Passive efficiency calibration detector parameter generation method based on artificial intelligence

Abstract

The invention relates to the technical field of passive efficiency scales of detectors, in particular to a method, a device and equipment for generating parameters of a passive efficiency scale detector based on artificial intelligence and a computer storage medium. According to the detector parameter generation method for the passive efficiency scale, the complex relation between the geometric parameters and the detection efficiency of the detector is reflected based on machine learning, the mapping relation is obtained through the neural network, and parameter adjustment can be directly conducted on the detector through an experimental efficiency curve and the parameters of the detector before adjustment; compared with manual parameter adjustment, the parameter adjustment process based on artificial intelligence saves time and computing resources, the process of manual repeated attempt, adjustment and Monte Carlo calculation is directly omitted, direct prediction is performed by the model based on the experimental curve, and the efficiency is greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of passive efficiency calibration of detectors, and in particular to a method, device, equipment and computer storage medium for generating detector parameters for passive efficiency calibration based on artificial intelligence. Background Art

[0002] High-purity germanium (HPGe) detectors are one of the core devices in the field of nuclear radiation detection. Based on the ionization effect of semiconductor materials, they can accurately measure the energy of photons. At present, HPGe detectors play an irreplaceable role in nuclear safety monitoring, radioactive waste management, nuclear medicine diagnosis and basic scientific research with their ultra-high energy resolution, wide energy range coverage ability and flexible multi-dimensional measurement characteristics.

[0003] When a detector detects particles, its detection efficiency changes with the energy of the detected particles. This "efficiency-energy" relationship is called the efficiency curve of the detector. In practical applications, the activity of particles at different energy points is generally calculated from the measured activity of the detector and the detection efficiency at that energy point. Therefore, achieving accurate calibration of the detector efficiency curve is of great significance for the use of detectors in various scenarios.

[0004] The passive efficiency calibration of a detector refers to, without using an actual radiation source for experiments, modeling a detector with known material and geometric parameters, and using the Monte Carlo method to simulate the particle transport process in efficiency calibration, and obtaining the efficiency curve of the detector based on the statistical results. This effectively avoids the safety risks and cost problems brought by the preparation, transportation and storage of standard sources.

[0005] With the use of the detector, some physical properties in the detector also change, resulting in a change in the detector efficiency curve. This process affects the accuracy of activity measurement. When the deviation reaches a certain level, the efficiency curve needs to be re-calibrated. This calibration often uses an active (i.e., experimental) method. After obtaining the new efficiency curve, it is necessary to manually adjust the geometric parameters of the detector in the passive efficiency calibration according to the deviation between the passive calibration efficiency and the active calibration efficiency, so that the result of the passive efficiency calibration can match the experimental curve.

[0006] Since the relationship between the detector efficiency curve and the geometric parameters of the detector model is relatively complex, in the process of manually adjusting the parameters according to the experimental curve, it is often necessary to make repeated attempts based on experience and the approximate relationship between different parameters and the efficiency curve. After each adjustment, a Monte Carlo simulation needs to be performed to obtain the passive efficiency calibration curve and compare it with the experimental results for verification.

[0007] In summary, this process of adjusting parameters based on experimental results depends on manual experience and has low efficiency. Summary of the Invention

[0008] To this end, the technical problem to be solved by the present invention is to overcome the problem of low efficiency in adjusting the detector geometric parameters in passive efficiency calibration in the prior art.

[0009] To solve the above technical problem, the present invention provides a method for generating detector parameters for passive efficiency calibration, including:

[0010] Taking the first detector parameters and the detection efficiency of the energy point as inputs, and the adjusted second detector parameters as outputs, a detector parameter prediction neural network model is established based on the mapping relationship between the first detector parameters, the detection efficiency curve, and the adjusted second detector parameters;

[0011] Obtain a set of detector parameters that have been adjusted, and randomly sample within the preset range of the set of detector parameters to generate detector parameter training samples;

[0012] Input the detector parameter training samples into the passive efficiency calibration program for Monte Carlo calculation to obtain the detection efficiency training samples of the energy point, and summarize them to obtain a training set;

[0013] Train the detector parameter prediction neural network model based on the training set to obtain a target detector parameter prediction model;

[0014] Adjust the second detector parameters to be adjusted based on the target detector parameter prediction model.

[0015] Preferably, the method of establishing a detector parameter prediction neural network model by taking the first detector parameters and the detection efficiency of the energy point as inputs, and the adjusted second detector parameters as outputs, based on the mapping relationship between the first detector parameters, the detection efficiency curve, and the adjusted second detector parameters includes:

[0016] Based on the first detector parameters and the detection efficiency of the energy point, establish multiple input neuron nodes in the input layer;

[0017] Based on the adjusted second detector parameters, establish multiple output neuron nodes in the output layer;

[0018] Based on the mapping relationship between the first detector parameters, the detection efficiency curve, and the adjusted second detector parameters, establish several intermediate layers and the network connection structure between the input and output layers.

[0019] Preferably, the method of establishing several intermediate layers and the network connection structure between the input and output layers based on the mapping relationship between the first detector parameters, the detection efficiency curve, and the adjusted second detector parameters includes:

[0020] Connect the energy points in different preset intervals and the first detector parameters to the corresponding output neuron nodes through multiple intermediate layers.

[0021] Preferably, the connecting the energy points in different preset intervals and the first detector parameters to the corresponding output neuron nodes through multiple intermediate layers includes:

[0022] Connect the input neuron nodes corresponding to the first detector parameters and the energy points in the first preset interval to the first output neuron node through multiple intermediate layers,

[0023] Connect the input neuron nodes corresponding to the first detector parameters and the energy points in the second preset interval to the second output neuron node through multiple intermediate layers;

[0024] Connect the input neuron nodes corresponding to the first detector parameters and the energy points in the third preset interval to the third output neuron node through multiple intermediate layers.

[0025] Preferably, the randomly sampling within the preset range of the detector parameter group to generate detector parameter training samples includes:

[0026] Randomly sample the adjusted second detector parameters within the preset range of the detector parameter group to generate detector parameter training samples.

[0027] Preferably, the randomly sampling within the preset range of the detector parameter group to generate detector parameter training samples includes:

[0028] Randomly sample the adjusted second detector parameters and the selected parameters among the first detector parameters within the preset range of the detector parameter group to generate detector parameter training samples.

[0029] Preferably, the selected parameters include the height of the detector electrode hole and the radius of the detector electrode hole.

[0030] The present invention also provides a detector parameter generation device for passive efficiency calibration, including:

[0031] A model construction module, configured to use the first detector parameters and the detection efficiency of the energy points as inputs, and the adjusted second detector parameters as outputs, and establish a detector parameter prediction neural network model based on the mapping relationship between the first detector parameters, the detection efficiency curve, and the adjusted second detector parameters;

[0032] A first training set construction module, configured to obtain the adjusted detector parameter group and randomly sample within the preset range of the detector parameter group to generate detector parameter training samples;

[0033] The second training set construction module is used to input the detector parameter training samples into a passive efficiency calibration program for Monte Carlo calculation, obtain the detection efficiency training samples at energy points, and summarize them to obtain a training set;

[0034] The model training module is used to train the detector parameter prediction neural network model based on the training set to obtain a target detector parameter prediction model;

[0035] The parameter adjustment module is used to adjust the second detector parameters to be adjusted based on the target detector parameter prediction model.

[0036] The present invention also provides a detector parameter generation device for passive efficiency calibration, including:

[0037] A memory for storing a computer program;

[0038] A processor for implementing the steps of the above-mentioned method for generating detector parameters for passive efficiency calibration when executing the computer program.

[0039] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned method for generating detector parameters for passive efficiency calibration are implemented.

[0040] The above technical solution of the present invention has the following advantages compared with the prior art:

[0041] The method for generating detector parameters for passive efficiency calibration according to the present invention reflects the complex relationship between the detector geometric parameters and the detection efficiency based on machine learning, and obtains the mapping relationship through a neural network, and can realize directly adjusting the parameters of the detector through the experimental efficiency curve and the detector parameters before adjustment; the parameter adjustment process based on artificial intelligence saves time and computing resources compared with manual parameter adjustment, directly omits the process of manual repeated attempts, adjustment and Monte Carlo calculation, but is directly predicted by the model based on the experimental curve, greatly improving the efficiency. Description of the Drawings

[0042] In order to make the content of the present invention easier to be clearly understood, the following further details the present invention according to specific embodiments of the present invention in conjunction with the drawings, where:

[0043] Figure 1 is the implementation flowchart of a method for generating detector parameters for passive efficiency calibration provided by the present invention;

[0044] Figure 2 is the flowchart block diagram of a method for generating detector parameters for passive efficiency calibration provided by an embodiment of the present invention;

[0045] Figure 3 Schematic diagram of the average error of the training set during the neural network training process;

[0046] Figure 4 Schematic diagram of the average error of the validation set during the neural network training process;

[0047] Figure 5 Schematic diagram of a comparison example of the model parameter tuning efficiency curve and the original efficiency curve;

[0048] Figure 6 Schematic diagram of the deviation between the calculation efficiency of each energy point of the test set (model parameter tuning) and the original efficiency. Specific implementation manners

[0049] The core of the present invention is to provide a method, device, equipment and computer storage medium for generating detector parameters for passive efficiency calibration, which effectively improves the efficiency of adjusting detector parameters.

[0050] In order to enable those skilled in the art to better understand the solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation manners. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0051] Please refer to Figure 1 and Figure 2 , Figure 1 which is a flowchart of the implementation of a method for generating detector parameters for passive efficiency calibration provided by the present invention, Figure 2 which is a flowchart of a method for generating detector parameters for passive efficiency calibration provided by an embodiment of the present invention; the specific operation steps are as follows:

[0052] S101: Using the first detector parameters and the detection efficiency of the energy points as inputs, and the adjusted second detector parameters as outputs, establish a detector parameter prediction neural network model based on the mapping relationship between the first detector parameters, the detection efficiency curve, and the adjusted second detector parameters;

[0053] S102: Obtain a set of detector parameters that have been adjusted, and randomly sample within the preset range of the set of detector parameters to generate detector parameter training samples;

[0054] S103: Input the detector parameter training samples into the passive efficiency calibration program for Monte Carlo calculation to obtain the detection efficiency training samples of the energy points, and summarize them to obtain a training set;

[0055] S104: Train the detector parameter prediction neural network model based on the training set to obtain a target detector parameter prediction model;

[0056] S105: Adjust the second detector parameters to be adjusted based on the target detector parameter prediction model.

[0057] Based on the above embodiments, this embodiment details step S101:

[0058] In some embodiments, the first detector parameters are detector parameters that do not need to be adjusted, and the second detector parameters are detector parameters to be adjusted, which can be set and adjusted according to the actual situation. This embodiment does not limit the specific parameters and types of the first detector parameters and the second detector parameters.

[0059] Since an increase in the crystal dead layer thickness will reduce the detection efficiency in the low energy region, an increase in the radius will increase the detection efficiency for all energies, and an increase in the crystal height will increase the detection efficiency in the high energy region. During the manual parameter adjustment process, often only these three parameters are adjusted without changing other detector parameters;

[0060] Therefore, in some embodiments, the neural network also only predicts these three parameters, with the remaining parameters and detection efficiency as inputs. That is, the second detector parameters can be, for example, the crystal dead layer thickness, the crystal radius, and the crystal height;

[0061] The first detector parameters can be, for example, 24 detector parameters other than the second detector parameters.

[0062] In some embodiments, the energy points can be, for example, 16 standard source energy points often used in the active efficiency calibration experiment, which can be set and adjusted according to the actual situation. This embodiment does not limit the specific number of energy points.

[0063] In some embodiments, using the first detector parameters and the detection efficiency of the energy points as inputs, and the adjusted second detector parameters as outputs, establishing a detector parameter prediction neural network model based on the mapping relationship between the first detector parameters, the detection efficiency curve, and the adjusted second detector parameters includes:

[0064] Establish multiple input neuron nodes in the input layer based on the first detector parameters and the detection efficiency of the energy points;

[0065] Establish multiple output neuron nodes in the output layer based on the adjusted second detector parameters;

[0066] Establish several intermediate layers and the network connection structure between the input and output layers based on the mapping relationship between the first detector parameters, the detection efficiency curve, and the adjusted second detector parameters.

[0067] In some embodiments, according to different detection parameters, the trend and degree of the change in the efficiency curve are different. The present invention designs a non-fully connected network to reflect the relationship between the detection parameters to be adjusted and the efficiency curve, so that the model can better master this "parameter - efficiency" mapping relationship:

[0068] Divide the energy point interval into a high-energy region, a low-energy region, and other energy regions;

[0069] Connect the energy points in different preset intervals and the first detector parameters to the corresponding output neuron nodes through multiple intermediate layers.

[0070] In some embodiments, connecting the energy points in different preset intervals and the first detector parameters to the corresponding output neuron nodes through multiple intermediate layers includes:

[0071] Connect the input neuron node corresponding to the first detector parameter and the energy points in the first preset interval to the first output neuron node through multiple intermediate layers,

[0072] Connect the input neuron node corresponding to the first detector parameter and the energy points in the second preset interval to the second output neuron node through multiple intermediate layers;

[0073] Connect the input neuron node corresponding to the first detector parameter and the energy points in the third preset interval to the third output neuron node through multiple intermediate layers.

[0074] In some embodiments, the energy points in the first preset interval may be, for example, the energy points in the low-energy region and part of other energy regions, the energy points in the second preset interval may be, for example, all the energy points, and the energy points in the third preset interval may be, for example, the energy points in the high-energy region and part of other energy regions; specifically:

[0075] Connect the input neuron node corresponding to the second detector parameter and the energy points in the low-energy region and part of other energy regions to the output neuron node of the crystal dead layer thickness through multiple intermediate layers,

[0076] Connect the multiple input neuron nodes to the output neuron node of the crystal radius through multiple intermediate layers;

[0077] Connect the input neuron node corresponding to the second detector parameter and the energy points in the high-energy region and part of other energy regions to the output neuron node of the crystal height through multiple intermediate layers.

[0078] In a specific embodiment, the three outputs respectively correspond to three parts of the neural network and are combined with different outputs. Among them, 24 input parameters and the first 11 energy points (low energy region and some other energies) are connected to the output of the crystal dead layer thickness through multiple intermediate layers, all inputs are connected to the output of the crystal radius through multiple intermediate layers, and 24 input parameters and the last 13 energy points (high energy region and some other energies) are connected to the output layer of the crystal height through multiple intermediate layers.

[0079] Based on the above embodiments, this embodiment details step S102:

[0080] The detector parameter training samples are generated based on the parameter combinations that have been adjusted currently. The present invention tests the influence degree of different parameter changes on the efficiency, and randomly samples these parameters near the original detector parameters to obtain the parameter part of the training set.

[0081] In some embodiments, random sampling is performed on the adjusted second detector parameters within the preset range of the detector parameter group to generate detector parameter training samples.

[0082] The parameter part of the detector in the generated training set is generated based on the parameter combinations that have been adjusted currently rather than random sampling. This not only ensures that there are no geometric conflicts in the detector model, but also enables the model to focus on the efficiency curve differences brought about by the specific parameter changes of mainstream detector devices, prevents model overfitting, and improves the learning effect of the model on the parameter adjustment process.

[0083] According to calculations and simulations, in addition to the crystal dead layer thickness, radius, and height, detector parameters not involved in the parameter adjustment process may also affect the efficiency curve of the detector;

[0084] Therefore, in some other embodiments, random sampling is performed on the adjusted second detector parameters and the selected parameters in the first detector parameters within the preset range of the detector parameter group to generate detector parameter training samples.

[0085] In a specific embodiment, since changes in the detector electrode hole height and radius will also affect the efficiency curve of the detector, in order for the model to better adapt to the changes in these two inputs, the values of these two parameters are also sampled near the original detector parameter group when extracting the training set.

[0086] Based on the above embodiments, this embodiment details step S103:

[0087] Input the generated detector parameters into the passive efficiency calibration program for Monte Carlo calculation, summarize the corresponding efficiency calculation results, and form a complete training set.

[0088] In some embodiments, since the previous passive efficiency calibration Monte Carlo program was calibrated only for a set of detector parameters, i.e., one detector, the present invention changes the input and output methods of the program to adapt to large-scale sampling calculations, and summarizes the detection efficiencies corresponding to the parameters in the training set into the training set format.

[0089] Based on the above embodiments, this embodiment details step S104:

[0090] In a specific embodiment, the inputs and outputs in the training set, and the training set and the test set are split to train the model. Among them, 10% of the data in the training set is randomly selected as the test set, and the remaining 90% is used for training. During the training process, the learning rate will be gradually decreased as the number of iterations increases. Specifically, every 30 epoches, it becomes 10% of the original learning rate to enhance the learning effect. Figure 3 It is a schematic diagram of the average error of the training set during the neural network training process; Figure 4 It is a schematic diagram of the average error of the validation set during the neural network training process.

[0091] Based on the above embodiments, this embodiment details step S105:

[0092] In some embodiments, after the present invention inputs the test set into the model to obtain the predicted parameters, the predicted parameters and the original parameters are respectively input back into the passive efficiency calibration Monte Carlo program for calibration, and the differences between the efficiency curves calculated by the two are compared to analyze the effects of the predicted parameters.

[0093] In a specific embodiment, the trained model is tested to analyze the accuracy of the model's predicted parameters. The specific operation method is to input the 3 parameters predicted by the model and the corresponding 24 non-adjusted parameters into the passive efficiency program for Monte Carlo calculation, and compare them with the efficiency curves given by the original 27 parameters, and analyze and obtain the results as Figure 5 , Figure 6 shown.

[0094] The process of adjusting parameters based on the experimental results is essentially a learning of the mapping relationship between the detector geometric parameters and the efficiency curve. Therefore, the present invention performs efficiency calibration on the detectors of the custom parameter set based on the passive efficiency calibration program to obtain the efficiency curve, and designs a neural network to learn the relationship between the parameters and the efficiency, so that the network can output the full set of detector parameters based on part of the detector parameters and the experimental efficiency curve, that is, replace manual adjustment of the detector parameters.

[0095] Based on the machine learning method, the present invention designs a set of detector geometric parameters, conducts Monte Carlo simulation calculations on them, obtains the corresponding efficiency curve as the training set, and designs a set of neural networks to learn the "parameters - efficiency", enabling the adjustment of detector parameters based on the experimental efficiency curve through artificial intelligence. The results show that this invention can meet the accuracy requirements of the parameter adjustment process. The efficiency results calculated based on the parameters adjusted by artificial intelligence reach the accuracy level of manual parameter adjustment, and this process does not require repeated Monte Carlo calculations. Instead, the results are directly output by the already trained network, greatly improving the efficiency of the parameter adjustment process.

[0096] The parameter adjustment process of the present invention based on artificial intelligence saves time and computing resources compared with manual parameter adjustment. It directly eliminates the process of manual repeated attempts, adjustments, and Monte Carlo calculations. Instead, the model directly predicts based on the experimental curve, greatly improving the efficiency. Based on machine learning, it reflects the complex relationship between the detector geometric parameters and the detection efficiency, and obtains the mapping relationship through the neural network, enabling the adjustment of the detector parameters directly through the experimental efficiency curve and the detector parameters before adjustment. A neural network is specifically designed for the influence of the crystal dead layer thickness, radius, and size on the detection efficiency at different energy points. The model can better master the specific relationship between these parameters and different energy points, enhancing the learning efficiency and achieving better output results.

[0097] The embodiment of the present invention also provides a device for generating detector parameters for passive efficiency calibration; the specific device may include:

[0098] A model construction module, configured to use the first detector parameters and the detection efficiency of the energy point as inputs, and the adjusted second detector parameters as outputs, and establish a detector parameter prediction neural network model based on the mapping relationship between the first detector parameters, the detection efficiency curve, and the adjusted second detector parameters;

[0099] A first training set construction module, configured to obtain a set of detector parameters that have been adjusted, and randomly sample within the preset range of the set of detector parameters to generate detector parameter training samples;

[0100] A second training set construction module, configured to input the detector parameter training samples into a passive efficiency calibration program for Monte Carlo calculation, obtain the detection efficiency training samples of the energy point, and summarize them to obtain a training set;

[0101] A model training module, configured to train the detector parameter prediction neural network model based on the training set to obtain a target detector parameter prediction model;

[0102] A parameter adjustment module, configured to adjust the second detector parameters to be adjusted based on the target detector parameter prediction model.

[0103] The detector parameter generation device for passive efficiency calibration in this embodiment is used to implement the aforementioned method for generating detector parameters for passive efficiency calibration. Therefore, the specific implementation in the detector parameter generation device for passive efficiency calibration can be seen in the embodiment part of the aforementioned method for generating detector parameters for passive efficiency calibration. For example, the model construction module, the first training set construction module, the second training set construction module, the model training module, and the parameter adjustment module are respectively used to implement steps S101, S102, S103, S104, and S105 in the aforementioned method for generating detector parameters for passive efficiency calibration. Therefore, the specific implementation can refer to the descriptions of the corresponding individual embodiments and will not be elaborated here.

[0104] The specific embodiment of the present invention also provides a device for generating detector parameters for passive efficiency calibration, including: a memory for storing a computer program; a processor for implementing the steps of the aforementioned method for generating detector parameters for passive efficiency calibration when executing the computer program.

[0105] The specific embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the aforementioned method for generating detector parameters for passive efficiency calibration are implemented.

[0106] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0107] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0108] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to work in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction device that implements the functions specified in one or more of the processes and / or blocks Figure 1 one or more of the processes and / or blocks Figure 1 specified in the block or blocks.

[0109] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one or more of the processes and / or blocks Figure 1 one or more of the processes and / or blocks Figure 1 specified in the block or blocks.

[0110] Obviously, the above embodiments are only examples for clear illustration and are not limitations on the implementation. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to exhaustively list all the implementation manners here. And the obvious changes or modifications derived therefrom are still within the protection scope of the present invention.

Claims

1. A method for generating detector parameters for passive efficiency calibration, characterized in that: include: Taking the first detector parameter and the detection efficiency of the energy point as input and the adjusted second detector parameter as output, a detector parameter prediction neural network model is established based on the mapping relationship between the first detector parameter, the detection efficiency curve and the adjusted second detector parameter; Acquire a detector parameter group that has been adjusted, and randomly sample within a preset range of the detector parameter group to generate a detector parameter training sample; Input the detector parameter training samples into the passive efficiency calibration program to perform Monte Carlo calculation to obtain the detection efficiency training samples of the energy points, and summarize them to obtain a training set; Training the detector parameter prediction neural network model based on the training set to obtain a target detector parameter prediction model; The second detector parameter to be adjusted is adjusted based on the target detector parameter prediction model.

2. The method for generating detector parameters for passive efficiency calibration according to claim 1, characterized in that: The method of taking the first detector parameter and the detection efficiency of the energy point as input and the adjusted second detector parameter as output, and establishing the detector parameter prediction neural network model based on the mapping relationship between the first detector parameter, the detection efficiency curve and the adjusted second detector parameter comprises: Establishing a plurality of input neuron nodes of an input layer based on the first detector parameter and the detection efficiency of the energy point; Establishing a plurality of output neuron nodes of an output layer based on the adjusted second detector parameters; Based on the mapping relationship between the first detector parameter, the detection efficiency curve and the adjusted second detector parameter, a plurality of intermediate layers and a network connection structure between the intermediate layers and the input and output layers are established.

3. The method for generating detector parameters for passive efficiency calibration according to claim 2, characterized in that: The network connection structure between the plurality of intermediate layers and the input and output layers based on the mapping relationship between the first detector parameter, the detection efficiency curve and the adjusted second detector parameter includes: The energy points in different preset intervals and the first detector parameters are connected to the corresponding output neuron nodes through multiple intermediate layers.

4. The method for generating detector parameters for passive efficiency calibration according to claim 3, characterized in that: The step of establishing a connection between energy points and first detector parameters in different preset intervals and corresponding output neuron nodes through multiple intermediate layers includes: The input neuron node corresponding to the first detector parameter and the first preset interval energy point is connected to the first output neuron node through multiple intermediate layers, Establishing a connection between the input neuron node corresponding to the first detector parameter and the second preset interval energy point through a plurality of intermediate layers and a second output neuron node; The input neuron nodes corresponding to the first detector parameter and the third preset interval energy point are connected through multiple intermediate layers and the third output neuron node.

5. The method for generating detector parameters for passive efficiency calibration according to claim 1, characterized in that: The randomly sampling and generating detector parameter training samples within the preset range of the detector parameter group includes: The adjusted second detector parameters are randomly sampled within a preset range of the detector parameter group to generate detector parameter training samples.

6. The method for generating detector parameters for passive efficiency calibration according to claim 1, characterized in that: The randomly sampling and generating detector parameter training samples within the preset range of the detector parameter group includes: The adjusted second detector parameters and selected parameters of the first detector parameters are randomly sampled within a preset range of the detector parameter group to generate detector parameter training samples.

7. The method for generating detector parameters for passive efficiency calibration according to claim 6, characterized in that: The selected parameters include the detector electrode hole height and the detector electrode hole radius.

8. A detector parameter generation device for passive efficiency calibration, characterized in that: include: A model building module, for taking the first detector parameter and the detection efficiency of the energy point as input and the adjusted second detector parameter as output, and establishing a detector parameter prediction neural network model based on a mapping relationship between the first detector parameter, the detection efficiency curve and the adjusted second detector parameter; A first training set construction module is used to obtain a detector parameter group that has been adjusted, and randomly sample within a preset range of the detector parameter group to generate a detector parameter training sample; A second training set construction module is used to input the detector parameter training samples into a passive efficiency calibration program to perform Monte Carlo calculations, obtain detection efficiency training samples of energy points, and summarize them to obtain a training set; A model training module, used to train the detector parameter prediction neural network model based on the training set to obtain a target detector parameter prediction model; The parameter adjustment module is used to adjust the second detector parameter to be adjusted based on the target detector parameter prediction model.

9. A detector parameter generation device for passive efficiency calibration, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of a method for generating detector parameters for passive efficiency calibration as claimed in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for generating detector parameters for passive efficiency calibration as claimed in any one of claims 1 to 7 are implemented.

Images