A method for detecting lithium ore content

Abstract

The application discloses a kind of detection methods based on lithium ore content, it is related to mineral detection field, comprising the following steps: step one: collection geological data in different strata in mining area, and extract lithium ore sample in data, to simulate ore particle distribution;The present application can identify the high probability area of lithium content in ore, and by maximizing information gain, ensure that the amount of information provided by each sampling point is maximum, reduce the case of redundant sampling, greatly reduce the cost and time of sampling, effectively improve the quality of data, reduce the necessity of manual intervention, so that the exploration process can realize higher degree of automation;Can ensure that the model can better reflect the lithium content in ore sample, enhance the robustness and adaptability of model, improve the calculation efficiency, reduce the consumption of computing resources, so as to reduce the time cost and hardware overhead of the whole detection process, can provide more comprehensive optimization solution for practical application.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of mineral detection, and particularly relates to a detection method based on lithium ore content. BACKGROUND

[0002] The distribution of lithium ore has high complexity, and the lithium content in the ore often has spatial heterogeneity, which makes the traditional lithium ore exploration method face the challenges of accuracy and efficiency. Therefore, developing a detection method based on lithium ore content has become a key to improving lithium ore mining accuracy, reducing environmental impact and improving resource utilization;

[0003] The existing detection method based on lithium ore content cannot identify the high probability area of lithium content in the ore, and the redundant sampling is more, and the cost and time required for sampling are higher, and the necessity of manual intervention is increased. In addition, the existing detection method based on lithium ore content cannot better reflect the lithium content in the ore sample, reduce the robustness and adaptability of the model, and increase the consumption of computing resources. Therefore, we propose a detection method based on lithium ore content. SUMMARY

[0004] The purpose of the present application is to solve the defects in the prior art and provide a detection method based on lithium ore content.

[0005] The present application provides a detection method based on lithium ore content, which solves the technical problems by adopting the following technical solutions:

[0006] Step one: collect geological data in different strata of the mining area, and extract lithium ore samples in the data to simulate the distribution of ore particles;

[0007] Step two: obtain the spectral data of the sample, extract the features of each group of spectral data, and identify the characteristic peaks of the lithium element;

[0008] Step three: store the spectral data and sample information of multiple mining areas in a distributed manner, and according to the existing sample data and the corresponding lithium content, construct a detection model to predict the lithium content of unknown samples;

[0009] Step four: according to the prediction result, optimize the detection model parameters, and then perform real-time lithium content detection on the newly collected samples, and dynamically optimize the detection result.

[0010] As a further scheme of the present application, the specific steps of simulating the distribution of ore particles in step one are as follows:

[0011] S101: Sort the collected geological data according to the small to large order respectively, use the quartile algorithm to divide each group of sorted geological data into 4 parts according to the proportion of 25%, and calculate the first quartile and the third quartile of the data;

[0012] S102: Based on the first quartile and the third quartile , calculate the interquartile range of various geological data , and then Set to the upper limit of the normal data range, Set it as the lower limit of the normal data range, identify all data outside the normal data range in the geological data, regard them as outliers and eliminate them, and after eliminating the outliers, re-examine the distribution of the data;

[0013] S103: Standardize and process the filtered geological data, extract lithium ore samples from the geological data, and establish a three-dimensional simulation model of ore particle distribution based on each set of geological data and lithium ore samples. Use the constructed three-dimensional simulation model to generate particle distribution probability maps at different depths and locations in the mining area, and simultaneously obtain high-probability areas for lithium ore particles. The specific expression of the high-probability areas for lithium ore particles is as follows:

[0014]

[0015]

[0016] Where, Representative position Particle density function at ; represents the diameter distribution of particles; Represents the density of the ore; The randomness parameter representing the particle distribution is adjusted by It can simulate the distribution of ore particles under different geological conditions; Representative The probability of particle distribution within a sampling area; Representative The volume of the sampling area; represents the particle density function;

[0017] S104: The obtained high-probability areas of each lithium ore particle are used as candidate sampling points, and the potential information gain of each candidate sampling point for lithium content detection is calculated. The number of sampling points is optimized based on the information gain distribution of the sampling points and the sampling cost.

[0018] As a further solution of the present invention, the specific steps of optimizing the number of sampling points in S104 are as follows:

[0019] S201: Establishing a sampling point location set based on the information gain distribution of the sampling points Maximize the total information gain as the goal, and set the objective function, and then determine the search range of the sampling points according to the two-dimensional and three-dimensional regions of the mining area, and randomly generate a plurality of groups of particle positions in the specified search range, wherein each particle represents a sampling point layout scheme;

[0020] S202: initialize the initial velocity of each particle, set the particle velocity range, calculate the information gain corresponding to the position of each particle according to the objective function, and record the calculation result as the fitness value of the corresponding particle, at the beginning of each iteration, each particle compares the current fitness value with the historical fitness value corresponding to it, and takes the fitness peak value as the historical individual optimal value, then compare the current fitness value of each particle, and take the position of the fitness peak particle as the global optimal point;

[0021] S203: update the velocity of each particle according to the individual optimal value of each particle and the global optimal point of the group, and then update the position of the particle according to the updated velocity, after updating, detect whether the position and velocity of each group of particles after updating are within the preset search range and velocity range, if not, correct the position and velocity of the particle to the boundary value of the range;

[0022] S204: repeat the historical individual optimal value selection, global optimal point selection, particle velocity update and particle position update until the fitness value of the global optimal point converges to the set threshold, stop iteration, and output the particle position of the global optimal point as the sampling point layout scheme of information gain peak.

[0023] As a further scheme of the present application, the specific steps of extracting the spectral data features of each group and identifying the characteristic peak value of lithium element in step two are as follows:

[0024] S301: Obtain the spectral data of the lithium ore sample through XRF and LIBS instruments, preprocess the collected spectral data by removing noise, correcting baseline and standardizing, and then segment the preprocessed spectral data into fixed length windows to construct corresponding input data;

[0025] S302: Construct a CNN network, input the input data into the CNN network, and perform forward propagation on the input data through the CNN network, and perform convolution operation on the input data through the convolution layer of the CNN network to extract the feature data in the spectral data, and then perform feature dimension reduction on the extracted feature data through the pooling layer;

[0026] S303: Repeat the convolution and pooling operations through the CNN network until the final dimension-reduced feature data is output after reaching the tail pooling layer, classify each wavelength point in the extracted feature data through the fully connected layer, and output the specific position of the lithium element characteristic peak, and then locate the specific wavelength position of the lithium characteristic peak by maximizing the prediction probability.

[0027] As a further scheme of the present application, the specific steps of predicting the lithium content of the unknown sample in step three are as follows:

[0028] S401: Collect spectrum data from different mining areas, remove noise signals in each group of spectrum data, scale the data to the range of [0, 1] through normalization processing, divide the preprocessed sample data according to 75% training set, 15% validation set and 15% test set, and construct a CNN-LSTM hybrid model as a detection model;

[0029] S402: Input the training set into the detection model in batches, the CNN module in the model receives the training set data, performs convolution operation on the spectrum data through the convolution layer in the CNN module, extracts local features in each group of spectrum data, and then uses the pooling layer to perform average pooling on the extracted local features to reduce the dimension of the local features. After multiple convolution layers and pooling layers, the ReLU activation function in the fully connected layer is used to model the processed local features;

[0030] S403: The CNN module transmits the modeling results as input data to the LSTM module, the LSTM module calculates the outputs of the input gate, the forgetting gate and the output gate according to the current input data and the hidden state at the last time, and updates the memory unit at the current time. Then, according to the value of the output gate, the hidden state at the current time is calculated and output as the final prediction result of the lithium ore content;

[0031] S404: Calculate the loss value of the prediction result and the true lithium ore content data through the mean square error loss function, then input the calculated loss value from the output layer of the detection model, and perform layer-by-layer transmission based on the chain rule, while updating the gradient descent of each layer parameter. After each round of training, evaluate the performance of the model through the validation set, and adjust the learning rate, batch size, CNN convolution layer number and LSTM layer number of the detection model based on the evaluation result;

[0032] S405: multiple rounds of training and verification are performed on the detection model until the loss value change value of the model on the verification set converges to a set threshold, after the training and verification process is completed, the trained model is finally evaluated using the test set, and the difference between the predicted value and the actual value is calculated using the mean square error and mean absolute error functions to evaluate the accuracy of the model, if the model meets the preset performance, it is deployed to the actual detection platform, otherwise the model is retrained and verified;

[0033] S406: input the actually collected mine sample data into the detection model, the input mine sample data is processed through forward propagation and the CNN and LSTM modules in the detection model, and the lithium ore content prediction results of each mine are output.

[0034] As a further scheme of the present application, the specific steps of optimizing the detection model parameters in step four are as follows:

[0035] S501: initialize a population containing multiple combinations of detection model parameters, wherein each individual represents a parameter combination, calculate the prediction accuracy of the current lithium ore content of each individual, and traverse the prediction accuracy of each individual in the initial population, and select the individual with high prediction accuracy as the qualified solution of the current population;

[0036] S502: set a set of coefficient vectors, and at the same time in each iteration, generate a random number between 0 and 1, if the generated random number is greater than the preset probability parameter, calculate the coefficient of controlling position update according to the coefficient vector and step length , according to the size of the direction to determine whether the remaining individuals are surrounded around the qualified solution;

[0037] S503: if <1, it indicates that the remaining individuals are surrounded around the qualified solution, and the position of the remaining individuals is updated according to the qualified solution, that is, the detection model parameter value of each individual is adjusted, if ≥1, it indicates that the remaining individuals are away from the qualified solution, through random search, a random position is selected in the population space to update the position of the remaining individuals, and after each iteration is completed, the coefficient vector is updated based on the linear decreasing rule;

[0038] S504: if the generated random number is less than or equal to the preset probability parameter, calculate the distance between the positions of the remaining individuals in the population and the qualified solution, and simulate the motion law of each individual approaching the qualified solution along the spiral trajectory through the spiral motion formula to update the position of the individual;

[0039] S505: Repeat the eligible solution selection and the iterative update of the position until the fitness value converges to within a preset threshold range, and then compare the prediction accuracy in each group of individuals, and output the detection model parameter combination with high prediction accuracy, and adjust the current detection model parameters based on the detection model parameter combination.

[0040] As a further scheme of the present application, the specific steps of dynamically optimizing the detection result in step four are as follows:

[0041] S601: Obtain each data of the detection platform through historical data and expert experience, calculate the prior distribution of the current detection platform based on the collected data, and simultaneously collect lithium ore detection sample data of each mining area in real time, and express the new lithium ore content detection result as , wherein represents the lithium content measurement value of the i-th sample, ; ;

[0042] S602: Update the likelihood function corresponding to the detection platform according to the newly obtained lithium ore content detection result, and then calculate the posterior distribution of the detection platform of the mining area according to the corresponding prior distribution and the likelihood function, so as to obtain the probability distribution of the model parameters under the new data, visualize the posterior distribution result of the detection model, and select each peak value in the posterior distribution to select the required control scheme;

[0043] S603: Repeat the posterior distribution update and scheme adjustment until the prediction accuracy change value of the subsequent control scheme converges to within a set range, and adjust each detection platform according to the obtained optimal control scheme, and then each time new sample arrives, the detection platform updates the posterior distribution and optimizes the control scheme according to the new data, and then the detection platform updates the detection result according to the optimized control scheme, and visualizes the detection result based on the simulation model.

[0044] Advantages of the present application:

[0045] 1. Compared with the traditional sampling method based on experience or rules, which may cause too many or too few sampling points, wasting time and resources, the present application can analyze the mining area more finely by simulating the distribution of ore particles, predict the lithium content change in different regions, identify the high-probability region of lithium content in ore, and maximize the information gain to ensure that each sampling point provides the maximum amount of information, reduce the case of redundant sampling, greatly reduce the cost and time of sampling, effectively improve the quality of data, reduce the necessity of manual intervention, and make the exploration process more automated.

[0046] 2. The application can ensure that the model can better reflect the lithium content in the ore sample, enhance the robustness and adaptability of the model, improve the calculation efficiency, reduce the consumption of computing resources, thereby reducing the time cost and hardware overhead of the entire detection process, and can provide a more comprehensive optimization solution for practical applications. BRIEF DESCRIPTION OF DRAWINGS

[0047] The application will be further described below with reference to the accompanying drawings.

[0048] Figure 1 A framework diagram of a lithium ore content detection method. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0050] Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0051] Embodiment 1

[0052] The embodiments of the application provide a lithium ore content detection method. Referring to Figure 1 , Figure 1 A framework diagram of a lithium ore content detection method provided by the embodiments of the application. The system includes the following steps

[0053] Collect geological data in different strata in the mining area, and extract lithium ore samples in the data to simulate ore particle distribution.

[0054] Specifically, the collected geological data is sorted in ascending order, and the sorted geological data is divided into four parts according to a 25% proportion using a quartile algorithm, and the first quartile and the third quartile of the data are calculated, the quartile range of each type of geological data is calculated according to the first quartile and the third quartile , and then is set as the upper limit of the normal data range, and The lower limit of the normal data range is set, all data outside the normal data range in the geological data is identified as an abnormal value and is removed, after removing the abnormal value, the distribution of the data is rechecked, the standardized processing of each geological data after screening is performed, and the lithium ore sample in the geological data is extracted, a three-dimensional simulation model of ore particle distribution is established according to each group of geological data and the lithium ore sample, and the three-dimensional simulation model is used to generate a particle distribution probability graph of different depths and positions in the mining area, and a high probability area of the lithium ore particle is obtained, each high probability area of the lithium ore particle obtained is used as a candidate sampling point, and the potential information gain of each candidate sampling point to the lithium content detection is calculated, the information gain distribution of the sampling point is used as a basis, and the sampling cost is combined to optimize the number of sampling points.

[0055] It needs to be further explained that the sampling point position set is established according to the information gain distribution of the sampling point The target function is set to maximize the total information gain, then the search range of the sampling point is determined according to the two-dimensional and three-dimensional areas of the mining area, and a plurality of particle positions are randomly generated in the specified search range, wherein each particle represents a sampling point layout scheme, the initial velocity of each particle is initialized, and the particle velocity range is set, then the information gain corresponding to the position of each particle is calculated according to the target function, and the calculation result is used as the fitness value of the corresponding particle and is recorded, at the beginning of each iteration, each particle compares the current fitness value with the corresponding historical fitness value, and the fitness peak value is used as the historical individual optimal value, then the current fitness values of each particle are compared, and the position of the fitness peak particle is used as the global optimal point, the velocity of each particle is updated according to the individual optimal value of each particle and the global optimal point of the group, then the position of each particle is updated according to the updated velocity, after the update, it is detected whether the position and velocity of each group of particles after the update are within the preset search range and velocity range, if not, the position and velocity of the particle are corrected to the boundary value of the range, the historical individual optimal value selection, global optimal point selection, particle velocity update and particle position update are repeatedly performed, until the fitness value of the global optimal point converges to a set threshold, the iteration is stopped, and the particle position of the global optimal point is output as the sampling point layout scheme of the information gain peak.

[0056] In this embodiment, the specific expression form of the high probability area of the lithium ore particle is as follows:

[0057]

[0058]

[0059] In the formula, represents the position of the particle density function; represents the diameter distribution of the particle; representing the density of the ore; representing the randomness parameter of the particle distribution, by adjusting The particle distribution state of the ore under different geological conditions can be simulated. representing the particle distribution probability in the first sampling area; representing the volume of the first sampling area; representing the particle density function.

[0060] Obtain the spectral data of the sample, extract the features of each group of spectral data, and identify the characteristic peak value of the lithium element.

[0061] Specifically, the spectral data of the lithium ore sample is obtained by XRF and LIBS instrument, the collected spectral data is preprocessed by noise removal, baseline correction and standardization, and then the preprocessed spectral data is segmented into a fixed length window to construct the corresponding input data. Construct the CNN network, input the input data into the CNN network, and perform forward propagation on the input data through the CNN network. The input data is convolved by the convolution layer of the CNN network to extract the feature data in the spectral data. The extracted feature data is reduced in dimension through the pooling layer, and the convolution and pooling operations are repeated through the CNN network until the final reduced feature data is output after reaching the tail pooling layer. Each wavelength point in the extracted feature data is classified through the full connection layer, and the specific position of the lithium element characteristic peak is output. Then, by maximizing the prediction probability, the specific wavelength position of the lithium characteristic peak is located.

[0062] Embodiment 2

[0063] The embodiment of the present application provides a detection method based on the content of lithium ore. Referring to Figure 1 , Figure 1 The embodiment of the present application provides a framework diagram of a detection method based on the content of lithium ore. The system comprises:

[0064] The spectral data and sample information of multiple mining areas are stored in a distributed manner, and a detection model is constructed according to the existing sample data and the corresponding lithium content to predict the lithium content of unknown samples.

[0065] Specifically, spectral data from different mining areas are collected, noise signals in each group of spectral data are removed, the data is scaled to the range of [0, 1] through normalization processing, the preprocessed sample data is divided into 75% training set, 15% validation set and 15% test set, a CNN-LSTM hybrid model is constructed as a detection model, the training set is input into the detection model in batches, the CNN module in the model receives the training set data, and the spectral data is convolved through the convolution layer in the CNN module, and the local features in each group of spectral data are extracted, then the average pooling is performed on the extracted local features through the pooling layer to reduce the dimension of the local features, after the convolution layer and the pooling layer are processed for multiple times, the ReLU activation function in the full connection layer is used to model the processed local features, the CNN module transmits the modeling results as input data to the LSTM module, the LSTM module calculates the outputs of the input gate, the forgetting gate and the output gate according to the current input data and the hidden state at the last moment, and updates the memory unit at the current moment, then the hidden state at the current moment is calculated according to the value of the output gate, and the final prediction result of the lithium ore content is output, the loss value of the prediction result and the true lithium ore content data is calculated through the mean square error loss function, then the calculated loss value is input from the output layer of the detection model, and is transmitted layer by layer based on the chain rule, and the gradient descent update is performed on the parameters of each layer, after each round of training, the performance of the model is evaluated through the validation set, and the learning rate, batch size, number of CNN convolution layers and number of LSTM layers of the detection model are adjusted based on the evaluation result, the detection model is trained and verified for multiple rounds, until the change value of the loss value of the model on the validation set converges to a set threshold, after the training and verification process is completed, the trained model is finally evaluated using the test set, and the difference between the predicted value and the actual value is calculated using the mean square error and the mean absolute error function to evaluate the accuracy of the model, if the model meets the preset performance, it is deployed to an actual detection platform, otherwise the model is retrained and verified, the actual collected sample data of the mining area is input into the detection model, the input sample data of the mining area is processed through the forward propagation and the CNN and LSTM modules in the detection model, and the lithium ore content prediction results of each mining area are output.

[0066] According to the prediction result, the parameters of the detection model are optimized, and the real-time lithium content detection of the newly collected samples is performed, and the detection result is dynamically optimized.

[0067] Specifically, a population containing multiple sets of detection model parameter combinations is initialized, where each individual represents a parameter combination, the prediction accuracy of the current lithium ore content of each individual is calculated, and the prediction accuracy of each individual in the initial population is traversed, and the individual with high prediction accuracy is selected as the qualified solution of the current population. A set of coefficient vectors is set, and at the same time, a random number between 0 and 1 is generated in each round of iteration. If the generated random number is greater than the preset probability parameter, the coefficient for controlling the position update is calculated based on the coefficient vector. and step length , according to the size of the direction, determine whether the remaining individuals are shrinking around the qualified solution. If <1, it means that the remaining individuals are surrounded by the solution that meets the conditions, and the positions of the remaining individuals are updated according to the solution that meets the conditions, that is, the detection model parameter values ​​of each individual are adjusted. If ≥1, it means that the remaining individuals are far away from the qualified solution. Through random search, random positions are selected in the population space, and the positions of the remaining individuals are updated. After each iteration, the coefficient vector is updated based on the linear decreasing rule. If the generated random number is less than or equal to the preset probability parameter, the distance between the positions of the remaining individuals in the population and the qualified solution position is calculated, and the spiral motion formula is used to simulate the movement law of each individual approaching the qualified solution along the spiral trajectory, update the individual position, repeat the qualified solution selection and iterative update of the position until the fitness value converges to the preset threshold range, then compare the prediction accuracy of each group of individuals, and output the detection model parameter combination with high prediction accuracy, and adjust the current detection model parameters based on the detection model parameter combination.

[0068] Specifically, the data of the detection platform are obtained through historical data and expert experience, and the prior distribution of the current detection platform is calculated based on the collected data. At the same time, the lithium ore detection sample data of each mining area is collected in real time, and the new lithium ore content detection results are expressed as ,in Indicates the The lithium content of the samples was measured. According to the newly acquired lithium ore content detection result, the likelihood function corresponding to the detection platform is updated, then according to the corresponding prior distribution and the likelihood function, the posterior distribution of the detection platform in the mine area is calculated, so as to obtain the probability distribution of the model parameters under the new data, the posterior distribution result of the detection model is visualized, and each peak value in the posterior distribution is selected to select the required control scheme, the posterior distribution updating and scheme adjustment are repeated until the prediction accuracy change value of the subsequent control scheme converges to a set range, and the obtained optimal control scheme is used to regulate and control each detection platform, then each time new samples arrive, the detection platform updates the posterior distribution according to the new data and optimizes the control scheme, then the detection platform updates the detection result according to the optimized control scheme, and the detection result is visualized based on the simulation model.

[0069] The above describes one embodiment of the present application in detail, but the content is only the preferred embodiment of the present application, and cannot be considered as limiting the scope of the present application. Any equivalent changes and improvements made according to the scope of the present application should still belong to the patent scope of the present application.

Claims

1. A detection method based on lithium ore content, characterized in that: The following steps are involved: Step 1: Collect geological data from different strata in the mining area and extract lithium ore samples from the data to simulate the distribution of ore particles; The specific steps of simulating the ore particle distribution in step 1 are as follows: S101: Sort the collected geological data from small to large, divide each group of sorted geological data into 4 parts according to the ratio of 25% using the quartile algorithm, and calculate the first quartile of the data and the third quartile ; S102: Based on the first quartile and the third quartile , calculate the interquartile range of various geological data , and then Set to the upper limit of the normal data range, Set it as the lower limit of the normal data range, identify all data outside the normal data range in the geological data, regard them as outliers and eliminate them, and after eliminating the outliers, re-examine the distribution of the data; S103: Standardize and process the filtered geological data, extract lithium ore samples from the geological data, and establish a three-dimensional simulation model of ore particle distribution based on each set of geological data and lithium ore samples. Use the constructed three-dimensional simulation model to generate particle distribution probability maps at different depths and locations in the mining area, and simultaneously obtain high-probability areas for lithium ore particles. The specific expression of the high-probability areas for lithium ore particles is as follows: ; ; Where, Representative position The particle density function at represents the particle diameter distribution, Represents the density of the ore, The randomness parameter representing the particle distribution is adjusted by It can simulate the distribution of ore particles under different geological conditions. Representative The probability of particle distribution in a sampling area, Representative The volume of the sampling area, represents the particle density function; S104: The obtained high probability areas of lithium ore particles are taken as candidate sampling points, and the lithium content detection probability of each candidate sampling point is calculated. The potential information gain of the measurement is calculated, and the number of sampling points is optimized according to the information gain distribution of the sampling points and the sampling cost; Step 2: Obtain the spectrum data of the sample, extract the characteristics of each group of spectrum data, and identify the characteristic peak of lithium element; Step 3: Distribute and store the spectrum data and sample information of multiple mining areas, and build a detection model to predict the lithium content of unknown samples based on the existing sample data and its corresponding lithium content; Step 4: Based on the prediction results, optimize the detection model parameters, then perform real-time lithium content detection on the newly collected samples and dynamically optimize the detection results.

2. The method for detecting lithium ore content according to claim 1, wherein: The specific steps for optimizing the number of sampling points in S104 are as follows: S201: Establishing a sampling point location set based on the information gain distribution of the sampling points , with the goal of maximizing the total information gain and setting the objective function, then determining the search range of the sampling points according to the two-dimensional and three-dimensional areas of the mining area, and randomly generating multiple groups of particle positions within the specified search range, where each particle represents a sampling point layout scheme; S202: Initialize the initial velocity of each particle and set the particle velocity range. Then, calculate the information gain corresponding to the position of each particle according to the objective function. At the same time, the calculation result is used as the fitness value of the corresponding particle and recorded. At the beginning of each iteration, each particle compares its current fitness value with its corresponding historical fitness value, and takes the fitness peak as the historical individual merit. Then, compare the current fitness values ​​of each particle, and take the position of the fitness peak particle as the global optimal position. S203: Update the speed of each particle based on the individual merit of each particle and the global optimal position of the group, and then update the position of the particle based on the updated speed. After the update is completed, check whether the updated position and speed of each group of particles are within the preset search range and speed range. If not, correct the position and speed of the particle to the boundary value of that range; S204: Repeat the historical individual merit selection, global optimal position selection, particle speed update and particle position update until the fitness value change of the global optimal position converges to the set threshold, stop the iteration, and output the particle position of the global optimal position as the sampling point layout plan for information gain peaking.

3. The method for detecting lithium ore content according to claim 2, wherein: The specific steps for extracting the features of each set of spectrum data and identifying the characteristic peaks of lithium elements in step 2 are as follows: S301: Acquire spectrum data of a lithium ore sample using XRF and LIBS instruments, perform noise removal, baseline correction, and standardization preprocessing on the acquired spectrum data, and then segment the preprocessed spectrum data into windows of fixed length to construct corresponding input data; S302: Construct a CNN network, input the input data into the CNN network, and forward propagate the input data through the CNN network. Use the CNN network convolution layer to perform a convolution operation on the input data to extract feature data from the spectrogram data, and then perform feature dimensionality reduction on the extracted feature data through the pooling layer; S303: Repeat convolution and pooling operations through the CNN network until reaching the tail pooling layer and outputting the final reduced-dimensional feature data. Use the fully connected layer to classify each wavelength point in the extracted feature data and output the specific location of the lithium element characteristic peak. Then, by maximizing the prediction probability, locate the specific wavelength position of the lithium characteristic peak.

4. The method for detecting lithium ore content according to claim 3, wherein: The specific steps for predicting the lithium content of the unknown sample in step 3 are as follows: S401: Spectral data from different mining areas are collected and noise signals are removed from each set of spectral data. The data are normalized to the range of [0, 1]. The preprocessed sample data is divided into 75% training set, 15% validation set, and 15% test set. A CNN-LSTM hybrid model is constructed as the detection model. S402: The training set is input into the detection model in batches. The CNN module in the model receives the training set data and performs convolution operations on the spectrogram data through the convolution layer in the CNN module. Local features in each group of spectrogram data are extracted. The extracted local features are then averaged and pooled using the pooling layer to reduce the local feature dimension. After multiple convolutional and pooling layer processing, the processed local features are modeled using the ReLU activation function in the fully connected layer. S403: The CNN module uses the modeling results as input data and transmits them to the LSTM module. The LSTM module calculates the outputs of the input gate, forget gate, and output gate based on the current input data and the hidden state at the previous moment, and updates the memory unit at the current moment. Then, based on the value of the output gate, it calculates the hidden state at the current moment and outputs it as the final prediction result of the lithium ore content. S404: Calculate the loss value between the predicted result and the actual lithium ore content data through the mean square error loss function, then input the calculated loss value from the detection model output layer, and transfer it layer by layer based on the chain rule. At the same time, perform gradient descent updates on the parameters of each layer. After each round of training, evaluate the performance of the model through the validation set, and adjust the detection model learning rate, batch size, number of CNN convolutional layers, and number of LSTM layers based on the evaluation results. S405: The detection model is trained and validated for multiple rounds until the change in the model's loss on the validation set converges to a set threshold. After the training and validation process is completed, the trained model is finally evaluated using the test set. The mean square error and mean absolute error functions are used to calculate the difference between the predicted value and the actual value to evaluate the model's accuracy. If the model meets the preset performance, it is deployed to the actual detection platform. Otherwise, the model is retrained and validated. S406: The actual collected mining area sample data is input into the detection model. The input mining area sample data is forward propagated and processed by the CNN and LSTM modules in the detection model to output the lithium ore content prediction results of each mining area.

5. The method for detecting lithium ore content according to claim 4, wherein: The specific steps for optimizing the detection model parameters described in step 4 are as follows: S501: Initialize a population containing multiple sets of detection model parameter combinations, where each individual represents a parameter combination, calculate the prediction accuracy of the current lithium ore content of each individual, traverse the prediction accuracy of each individual in the initial population, and select the individual with high prediction accuracy as the qualified solution of the current population; S502: Set a set of coefficient vectors, and generate a random number between 0 and 1 in each round of iteration. If the generated random number is greater than the preset probability parameter, the coefficient of the control position update is calculated based on the coefficient vector. and step length , according to the size of the direction, determine whether the remaining individuals are shrinking around the qualified solution; S503: If <1, it means that the remaining individuals are surrounded by the solution that meets the conditions, and the positions of the remaining individuals are updated according to the solution that meets the conditions, that is, the detection model parameter values ​​of each individual are adjusted. ≥1, it means that the remaining individuals are far away from the qualified solution. Through random search, random positions are selected in the population space, and the positions of the remaining individuals are updated. After each iteration, the coefficient vector is updated based on the linear decreasing rule. S504: If the generated random number is less than or equal to the preset probability parameter, the distance between the positions of the remaining individuals in the population and the position of the solution that meets the conditions is calculated, and the motion law of each individual along the spiral trajectory approaching the solution that meets the conditions is simulated using the spiral motion formula to update the individual position; S505: Repeatedly select solutions that meet the conditions and iteratively update the positions until the fitness value converges to a preset threshold range, then compare the prediction accuracy of individuals in each group, and output a detection model parameter combination with high prediction accuracy, and adjust the current detection model parameters based on the detection model parameter combination.

6. The method for detecting lithium ore content according to claim 1, wherein: The specific steps for dynamically optimizing the test results described in step 4 are as follows: S601: Obtain the data of the detection platform through historical data and expert experience, and calculate the prior distribution of the current detection platform based on the collected data. At the same time, collect the lithium ore detection sample data of each mining area in real time, and express the new lithium ore content detection results as ,in Indicates the The lithium content of the samples was measured. ; S602: Update the likelihood function corresponding to the detection platform based on the newly acquired lithium ore content detection results. Then, calculate the posterior distribution of the mining area detection platform based on the corresponding prior distribution and the likelihood function to obtain the probability distribution of the model parameters under the new data. Visualize the posterior distribution results of the detection model and select the desired control solution by selecting each peak in the posterior distribution. S603: Repeat the posterior distribution update and scheme adjustment until the subsequent control scheme prediction accuracy change value converges to the set range, and adjust each detection platform according to the obtained optimal control scheme. After that, each time a new sample arrives, the detection platform updates the posterior distribution according to the new data and optimizes the control scheme. After that, the detection platform updates the detection results according to the optimized control scheme and visualizes the detection results based on the simulation model.

Images