The shared expert feature to be detected, the near-infrared spectrum data to be detected, and the specific expert feature to be detected are respectively loaded into the corresponding gating modules to perform gating weighted fusion using the gating modules and generate a sequence of gating weighted features to be detected, and the sequence of gating weighted features to be detected is loaded into the corresponding connected tower modules; The tower module performs regression prediction on the received sequence of gating weighted features to be detected to generate a corresponding quality parameter prediction value after the regression prediction.

[0012] The gating module includes a data to be detected processing unit, a first gating multiplier, a second gating multiplier, and a gating adder, where During gating weighted fusion, the data to be detected processing unit linearly activates the data to be detected near-infrared spectral data at least, so as to generate an activated feature to be detected after the data linear activation processing; The first gating multiplier is used to multiply the shared expert feature to be detected by the activated feature to be detected to generate a shared activated feature to be detected; meanwhile, the second gating multiplier is used to multiply the specific expert feature to be detected by the activated feature to be detected to generate a specific activated feature to be detected; The gating adder is used to perform an addition operation on the shared activated feature to be detected and the specific activated feature to be detected to generate a gating weighted feature sequence to be detected.

[0013] When constructing the quality parameter collaborative detection model, it includes: Constructing a quality parameter collaborative detection basic model and a basic model training data set for training the quality parameter collaborative detection basic model, where The basic model training data set includes a number of training samples. Each training sample includes a training near-infrared spectral data and a number of quality parameter labels. The number of quality parameter labels in the training sample is consistent with the number of quality parameter predicted values in the multi-quality parameter information, and the type of the quality parameter label corresponds one-to-one with the type of the quality parameter predicted value; Configuring the model training conditions for the quality parameter collaborative detection basic model until the quality parameter collaborative detection basic model is trained to reach the target state. After that, the quality parameter collaborative detection basic model trained to reach the target state is configured as the quality parameter collaborative detection model.

[0014] The configured model training conditions include a training loss function, and the training loss function includes:

[0015] Where is the training loss value, is the regression training loss, is the orthogonal training loss, is the number of training samples, is the number of tasks during collaborative detection, is the quality parameter label of the th training sample corresponding to the kth task, is the quality parameter predicted value of the th training sample corresponding to the kth task, is the batch number during model training, is the training shared expert feature matrix of the is the transposed matrix for training the shared expert feature matrix, is the training specific expert feature matrix corresponding to the k-th task for the -th batch training sample, and

[0016] A multi-quality parameter collaborative detection system based on near-infrared spectroscopy includes a multi-quality parameter collaborative detection device, and deploys the above-mentioned quality parameter collaborative detection model inside the multi-quality parameter collaborative detection device, where For the near-infrared spectroscopy data to be detected of any substance to be detected, the multi-quality parameter collaborative detection device performs collaborative detection processing by using the above-mentioned method to obtain the multi-quality parameter information of the substance to be detected after the collaborative detection processing.

[0017] Advantages of the present invention: The near-infrared spectroscopy data to be detected of the substance to be detected is collected and acquired. Thereafter, the quality parameter collaborative detection model is used for collaborative detection processing to obtain the multi-quality parameter information of the substance to be detected. It can be seen therefrom that during the collaborative detection processing, due to the use of the near-infrared spectroscopy data collected by near-infrared spectroscopy, compared with the existing detection methods, the cost and complexity of the multi-quality parameter collaborative detection can be effectively reduced; through the collaborative detection processing by the quality parameter collaborative detection model, the multi-quality parameter information of the substance to be detected can be obtained, so that the collaborative detection of the multi-quality parameters can be effectively realized, and the accuracy and reliability of the multi-quality parameter detection can be improved.

[0018] Since orthogonal constraints are adopted in model training, the correlation between the multi-quality parameters of the substance to be detected is fully considered. Therefore, when using the quality parameter collaborative detection model for collaborative detection processing, while maintaining high accuracy, the correlation between the multi-quality parameters of the substance to be detected can be fully considered, and the accuracy and reliability of generating the multi-quality parameter information can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flowchart of an embodiment of the multi-quality parameter collaborative detection method of the present invention.

[0020] Figure 2 is a schematic diagram of an embodiment of the existing near-infrared spectroscopy data.

[0021] Figure 3 is a schematic diagram of an embodiment of the quality parameter collaborative detection model of the present invention.

[0022] Figure 4 is a structural block diagram of an embodiment of the depth point convolution network of the present invention.

[0023] Figure 5 is a structural block diagram of an embodiment of the spatial dimension feature weighting module of the present invention.

[0024] Figure 6 It is a schematic diagram of an embodiment during the model training of the present invention.

[0025] Figure 7 It is a structural block diagram of an embodiment of the gating structure of the present invention.

[0026] Figure 8 It is a structural block diagram of an embodiment of the shared expert module of the present invention. Detailed implementation manners

[0027] The present invention will be further described below in conjunction with specific drawings and embodiments.

[0028] In order to effectively achieve the collaborative detection of multiple quality parameters, improve the accuracy and reliability of the detection of multiple quality parameters, and reduce the cost and complexity of the collaborative detection of multiple quality parameters, the present invention provides a method for collaborative detection of multiple quality parameters based on near-infrared spectroscopy. Specifically, the method for collaborative detection of multiple quality parameters includes: Providing the near-infrared spectral data to be detected of the substance to be detected, and loading the near-infrared spectral data to be detected into the constructed quality parameter collaborative detection model, so as to perform collaborative detection processing by using the quality parameter collaborative detection model and generate the multi-quality parameter information of the substance to be detected, wherein When performing collaborative detection processing, at least feature dimension expansion processing, feature extraction processing, and feature fusion prediction processing are performed on the near-infrared spectral data to be detected, and the multi-quality parameter information of the substance to be detected is generated after the feature fusion prediction processing.

[0029] It should be noted that the collaborative detection of multiple quality parameters of the present invention specifically refers to the detection of multiple quality parameters that can be realized simultaneously, that is, multiple quality parameters of the same substance can be obtained simultaneously. Figure 1 FIG. shows a flowchart of an embodiment of the collaborative detection of multiple quality parameters of the present invention. It can be seen from the figure that when performing the collaborative detection of multiple quality parameters, the near-infrared spectral data to be detected of the substance to be detected should be provided. It can be understood that the substance to be detected should be a substance suitable for near-infrared spectral acquisition. For example, the substance to be detected can be bauxite, coal, petroleum and other substances. The type of the substance to be detected can be selected according to needs and is not limited here.

[0030] After determining the type of the substance to be detected, the corresponding near-infrared spectrum data of the substance to be detected can be obtained by collecting the near-infrared spectrum of the substance to be detected. For example, the MicroNIR Pro handheld near-infrared spectrometer produced by VIAVI can be used to collect the near-infrared spectrum of the substance to be detected. The wavelength range for collection can be 908nm - 1676nm, the resolution is 6.24nm, and the number of wavelength points is 125. Of course, when collecting the near-infrared spectrum of the substance to be detected, the corresponding collection conditions can be configured according to the type of the substance to be detected, specifically based on the ability to effectively collect the near-infrared spectrum data of the substance to be detected.

[0031] Figure 2 An embodiment of the near-infrared spectrum data is shown. As can be seen from the figure, the near-infrared spectrum of the substance to be detected can be characterized as a waveform diagram. Figure 2 In this figure, the abscissa is the wavelength and the ordinate is the absorbance. Therefore, the characteristic dimension of the near-infrared spectrum data of the substance to be detected can be 1×L, where L is the number of characteristic points of the near-infrared spectrum data of the substance to be detected, and "1" is the channel dimension. It should be noted that when the above-mentioned 125 wavelength points are used, the number of characteristic points L of the near-infrared spectrum data of the substance to be detected should be 125. For other situations, reference can be made to the description here, and no further listing will be made here.

[0032] From Figure 1 it can be seen that when performing co-detection of multiple quality parameters, the near-infrared spectrum data of the substance to be detected should be loaded into the co-detection model of quality parameters to perform co-detection processing using the co-detection model of quality parameters, and after the co-detection processing, the multi-quality parameter information of the substance to be detected can be generated. Among them, the co-detection processing performed by the co-detection model of quality parameters generally should include feature dimension expansion processing, feature extraction processing, and feature fusion prediction processing. After the feature fusion prediction processing, the multi-quality parameter information of the substance to be detected is generated.

[0033] From the above description, it can be seen that the multi-quality parameter information should include multiple quality parameter prediction values. The number of quality parameter prediction values can generally be selected according to actual needs. The type of quality parameter prediction values should be related to the type of the substance to be detected. For example, when the substance to be detected is bauxite, the multi-quality parameter information can include the predicted value of the content of Al 2 O 3 , the predicted value of the content of SiO 2 , the predicted value of the content of Fe 2 O 3 , that is, the content of Al 2 O 3 , the content of SiO 2 , the content of Fe 2 O 3Quality parameters of bauxite; when the substance to be detected is coal, the quality parameters of coal can be moisture, ash, volatile matter, and / or calorific value. At this time, the multi-quality parameter information obtained can be the predicted value of moisture content, the predicted value of ash content, the predicted value of volatile matter, and / or the predicted value of calorific value; when the substance to be detected is petroleum, the quality parameters of petroleum can be density, sulfur content, and / or acid value. At this time, the multi-quality parameter information obtained can be the predicted value of density, the predicted value of sulfur content, or the predicted value of acid value. When the substance to be detected is of other types, corresponding multi-quality parameter information can be obtained, which will not be elaborated one by one here.

[0034] As can be seen from the above description, when the present invention performs collaborative detection of multi-quality parameters, the near-infrared spectral data of the substance to be detected collected by near-infrared spectroscopy can effectively reduce the cost and complexity of the collaborative detection of multi-quality parameters compared with the existing detection methods; through the quality parameter collaborative detection model for collaborative detection processing, multi-quality parameter information of the substance to be detected can be obtained, so as to effectively realize the collaborative detection of multi-quality parameters and improve the accuracy and reliability of the multi-quality parameter detection.

[0035] In one embodiment of the present invention, the quality parameter collaborative detection model includes an embedding layer, a feature extraction module, and a customized gated network connected in sequence, where The embedding layer performs feature dimension expansion processing on the near-infrared spectral data of the substance to be detected and generates the spectral data of the substance to be detected after dimension expansion; The feature extraction module performs feature extraction processing on the spectral data of the substance to be detected after dimension expansion and generates the multi-quality shared features of the substance to be detected; The customized gated network performs feature fusion prediction processing on the multi-quality shared features of the substance to be detected and the near-infrared spectral data of the substance to be detected to generate multi-quality parameter information, where When performing the feature fusion prediction processing, first perform feature linear activation processing on the multi-quality shared features of the substance to be detected to generate a shared expert feature of the substance to be detected and several specific expert features of the substance to be detected. Thereafter, each specific expert feature of the substance to be detected is respectively subjected to gated weighted fusion with the shared expert feature of the substance to be detected and the near-infrared spectral data of the substance to be detected, and after regression prediction, the corresponding quality parameter prediction value is generated, where the number of specific expert features of the substance to be detected is consistent with the number of quality parameter prediction values in the multi-quality parameter information; Based on all the quality parameter prediction values, multi-quality parameter information of the substance to be detected is formed.

[0036] Figure 3An embodiment of the quality parameter detection model of the present invention is shown. As can be seen from the figure, the quality parameter collaborative detection model may include an embedding layer, a feature extraction module, and a customized gating network. The part where the customized gating module is located forms the customized gating network. The embedding layer is connected to the customized gating module through the feature extraction module. Specifically, the feature dimension expansion process can be performed through the embedding layer, the feature extraction process can be performed through the feature extraction module, and the feature fusion prediction process can be performed through the customized gating network. The specific situations of the embedding layer, the feature extraction module, and the customized gating network will be specifically described below.

[0037] It can be understood that when loading the near-infrared spectral data to be detected into the quality parameter collaborative detection model, it specifically refers to loading the near-infrared spectral data to be detected into the embedding layer. As described above, the infrared spectral data to be detected is one-dimensional data. In order to meet the subsequent feature extraction process and feature fusion prediction process, the feature dimension expansion process should be performed through the embedding layer, and the spectral data after dimension expansion to be detected is generated. Among them, the feature dimension of the spectral data after dimension expansion to be detected can be C×L, where C is the channel dimension of the spectral data after dimension expansion to be detected. Generally, C can take 128. Of course, C can also take other values, which can be specifically selected according to needs. The embedding layer can adopt the existing common form, and the method of feature dimension expansion is consistent with the prior art, which will not be elaborated here.

[0038] After the embedding layer generates the spectral data after dimension expansion to be detected, the feature extraction module is configured to perform feature extraction processing on the spectral data after dimension expansion to be detected, so as to obtain the multi-quality shared features to be detected after the feature extraction process, and transmit the generated multi-quality shared features to be detected to the customized gating network.

[0039] After obtaining the multi-quality shared features to be detected, the customized gating network is configured to perform feature fusion prediction processing. Among them, when performing feature fusion prediction processing, first perform feature linear activation processing on the multi-quality shared features to be detected, so as to generate a shared expert feature to be detected and several specific expert features to be detected after the feature linear activation processing. Thereafter, each specific expert feature to be detected is respectively subjected to gated weighted fusion with the shared expert feature to be detected and the near-infrared spectral data to be detected, and the corresponding quality parameter prediction value is generated after regression prediction. Therefore, based on multiple specific expert features to be detected, multiple quality parameter prediction values can be generated after gated weighted fusion and regression prediction. Thereafter, based on all the quality parameter prediction values, the multi-quality parameter information of the substance to be detected can be formed, and thus the above-mentioned collaborative detection process is completed.

[0040] In order to be able to extract the above-mentioned multi-quality shared features to be detected, the present invention provides a feature extraction module. Specifically, the feature extraction module includes a plurality of sequentially connected feature extraction sub-modules, where For any feature extraction sub-module, it includes a depthwise point convolution network and a spatial dimension feature weighting module connected in series in sequence, and configures the depthwise point convolution network and the spatial dimension feature weighting module to form a residual connection; When performing feature extraction processing, for any feature extraction sub-module, first use the depthwise point convolution network to perform single-channel long-distance feature extraction processing and channel fusion processing on the basic data of the feature to be detected in sequence, and generate the fused feature after extraction of the feature to be detected; Use the spatial dimension feature weighting module to perform spatial dimension feature weighting processing on the fused feature after extraction of the feature to be detected to generate the weighted feature of the spatial dimension of the feature to be detected; thereafter, perform residual connection processing on the weighted feature of the spatial dimension of the feature to be detected and the basic data of the feature to be detected, and generate the data feature of the sub-module to be detected.

[0041] Specifically, the feature extraction module includes a number of feature extraction sub-modules connected in series in sequence. At the same time, each feature extraction sub-module adopts a residual connection. Figure 3 In this case, "×7" in the feature extraction module specifically means that the feature extraction module includes 7 feature extraction sub-modules connected in series in sequence. The number of feature extraction sub-modules can be selected according to needs, so as to meet the specific feature extraction requirements.

[0042] In addition, Figure 3 An embodiment of the feature extraction sub-module is also shown in this case. It can be seen from the figure that each feature extraction sub-module may include a depthwise point convolution network and a spatial dimension feature weighting module connected in series in sequence. When the feature extraction sub-module adopts a residual connection, the input end of the depthwise point convolution network is correspondingly connected to the output end of the spatial dimension feature weighting module. Generally, the adaptation connection between the input end of the depthwise point convolution network and the output end of the spatial dimension feature weighting module can be realized through a feature extraction adder, that is, the residual connection processing can be realized through the feature extraction adder. It should be noted that the feature extraction adder is not shown in Figure 3 this case.

[0043] It should be noted that when each feature extraction sub-module adopts a residual connection, the ability of each feature extraction sub-module to capture detailed information can be improved, ensuring that each feature extraction sub-module can retain and strengthen the information obtained, thereby effectively promoting the feature fusion in the deep network in the feature extraction module, enabling the quality parameter collaborative detection model to maintain sensitivity to the near-infrared spectrum data to be detected when capturing the key information in the near-infrared spectrum data to be detected, thereby improving the feature extraction ability.

[0044] Within the feature extraction module, the feature extraction sub-module at the head of the concatenation should be connected to the input end of the embedding layer to receive the corresponding spectral data after dimension expansion of the dimension to be detected. The feature extraction sub-module at the tail of the concatenation should serve as the output layer of the current feature extraction module, that is, the corresponding specific expert features to be detected or the specific expert features to be detected can be output through the feature extraction sub-module at the tail of the concatenation. In addition, along the concatenation direction of the feature extraction sub-modules, the output end of the feature extraction adder of the previous feature extraction sub-module is connected to the input end of the depthwise point convolution network within the next feature extraction sub-module. Here, the concatenation direction of the feature extraction sub-modules specifically refers to the direction from the head of the concatenation to the tail of the concatenation.

[0045] When performing feature extraction processing, for the feature extraction sub-module within any feature extraction module, first use the depthwise point convolution network to sequentially perform single-channel long-distance feature extraction processing and channel fusion processing on the basic data for feature extraction to be detected, and generate the fused features after extraction to be detected; thereafter, use the spatial dimension feature weighting module to perform spatial dimension feature weighting processing on the fused features after extraction to be detected to generate the weighted features in the spatial dimension to be detected. When using residual connection, load the basic data for feature extraction to be detected and the weighted features in the spatial dimension to be detected into the feature extraction adder, and after performing addition operations through the feature extraction adder, the data features of the sub-module to be detected can be generated. Therefore, the above-mentioned residual connection processing is the addition operation processing performed by the feature extraction adder.

[0046] As can be seen from the above description, within the feature extraction module, for the feature extraction sub-module at the head of the concatenation, the basic data for feature extraction to be detected should be the spectral data after dimension expansion of the dimension to be detected generated by the embedding layer; the remaining basic data for feature extraction to be detected should be the data features of the sub-module to be detected output by the previous feature extraction sub-module. Based on the data features of the sub-module to be detected generated by the feature extraction sub-module at the tail of the concatenation, the corresponding multi-quality shared features to be detected can be formed. After forming the multi-quality shared features to be detected, the above-mentioned feature extraction processing is realized.

[0047] In an embodiment of the present invention, the depthwise point convolution network includes a depth convolution layer using a large-size convolution kernel, and a batch normalization layer, a first point convolution layer, a GeLU activation function, and a second point convolution layer sequentially connected to the depth convolution layer, where Perform single-channel long-distance feature extraction processing on the basic data for feature extraction to be detected through the depth convolution layer with a large-size convolution kernel; Form an inverted bottleneck structure through the first point convolution layer and the second point convolution layer, and use the formed inverted bottleneck structure to perform channel fusion processing. During channel fusion processing, first expand the channel dimension by r times through the first point convolution layer, and then restore the expanded channel dimension through the second point convolution layer.

[0048] Figure 4 Figure 1 shows a schematic diagram of an embodiment of a depthwise point convolution network. Depthwise convolution can be performed using depthwise convolution, which is a lightweight convolution operation that independently convolves each input channel to effectively extract feature information for each channel. In a specific implementation, the depthwise convolution layer uses a large-sized convolution kernel, which can not only effectively capture long-range dependencies but also enhance the feature extraction ability. The large convolution kernel can cover a larger spectral range in a single operation and effectively capture the relevant features between multiple bands. Figure 4 Figure 2 shows an embodiment where the size of the convolution kernel in the depthwise convolution layer is 51 (k = 51). Therefore, the large-sized convolution kernel here specifically refers to a relatively large size of the convolution kernel. When the size of the convolution kernel in the depthwise convolution layer is 51, it can accurately capture the peak regions in the near-infrared spectral data and concentrate the activation at these key bands. In addition, the number of convolution kernels in the depthwise convolution layer can be selected according to needs. Compared with the currently popular Transformer model, configuring the depthwise convolution layer with a large-sized convolution kernel has the advantages of simple design, fewer parameters, and helping to reduce the risk of overfitting.

[0049] As can be seen from the above description, the depthwise convolution layer is used to perform single-channel long-range feature extraction processing on the basic data of the feature to be detected, so as to generate the single-channel long-range related features to be detected after the single-channel long-range feature extraction processing. Among them, the channel dimension of the single-channel long-range related features to be detected is still C.

[0050] From Figure 4 After obtaining the single-channel long-range related features to be detected, batch normalization processing is performed through the batch normalization layer, and the features after batch normalization to be detected are generated after the batch normalization processing. The channel dimension of the batch normalization layer to be detected is C. Figure 4 In [reference], the connection relationship between the batch normalization layer and the depthwise convolution layer is not shown. Figure 4 BN in [reference] represents the batch normalization layer. It can be understood that when the batch normalization layer is set in the depthwise point convolution network, it can effectively stabilize the learning process of training and generating the quality parameter collaborative detection basic model, improve the convergence speed of the quality parameter collaborative detection basic model. For the situation of the quality parameter collaborative detection basic model, please refer to the corresponding description below.

[0051] The point convolution operation can be performed on the features after batch normalization to be detected through the first point convolution layer, so as to generate the first point convolution features to be detected after the point convolution operation. Among them, the channel dimension of the first point convolution features to be detected is r × C. When the channel dimension of the first point convolution features to be detected is expanded to r × C, the expression ability of the features can be enhanced. Generally, r can take the value of 7, and the corresponding first point convolution layer can be selected according to the expansion multiple r of the channel dimension.

[0052] After obtaining the feature after the first point convolution to be inspected, activation processing is performed through the GeLU activation function to enhance the learning ability of non-linear features. It should be noted that Figure 4 the connection between the GeLU activation function and the first point convolution layer and the second point convolution layer is not shown in Figure 4 The GeLU in

[0053] is the GeLU activation function here. It can be understood that after the activation processing by the GeLU activation function, the feature after the GeLU activation function to be inspected can be obtained, and the channel dimension of the feature after the GeLU activation function to be inspected is still r×C.

[0054] As can be seen from the above description, although the depth convolution layer can independently extract the information of each channel, it cannot achieve the information fusion between channels. The second point convolution layer integrates the information of different channels through a convolution kernel with a size of 1 (k = 1), so as to achieve the effective fusion of information of different channels and generate a more representative feature expression. Specifically, the first point convolution layer also uses a convolution kernel with a size of 1 (k = 1). In addition, according to the above situation of the channel dimension, the number of corresponding convolution kernels in the first point convolution layer and the second point convolution layer can be determined.

[0055] It should be noted that the channel expansion can be achieved through the first point convolution layer, so as to capture more fine-grained information. The channel compression can be achieved through the second point convolution layer, and the channel compression selectively retains important features by learning adaptive weights to avoid the loss of useful information. It can be seen from this that an inverted bottleneck structure is formed by the first point convolution layer and the second point convolution layer. This inverted bottleneck structure ensures high-quality feature extraction while maintaining high computational efficiency, and achieves high-quality feature fusion effects at low computational costs.

[0056] In an embodiment of the present invention, the spatial dimension feature weighting module includes a spatial dimension convolution block, a spatial dimension Sigmoid layer, and a spatial dimension multiplier, where when performing spatial dimension feature weighting processing, for the feature after extraction and fusion to be inspected, first, a convolution operation is performed using the spatial dimension convolution block to generate a single-channel feature sequence to be weighted. Thereafter, the single-channel feature sequence to be weighted is converted into a probability distribution using the spatial dimension Sigmoid layer to form single-channel probability distribution information, where the feature after extraction and fusion to be inspected is generated by the depth point convolution network within the same feature extraction sub-module; Multiply the single-channel probability distribution information by the feature after fusion of the feature to be detected and extraction through a spatial dimension multiplier to generate a weighted feature of the spatial dimension to be detected.

[0057] Figure 5 An embodiment diagram of the spatial dimension feature weighting module is shown. It can be seen from the figure that the spatial dimension feature weighting module may include a spatial dimension convolution block, a spatial dimension Sigmoid layer, and a spatial dimension multiplier. Among them, the number of convolution kernels in the spatial dimension convolution block is 1. At this time, the channel dimension of the generated single-channel feature sequence to be weighted is 1. Use the spatial dimension Sigmoid layer to convert the single-channel feature sequence to be weighted into a probability distribution to form single-channel probability distribution information and weight the importance of each band.

[0058] Figure 5 In it, E0 is the feature after fusion of the feature to be detected and extraction, E1 is the single-channel feature sequence to be weighted, E2 is the single-channel probability distribution information, E3 is the weighted feature of the spatial dimension to be detected, and the feature dimension of the weighted feature of the spatial dimension to be detected is C×L.

[0059] It can be seen from the above description that the spatial dimension feature weighting module can enhance the response to the significant features in the near-infrared spectrum data to be detected through the learned weights, especially the peak and trough regions closely related to the quality parameters. This not only maintains the sensitivity of the quality parameter collaborative detection model to key features but also improves the prediction accuracy of the regression task.

[0060] In an embodiment of the present invention, the customized gating network includes a customized gating module and a tower network adaptively connected to the customized gating module, where The customized gating module includes an expert unit and a gating unit. The expert unit includes a shared expert module and several specific expert modules. The gating unit includes several gating modules. The number of specific expert modules is the same as the number of gating modules, and the number of gating modules is not less than the number of quality parameter prediction values in the multi-quality parameter information; The tower network includes several tower modules for regression prediction, and the tower modules are connected in one-to-one correspondence with the gating modules; During feature fusion prediction processing, use the expert unit to perform feature linear activation processing on the multi-quality shared feature to be detected, so as to generate a shared expert feature to be detected through the shared expert module and a corresponding specific expert feature to be detected through a specific expert module; Load the shared expert feature to be detected, the near-infrared spectrum data to be detected, and the specific expert feature to be detected into the corresponding gating modules respectively, so as to perform gating weighted fusion by the gating modules and generate a gating weighted feature sequence to be detected, and load the gating weighted feature sequence to be detected into the corresponding connected tower modules; The tower module performs regression prediction on the received gated weighted feature sequence to be inspected, so as to generate a corresponding predicted value of the quality parameter after the regression prediction.

[0061] Figure 3 An embodiment of the customized gating network is shown. It can be seen from the figure that the customized gating network may include a customized gating module and a tower network. Among them, the customized gating module may include an expert unit and a gating unit. The expert unit can perform the above-mentioned feature linear activation processing, and the gating unit can perform the above-mentioned gated weighted fusion. In addition, the customized gating module is connected to the tower network, so that the tower network can perform the above-mentioned regression prediction.

[0062] Specifically, the expert unit generally should include a shared expert module and several specific expert modules. The number of specific expert modules should be no less than the number of predicted values of the quality parameter. Preferably, the number of specific expert modules should be consistent with the number of predicted values of the quality parameter. When performing the feature fusion prediction process, the multi-quality shared features to be inspected should be loaded into the shared expert module and all specific expert modules at the same time. After that, the shared expert module can generate the shared expert feature to be inspected, and a specific expert module can generate a specific expert feature to be inspected. That is, the specific expert features to be inspected correspond one-to-one with the specific expert modules, and the specific expert features to be inspected also correspond one-to-one with the predicted values of the quality parameter.

[0063] It can be seen from the above description that when the expert unit performs the feature linear activation processing, it specifically includes the linear activation processing performed by the shared expert module and the linear activation processing performed by each specific expert module.

[0064] It should be noted that Figure 3 An embodiment in which the number of predicted values of the quality parameter is 3 is shown. At this time, the expert unit should include at least 3 specific expert modules. Figure 3 In this case, the specific expert module A, the specific expert module B, and the specific expert module C are respectively the 3 corresponding specific expert modules in the expert unit. When the multi-quality shared features to be inspected are in other situations, reference can be made to the description here, and no further examples will be given one by one.

[0065] The gating unit may include multiple gating modules. Generally, the number of gating modules should be no less than the number of predicted values of the quality parameter, and the number of gating modules should be at least consistent with the number of specific expert modules. That is, the number of gating modules should be no less than the number of specific expert modules, so that the specific expert modules and the gating modules can be connected in a one-to-one correspondence. It can be seen from the above description that when the number of predicted values of the quality parameter is 3, the number of gating modules in the gating unit should be no less than 3. Figure 3In an embodiment where the gating unit includes three gating modules, as shown in the figure, the three gating modules are respectively the gating module G1, the gating module G2, and the gating module G3. In addition, Figure 3 The situation of three predicted quality parameter values is also shown. The three predicted quality parameter values are respectively the predicted quality parameter value Pz1, the predicted quality parameter value Pz2, and the predicted quality parameter value Pz3. Among them, the predicted quality parameter value Pz1 corresponds to the gating module G1, the predicted quality parameter value Pz2 corresponds to the gating module G2, and the predicted quality parameter value Pz3 corresponds to the gating module G3.

[0066] In specific implementation, the tower network generally should include several tower modules. The number of tower modules should be no less than the number of gating modules. Preferably, the number of tower modules is configured to be the same as the number of gating modules. Generally, a one-to-one correspondence connection should be adopted between the gating modules and the tower modules. When Figure 3 it is shown that there are three gating modules, then the number of tower modules should also be 3. Figure 3 In the figure, the three tower modules are respectively the tower module TowerA, the tower module TowerB, and the tower module TowerC. Among them, the tower module TowerA is correspondingly connected to the gating module G1, the tower module TowerB is correspondingly connected to the gating module G2, and the tower module TowerC is correspondingly connected to the gating module G3.

[0067] As can be seen from the above description, for each multi-quality shared feature to be detected, after the expert module performs feature linear activation processing, a shared expert feature to be detected and several specific expert features to be detected can be obtained. Thereafter, for each specific expert feature to be detected, the specific expert feature to be detected should be subjected to gated weighted fusion with the shared expert feature to be detected and the near-infrared spectral data to be detected. Among them, when performing gated weighted fusion, a specific expert feature to be detected, the shared expert feature to be detected, and the near-infrared spectral data to be detected should be loaded into a corresponding gating module, and then the gating module is used to perform gated weighted fusion.

[0068] Figure 3An example of gated weighted fusion implemented by AVIC is shown in the figure. In the figure, the to-be-tested shared expert features generated by the shared expert module, the to-be-tested near-infrared spectral data, and the to-be-tested specific expert features generated by the specific expert module ExpertsA are simultaneously loaded into the gating module G1. At the same time, the to-be-tested shared expert features generated by the shared expert module, the to-be-tested near-infrared spectral data, and the to-be-tested specific expert features generated by the specific expert module ExpertsB are simultaneously loaded into the gating module G2, and the to-be-tested shared expert features generated by the shared expert module, the to-be-tested near-infrared spectral data, and the to-be-tested specific expert features generated by the specific expert module ExpertsC are simultaneously loaded into the gating module G3. After that, the gating module G1, the gating module G2, and the gating module G3 are used to perform corresponding gated weighted fusion respectively.

[0069] Figure 3 In the figure, after the gating module G1 performs the gated weighted fusion process, it can generate the to-be-tested gated weighted feature sequence A and load the to-be-tested gated weighted feature sequence A into the tower module TowerA. At the same time, after the gating module G2 performs the gated weighted fusion process, it can generate the corresponding to-be-tested gated weighted feature sequence B and load the to-be-tested gated weighted feature sequence B into the tower module TowerB. After the gating module G3 performs the gated weighted fusion process, it can generate the corresponding to-be-tested gated weighted feature sequence C and load the to-be-tested gated weighted feature sequence C into the tower module TowerC.

[0070] Specifically, the tower module can be composed of multiple cascaded regression prediction heads. The regression prediction head can adopt the existing common structural form. The way of cascading the regression prediction heads to form the tower module can be consistent with the prior art. When the tower module is formed by the regression prediction heads, the tower module performs regression prediction on the received to-be-tested gated weighted feature sequence to generate a corresponding quality parameter prediction value after the regression prediction. For example, the tower module TowerA performs regression prediction on the to-be-tested gated weighted feature sequence A and obtains the quality parameter prediction value Pz1 after the regression prediction. At the same time, the tower module TowerB performs regression prediction on the to-be-tested gated weighted feature sequence B and obtains the quality parameter prediction value Pz2 after the regression prediction. The tower module TowerC performs regression prediction on the to-be-tested gated weighted feature sequence C and obtains the quality parameter prediction value Pz3 after the regression prediction.

[0071] It should be noted that the shared expert module and the specific expert module can adopt the same structural form. Of course, the shared expert module and the specific expert module can also adopt different forms. Preferably, the shared expert module and the specific expert module adopt the same form. When the shared expert module and the specific expert module adopt the same form, Figure 8An embodiment of the shared expert module is shown. In the figure, the shared expert module includes a first linear layer of the expert module, a ReLU activation function layer, and a second linear layer of the expert module, which are connected in sequence. Specifically, when performing the above linear activation process, it specifically means using the first linear layer of the expert module, the ReLU activation function layer, and the second linear layer of the expert module to process the multi-quality shared features to be detected in sequence.

[0072] In specific implementation, the shared expert module can also adopt other implementation forms, which can be specifically selected according to needs. According to the forms adopted by the shared expert module and the specific expert module, the corresponding linear activation processing method can be determined, that is, the forms corresponding to generating the corresponding shared expert features to be detected and specific expert features to be detected can be determined.

[0073] In an embodiment of the present invention, the gating module includes a data processing unit to be detected, a first gating multiplier, a second gating multiplier, and a gating adder, where When performing gating weighted fusion, the data processing unit to be detected performs at least data linear activation processing on the near-infrared spectrum data to be detected, so as to generate activated features to be detected after the data linear activation processing; The first gating multiplier is used to multiply the shared expert features to be detected by the activated features to be detected to generate shared activated features to be detected; at the same time, the second gating multiplier is used to multiply the specific expert features to be detected by the activated features to be detected to generate specific activated features to be detected; The gating adder is used to perform an addition operation on the shared activated features to be detected and the specific activated features to be detected to generate a gating weighted feature sequence to be detected.

[0074] It should be noted that the gating modules for customizing the gating module preferably adopt the same structural form, such as including a data processing unit to be detected, a first gating multiplier, a second gating multiplier, and a gating adder, where the data processing unit to be detected can perform linear activation processing on the near-infrared spectrum data to be detected. Figure 7 An embodiment of the gating module of the present invention is shown. It can be seen from the figure that Figure 7 in corresponds to Figure 3 the gating module G1 in, Figure 7 the Input in specifically refers to the near-infrared spectrum data to be detected, the specific expert module A is the specific expert features to be detected, and the shared expert module is the shared expert features to be detected.

[0075] Figure 7An embodiment of the data to be inspected processing unit is shown. As can be seen from the figure, the data to be inspected processing unit may include a data activation linear layer and a Softmax activation function. Among them, the data activation linear layer can perform a linear transformation on the data to be inspected near-infrared spectral data, and the Softmax activation function is used to implement the activation processing. The data to be inspected near-infrared spectral data is sequentially processed by the data activation linear layer and the Softmax activation function, and the Softmax activation function can generate the data to be inspected activation features. Of course, the data to be inspected processing unit can also adopt other forms, as long as it can achieve the same linear activation processing on the data to be inspected near-infrared spectral data, which will not be listed one by one here.

[0076] Specifically, in implementation, the gated first multiplier is used to multiply the data to be inspected shared expert features and the data to be inspected activation features to generate the data to be inspected shared activation features; at the same time, the gated second multiplier multiplies the data to be inspected specific expert features and the data to be inspected activation features to generate the data to be inspected specific activation features; thereafter, the gated adder is used to perform an addition operation on the data to be inspected shared activation features and the data to be inspected specific activation features to generate the data to be inspected gated weighted feature sequence. From the above description, it can be seen that the generated data to be inspected gated weighted feature sequence should be loaded into the corresponding tower module, such as Figure 7 In the example shown in, the data to be inspected gated weighted feature sequence should be loaded into the tower module TowerA. Figure 7 In, CF2 is the gated first multiplier, CF3 is the gated second multiplier, and Ad1 is the gated adder.

[0077] From the above description, based on the dynamic fusion mechanism of the customized gated module, the balance between quality parameter prediction tasks can be effectively achieved, task conflicts and sample correlations can be better handled, the knowledge transfer and sharing of different quality parameter prediction tasks are promoted, and thus better performance is achieved in multi-quality parameter prediction.

[0078] In an embodiment of the present invention, when constructing a quality parameter collaborative detection model, it includes: Constructing a quality parameter collaborative detection basic model and a basic model training data set for training the quality parameter collaborative detection basic model, where The basic model training data set includes several training samples. Each training sample includes a training near-infrared spectral data and several quality parameter labels. The number of quality parameter labels in the training sample is consistent with the number of quality parameter prediction values in the multi-quality parameter information, and the types of quality parameter labels and the types of quality parameter prediction values are in one-to-one correspondence; Configuring the model training conditions for the quality parameter collaborative detection basic model until the quality parameter collaborative detection basic model is trained to the target state. Thereafter, the quality parameter collaborative detection basic model trained to the target state is configured as the quality parameter collaborative detection model.

[0079] It is understandable that the basic model for collaborative detection of quality parameters should have the same structure as the above-mentioned quality parameter collaborative detection model. Therefore, the corresponding basic model for collaborative detection of quality parameters can be constructed according to the description of the above quality parameter collaborative detection model. After constructing the basic model for collaborative detection of quality parameters, a training data set for the basic model should also be constructed, and the basic model for collaborative detection of quality parameters should be trained using the constructed training data set for the basic model. The model training process for the basic model for collaborative detection of quality parameters can refer to the corresponding description below.

[0080] In order to construct the required training data set for the basic model, training substances should be provided. Among them, the training substances should belong to the same substance as the substance to be detected. As can be seen from the above description, when the substance to be detected is bauxite, the training substances should also be bauxite. When the substance to be detected is other types of substances, the training substances can be correspondingly selected and determined. After selecting and determining the training substances, the near-infrared spectra of the training substances can be collected by referring to the above method for obtaining the near-infrared spectral data of the substance to be detected, so that the corresponding training near-infrared spectral data can be obtained after the near-infrared spectral data is collected. In addition, according to the type of the training substances, the quality parameter labels of the training substances can be measured by the methods in the technical field. It should be noted that the quality parameter labels are the measured values of the corresponding quality parameters of the training substances.

[0081] In order to achieve the collaborative detection of multiple quality parameters, multiple quality parameters of the training substances should be measured, and thus multiple quality parameter labels can be obtained, and the quality parameter types of the multiple quality parameter labels are completely different. It should be noted that after obtaining multiple quality parameter labels and the corresponding training near-infrared spectral data, a training sample can be formed. Among them, for each training sample, the number of quality parameter labels is consistent with the number of quality parameter predicted values in the multi-quality parameter information, and the type of the quality parameter labels corresponds one-to-one with the type of the quality parameter predicted values.

[0082] Taking the substance to be detected and the training substance as bauxite as an example, the method and process for constructing the training data set for the basic model will be illustrated below.

[0083] Specifically, 424 bauxite samples with a particle size of 0.15 mm after drying, crushing, grinding, and screening were collected, and the Al in the bauxite samples was determined by X-ray fluorescence spectrometry based on the standard "Chemical Analysis Methods for Bauxite Ores" 2 O 3 、SiO 2 、Fe 2The corresponding content, that is, the measured values of the corresponding quality parameters are obtained, and corresponding multi-quality parameter labels can be formed accordingly. It can be seen from the above description that for each bauxite sample, the near-infrared spectrum should also be collected to obtain the near-infrared spectrum data of each bauxite sample.

[0084] After obtaining the near-infrared spectrum data and the corresponding measured values of the quality parameters of each bauxite sample, in order to improve the quality of the constructed basic model training dataset, data cleaning processing should also be performed. For example, the criterion based on Mahalanobis distance can be used to eliminate abnormal data and the data preprocessing method of SNV. The method of abnormal data elimination based on the criterion and the data preprocessing method of SNV can be consistent with the existing technology and will not be elaborated here. After data cleaning processing, the corresponding basic model training dataset can be constructed.

[0085] It should be understood that when training the quality parameter collaborative detection basic model, the model training conditions should also be configured. The configured model training conditions generally include the training loss function and the training configuration parameters. Specifically, when implemented, the configured training configuration parameters can include: using the Adam optimizer with a regularization weight of 0.001, setting the initial learning rate to 0.00025, using the ReduceLROnPlateau learning rate decay strategy, and dynamically adjusting the learning rate according to the loss value of the basic model training validation set. The size of each batch is 32, and the maximum number of iterations is set to 150.

[0086] It should be noted that when constructing the basic model training dataset, the basic model training validation set should also be constructed. The method of constructing the basic model training validation set can refer to the corresponding description of the above basic model training dataset, and the learning rate is dynamically adjusted based on the loss value of the basic model training validation set. When using the above model training conditions, when the maximum number of iterations of the model training reaches 150, the training of the quality parameter collaborative detection basic model reaches the target state. After that, the quality parameter collaborative detection model is configured with the quality parameter collaborative detection basic model that has been trained for 150 generations.

[0087] In an embodiment of the present invention, the configured model training conditions include a training loss function, and the training loss function includes:

[0088] Among them, is the training loss value, is the regression training loss, is the orthogonal training loss, is the number of training samples, is the number of tasks during collaborative detection, is the The quality parameter label of the -th training sample corresponding to the k-th task, is the predicted value of the quality parameter of the -th training sample corresponding to the k-th task, is the batch size during model training, is the training shared expert feature matrix of the -th batch of training samples, is the transposed matrix of the training shared expert feature matrix, is the training specific expert feature matrix of the -th batch of training samples corresponding to the k-th task,

[0089] During specific implementation, the number of tasks during collaborative detection is the number of quality parameter predicted values that need to be generated during collaborative detection processing. For example, Figure 3 in the illustrated embodiment, the number of tasks during collaborative detection should be 3. It should be noted that for the predicted value of the quality parameter of the -th training sample corresponding to the k-th task, it can be directly obtained through the quality parameter collaborative detection basic model. After constructing the basic model training dataset, the batch size during model training can be determined according to the number of training samples in the basic model training dataset and the size of each batch in the above description .

[0090] It should be noted that since there are multiple specific expert modules, gating modules, and tower modules, during initial training, it is necessary to specify a tower module to predict the corresponding quality parameter to satisfy the calculation of the above training loss function. For example, tower module TowerA can be specified to predict a quality parameter, and during inference, the corresponding quality parameter predicted value can be obtained through the output of tower module TowerA.

[0091] To decompose the near-infrared spectral data to be detected to the greatest extent and obtain the specific expert features to be detected and the shared expert features to be detected, the present invention introduces an orthogonal constraint, that is, adding an orthogonal training loss to the training loss function to ensure that the trained quality parameter collaborative detection model can focus on the specific expert features to be detected and avoid redundant information.

[0092] It should be noted that, due to the adoption of orthogonal constraints in model training, the correlation between multiple quality parameters of the substance to be detected is fully considered. Therefore, when using the quality parameter collaborative detection model for collaborative detection processing, while maintaining high precision, the correlation between multiple quality parameters of the substance to be detected can be fully considered, improving the accuracy and reliability of generating multi-quality parameter information. In addition, when using the quality parameter collaborative detection model of the present invention for collaborative detection processing, it can also overcome the limitations of traditional detection methods, such as complex preprocessing of the substance to be detected, expensive analytical instruments, and the ability to only detect a single quality parameter.

[0093] During model training, for each batch of training samples, the training shared expert features and corresponding training specific expert features of each training sample can be obtained through the feature extraction module in the quality parameter collaborative detection basic model. The situations of the training shared expert features and training specific expert features can refer to the description of the to-be-detected shared expert features and to-be-detected specific expert features. The difference is that here they are generated based on the training near-infrared spectral data of the training samples, while the above-mentioned to-be-detected shared expert features and to-be-detected specific expert features are generated based on the to-be-detected near-infrared spectral data.

[0094] When the number of training samples in each batch is 32, for each batch of training samples, 32 corresponding training shared expert features can be obtained. At this time, a training shared expert feature matrix can be constructed based on the 32 training shared expert features. Similarly, a training specific expert feature matrix for each task of the current batch of training samples can be obtained. Figure 6 An embodiment of introducing orthogonal constraints is shown. It can be seen from the figure that when introducing orthogonal constraints, the orthogonal constraints of the training shared expert features and each training specific expert feature of each training sample are mainly calculated. In the figure, ZJ1 and ZJ2 represent the orthogonal constraints.

[0095] Based on the above description of the orthogonal constraints, the orthogonal training loss of each batch of training samples and the orthogonal training loss of all training samples can be calculated. It can be understood that for the loss value of the training validation set of the basic model, the above-mentioned training loss value can be referred to. The corresponding description is not repeated here.

[0096] It should be understood that after training the constructed quality parameter collaborative detection basic model using the above model training method to obtain the quality parameter collaborative detection model, for the to-be-detected near-infrared spectral data of the to-be-detected substance, when the quality parameter collaborative detection model performs collaborative detection processing, the corresponding to-be-detected multi-quality shared features can be effectively decomposed, and then after feature fusion prediction processing, all quality parameter prediction values can be obtained simultaneously, that is, the collaborative detection processing of multi-quality parameters is realized.

[0097] In summary, a multi-quality parameter collaborative detection system based on near-infrared spectroscopy can be obtained. Specifically, it includes a multi-quality parameter collaborative detection device, and the above-mentioned quality parameter collaborative detection model is deployed inside the multi-quality parameter collaborative detection device. Among them, For the near-infrared spectral data to be detected of any substance to be detected, the multi-quality parameter collaborative detection device performs collaborative detection and processing using the method described above, so as to obtain the multi-quality parameter information of the substance to be detected after the collaborative detection and processing.

[0098] It should be noted that the multi-quality parameter collaborative detection device can adopt commonly used existing computer terminal devices. Using the common methods in this technical field, the quality parameter collaborative detection model can be deployed in the multi-quality parameter collaborative detection device. After that, for the near-infrared spectral data to be detected of any substance to be detected, the multi-quality parameter collaborative detection device performs collaborative detection and processing using the method described above, so as to obtain the multi-quality parameter information of the substance to be detected. The manner and process of the multi-quality parameter collaborative detection device for performing collaborative detection and processing can refer to the above description and will not be elaborated here.

[0099] The above schematically describes the present invention and its implementation manners. This description is not restrictive. Without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. What is shown in the drawings is only one of the implementation manners of the present invention. The actual structure is not limited thereto. Any reference numerals should not limit the claimed rights. Therefore, if those of ordinary skill in the art are inspired by it and design similar structural manners and embodiments to this technical solution without creative efforts without departing from the purpose of this creation, they should all fall within the protection scope of this application. In addition, the word "including" does not exclude other elements or steps, and the word "a" before an element does not exclude including "a plurality" of such elements. The multiple elements stated in the product claims can also be implemented by one element through software or hardware. The words such as first and second are used to represent names and do not represent any specific order.

Claims

1. A multi-quality parameter collaborative detection method based on near infrared spectroscopy, characterized in that: The method comprises: Providing near infrared spectrum data of the substance to be tested, and loading the near infrared spectrum data to be tested into the constructed quality parameter collaborative detection model, so as to use the quality parameter collaborative detection model to perform collaborative detection processing, and generate multiple quality parameter information of the substance to be tested, wherein, When performing collaborative detection processing, the near-infrared spectrum data to be inspected is subjected to at least feature dimension expansion processing, feature extraction processing and feature fusion prediction processing, and multiple quality parameter information of the substance to be inspected is generated after the feature fusion prediction processing.

2. The multi-quality parameter collaborative detection method based on near infrared spectroscopy according to claim 1 is characterized in that: The quality parameter collaborative detection model includes an embedding layer, a feature extraction module and a customized gating network connected in sequence, wherein: Perform feature dimension expansion processing on the near-infrared spectrum data to be inspected through the embedding layer, and generate spectrum data after dimension expansion to be inspected; The feature extraction module performs feature extraction processing on the spectral data after the dimension expansion to be inspected, and generates multi-quality shared features to be inspected; Through a customized gating network, feature fusion prediction processing is performed on the multi-quality shared features to be tested and the near-infrared spectral data to be tested to generate multi-quality parameter information, where: When performing feature fusion prediction processing, first perform feature linear activation processing on the multiple quality shared features to be tested to generate a shared expert feature to be tested and several specific expert features to be tested. Thereafter, each specific expert feature to be tested is gated and weighted fused with the shared expert feature to be tested and the near-infrared spectrum data to be tested, and the corresponding quality parameter prediction value is generated after regression prediction, wherein the number of the specific expert features to be tested is consistent with the number of quality parameter prediction values ​​in the multiple quality parameter information; Based on all the quality parameter prediction values, multiple quality parameter information of the substance to be tested is formed.

3. The multi-quality parameter collaborative detection method based on near infrared spectroscopy according to claim 2 is characterized in that: The feature extraction module includes a plurality of feature extraction submodules connected in series, wherein: For any feature extraction submodule, a deep point convolutional network and a spatial dimension feature weighting module are sequentially connected, and the deep point convolutional network and the spatial dimension feature weighting module are configured to form a residual connection; When performing feature extraction processing, for any feature extraction submodule, first use the deep point convolution network to perform single-channel long-distance feature extraction processing and channel fusion processing on the basic data of the feature extraction to be tested, and generate the extracted fusion feature to be tested; The spatial dimension feature weighting module is used to perform spatial dimension feature weighting processing on the features after extraction and fusion to be inspected, so as to generate the spatial dimension weighted features to be inspected; thereafter, the spatial dimension weighted features to be inspected and the basic data of the feature extraction to be inspected are processed by residual connection, and the data features of the sub-module to be inspected are generated.

4. The multi-quality parameter collaborative detection method based on near infrared spectroscopy according to claim 3 is characterized in that: The deep point convolution network includes a deep convolution layer using a large-size convolution kernel, a batch normalization layer, a first point convolution layer, a GeLU activation function, and a second point convolution layer sequentially connected to the deep convolution layer, wherein: A single-channel long-distance feature extraction process is performed on the basic data of feature extraction to be inspected through a deep convolution layer with a large-size convolution kernel; An inverted bottleneck structure is formed by the first point convolution layer and the second point convolution layer, and channel fusion processing is performed using the formed inverted bottleneck structure. During the channel fusion processing, the channel dimension is first expanded r times through the first point convolution layer, and then the expanded channel dimension is restored through the second point convolution layer.

5. The multi-quality parameter collaborative detection method based on near infrared spectroscopy according to claim 3 is characterized in that: the spatial dimension feature weighting module includes a spatial dimension convolution block, a spatial dimension Sigmoid layer and a spatial dimension multiplier, wherein: When performing spatial dimension feature weighting processing, the extracted and fused features to be inspected are first convolved using the spatial dimension convolution block to generate a single-channel feature sequence to be weighted. Thereafter, the single-channel feature sequence to be weighted is converted into a probability distribution using the spatial dimension Sigmoid layer to form single-channel probability distribution information, wherein the extracted and fused features to be inspected are generated by a deep point convolutional network belonging to the same feature extraction submodule; The single-channel probability distribution information is multiplied by the spatial dimension multiplier with the extracted and fused features to be detected to generate the weighted features of the spatial dimension to be detected.

6. The multi-quality parameter collaborative detection method based on near infrared spectroscopy according to any one of claims 2 to 5, characterized in that: The customized gating network includes a customized gating module and a tower network adapted to be connected to the customized gating module, wherein: The customized gating module includes an expert unit and a gating unit, wherein the expert unit includes a shared expert module and a plurality of specific expert modules. The gating unit includes a plurality of gating modules, the number of the specific expert modules is consistent with the number of the gating modules, and the number of the gating modules is not less than the number of the quality parameter prediction values ​​in the multiple quality parameter information; The tower network includes several tower modules for regression prediction, and the tower modules are connected to the gate control modules in a one-to-one correspondence; During feature fusion prediction processing, the expert unit is used to perform feature linear activation processing on the multi-quality shared features to be tested, so as to generate shared expert features to be tested through the shared expert module, and generate corresponding specific expert features to be tested through a specific expert module; The shared expert features to be tested, the near-infrared spectral data to be tested, and the specific expert features to be tested are loaded into the corresponding gating modules respectively, so as to use the gating modules to perform gated weighted fusion and generate a gated weighted feature sequence to be tested, and the gated weighted feature sequence to be tested is loaded into the corresponding connected tower module; The tower module performs regression prediction on the received gated weighted feature sequence to be inspected, so as to generate a corresponding quality parameter prediction value after the regression prediction.

7. The multi-quality parameter collaborative detection method based on near infrared spectroscopy according to claim 6 is characterized in that: The gate control module includes a data processing unit to be inspected, a gated first multiplier, a gated second multiplier and a gated adder, wherein: During gated weighted fusion, the to-be-detected data processing unit performs at least data linear activation processing on the to-be-detected near-infrared spectrum data to generate to-be-detected activation features after the data linear activation processing; The shared expert feature to be checked and the activation feature to be checked are multiplied by a gated first multiplier to generate a shared activation feature to be checked; at the same time, the specific expert feature to be checked and the activation feature to be checked are multiplied by a gated second multiplier to generate a specific activation feature to be checked; The gated adder is used to add the shared activation features to be detected and the specific activation features to be detected to generate a gated weighted feature sequence to be detected.

8. The multi-quality parameter collaborative detection method based on near infrared spectroscopy according to claim 6 is characterized in that: When building a quality parameter collaborative detection model, it includes: Constructing a basic model for collaborative detection of quality parameters and a basic model training data set for training the basic model for collaborative detection of quality parameters, wherein: The basic model training data set includes a plurality of training samples, each of which includes a training near-infrared spectrum data and a plurality of quality parameter labels, the number of quality parameter labels in the training samples is consistent with the number of quality parameter prediction values ​​in the multiple quality parameter information, and the type of the quality parameter label is in one-to-one correspondence with the type of the quality parameter prediction value; The model training conditions for the quality parameter collaborative detection basic model are configured until the quality parameter collaborative detection basic model is trained to a target state, and thereafter, the quality parameter collaborative detection basic model trained to the target state is configured as the quality parameter collaborative detection model.

9. The multi-quality parameter collaborative detection method based on near infrared spectroscopy according to claim 6 is characterized in that: The configured model training conditions include a training loss function, which includes: in, is the training loss value, is the regression training loss, is the orthogonal training loss, is the number of training samples, is the number of tasks in collaborative detection, For the The quality parameter label of the kth task corresponding to each training sample, For the The quality parameter prediction value of the kth task corresponding to the training sample, is the number of batches during model training, For the The training shared expert feature matrix of batch training samples, is the transposed matrix of the training shared expert feature matrix, For the The batch of training samples corresponds to the training-specific expert feature matrix of the k-th task, is the square of the Frobenius norm.

10. A multi-quality parameter collaborative detection system based on near infrared spectroscopy, characterized in that: It includes a multi-quality parameter collaborative detection device, and deploys the above-mentioned quality parameter collaborative detection model in the multi-quality parameter collaborative detection device, wherein: For any near-infrared spectrum data of a substance to be tested, the multi-quality parameter collaborative detection device adopts the method described in any one of claims 1 to 9 to perform collaborative detection processing to obtain multi-quality parameter information of the substance to be tested after the collaborative detection processing.

Images