Nuclide identification method, system and electronic equipment

Abstract

The nuclide identification method, system and electronic device disclosed in the present invention relate to the field of nuclide identification. Based on the Bayesian method and sequential posterior probability, the present invention greatly improves the universality of the method by setting the value range of the time interval on the decision function instead of directly giving a preset time interval parameter; the background probability density function and the Compton probability density function in the ROI area are determined according to statistical characteristics, and the energy Bayes factor and the time Bayes factor are determined based on the energy and time interval information in the nuclear detection event sequence obtained by measurement. Combining the two factors, effective identification and differentiation of Compton plateau can be achieved. Under the same identification conditions and when the same confidence level is set, the nuclide identification method provided by the present invention can effectively identify the existence and type of radioactive nuclides faster than the energy spectrum decomposition analysis-characteristic peak matching method; compared with the existing nuclide identification methods, the nuclide identification method provided by the present invention is more universal.

Description

Technical Field

[0001] The present invention relates to the field of nuclide identification, and in particular to a nuclide identification method, system and electronic equipment. Background Art

[0002] At present, the common traditional nuclide identification methods are based on γ spectrum analysis technology and characteristic peak matching technology. These two technologies mainly include background subtraction, filtering smoothing and peak finding. They are all based on the Gaussian distribution assumption of the peak shape of the full energy peak and the matching of the peak position, statistically analyzing the characteristic γ rays released by the radioactive material, and realizing the qualitative and quantitative judgment of the radioactive material. However, these methods need to collect enough photons to reduce the statistical fluctuation of the characteristic peak, so there are certain requirements for the detection time and the intensity of the radiation release of the radioactive nuclide. In special security inspection scenarios such as ports, airports and borders, the flow of personnel and goods is large, and the detection target time can only be measured for a short time (seconds). At the same time, the detection target is in motion relative to the detector during the security inspection process, which will further affect the number of photons collected by the detector. In addition, in these scenarios, radioactive materials are usually placed in well-shielded containers or containers, which reduces gamma distortion and counting rate, and is affected by environmental background and Compton scattering, affecting the rapid and accurate identification of nuclides, and further increasing the difficulty of processing γ radiation signals. In summary, in special security inspection scenarios such as ports, airports and borders, the peak-matching nuclide identification method based on the Gaussian model becomes difficult and inaccurate when analyzing the gamma energy spectrum, and the contradiction between identification speed and accuracy is more prominent. Although the use of fixed large-volume detectors can alleviate this contradiction to a certain extent, the high price and large size limit its widespread application.

[0003] In addition, the new nuclide identification methods based on full spectrum analysis mainly include sequential Bayesian method, fuzzy mathematics method, neural network method and deconvolution method. These methods not only use characteristic peaks as useful information, but also incorporate information such as branching ratio, half-life and energy spectrum image into the analysis information to reduce the uncertainty of nuclide identification under low counting rate conditions and improve detection efficiency. The sequential Bayesian nuclide identification method has the advantages of high identification lower limit and fast identification speed. The fuzzy mathematics nuclide identification method can still accurately identify in complex environments. The neural network nuclide identification method has the characteristics of being able to simulate arbitrary functions, and can use the full spectrum information of radioactive nuclides for analysis. It can quickly and accurately identify nuclides without complex operations such as smoothing, peak finding, and peak fitting.

[0004] The detection of illegal transportation of radioactive materials in special scenarios such as airports, ports and nuclear radiation emergencies has the characteristics of non-static measurement, short detectable time and weak detected signal, which puts forward new requirements on measurement technology:

[0005] 1. The measurable time is short, usually ranging from sub-seconds to more than ten seconds.

[0006] 2. The amount of information obtained is rich, and it is necessary not only to determine the energy of the gamma rays, but also the type, intensity, dose and other information of the corresponding radioactive nuclides.

[0007] 3. The data credibility requirements are high. At a higher confidence level, the requirements for recognition efficiency and accuracy are high, while the false alarm rate and missed alarm rate should be as low as possible.

[0008] Traditionally, the γ spectrum characteristic peak matching technology is generally used to analyze radioactive nuclides, doses and other indicators. This analysis method usually requires the collection of enough photons for identification, that is, it is based on a large number of γ-ray measurement cases and long-term measurements to reduce statistical fluctuations to ensure the necessary measurement representativeness and measurement accuracy. In the radiation detection process recommended by the International Atomic Energy Agency (IAEA), there are three main categories based on this method: fixed portal radiation monitor system, handheld / portable γ spectrometer and laboratory high-resolution γ spectrometer. The newly developed fixed large-volume sodium iodide detector and high-purity germanium detector can accurately identify the type of radioactive nuclides, but the system is very expensive, poorly portable, and cannot be deployed quickly. Portable γ spectrometers mainly include scintillator γ spectrometers and semiconductor γ spectrometers. These detectors usually have low detection efficiency and require long-term measurements to collect enough photons to reduce the statistical fluctuations of the counts. Laboratory high-resolution γ spectrometers are mainly liquid nitrogen-cooled high-purity germanium γ spectrometers. The system has poor universality, cannot be deployed quickly, and is expensive. In summary, it is difficult to ensure the rapid measurement and identification of low-level nuclides in a short time using the traditional method, i.e., the method of searching from the nuclide database combined with the matching of characteristic peaks in the gamma spectrum. On the one hand, this method is difficult to meet the identification requirements of multiple nuclides in complex environments. On the other hand, it is difficult to achieve rapid nuclide identification due to the long formation time of the energy spectrum peak. At the same time, the traditional nuclide identification method based on gamma spectrum analysis and characteristic peak matching is prone to missed alarms when the total energy peak net count rate is not high enough.

[0009] In addition, the new nuclide identification methods based on full spectrum analysis mainly include the sequential Bayesian method based on Bayesian theory and sequential probability ratio test, fuzzy mathematics method, neural network method and deconvolution method, etc. These methods have some shortcomings in terms of universality on different types of γ spectrometers, the degree of influence of background and Compton plateau, false alarm rate, missed alarm rate and the amount of calculation required. Fuzzy mathematics method, neural network method and deconvolution method are based on full spectrum analysis, which requires a large number of particles and a long measurement time. At the same time, the neural network method and deconvolution method require a large amount of calculation, so they are not suitable for real-time online analysis.

[0010] The sequential Bayesian nuclide identification method based on Bayesian theory and sequential probability ratio test uses the three characteristics of radioactive nuclides, half-life, characteristic gamma ray energy and branching ratio, and updates the decision function by selecting appropriate prior functions and confidence levels, and uses sequential probability ratio tests to perform statistical inferences on hypothesis testing. However, this method requires the presupposition of some parameters related to the sample, which should be unknown quantities in the actual measurement process, which greatly limits the universality of the method; for example, due to the time interval parameter term in the test model, this method will cause the method to have a high false alarm rate for the Compton plateau generated by high-energy rays in the low-energy area of ​​interest; in addition, this method uses a Gaussian model for the distribution of the background, which is inconsistent with the actual situation. This leads to some limitations in the application of the existing sequential Bayesian method. Summary of the invention

[0011] In order to solve the above problems existing in the prior art, the present invention provides a nuclide identification method, system and electronic equipment.

[0012] To achieve the above object, the present invention provides the following solutions:

[0013] A method for nuclide identification, comprising:

[0014] Acquiring time-energy information of a ray, and determining whether the ray belongs to a certain ROI region based on energy information in the time-energy information of the ray;

[0015] When the ray belongs to a certain ROI area, describe all nuclear detection event sequences in this ROI area;

[0016] Determining an energy Bayes factor based on energy information in the described nuclear detection event sequence;

[0017] determining an energy decision function based on the energy Bayes factor;

[0018] Determine the time interval within the ROI region based on the time information in the described nuclear detection event sequence;

[0019] determining a temporal Bayes factor based on the time interval;

[0020] determining a temporal decision function based on the temporal Bayes factor;

[0021] Combining the energy decision function and the time decision function to obtain a joint decision function of the ROI area;

[0022] Based on the characteristic gamma rays corresponding to the ROI area, potential nuclides corresponding to the ROI area are retrieved, and joint decision functions of the ROI areas corresponding to the retrieved nuclides are combined to obtain a nuclide joint decision function;

[0023] The determination and identification of nuclides are completed based on the nuclide joint decision function.

[0024] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0025] The present invention is based on Bayesian factors and sequential posterior probabilities. By setting the value range of the time interval on the decision function instead of directly giving a preset time interval parameter, the universality of the method is greatly improved; the background probability density function and the Compton probability density function in the ROI area are determined according to statistical characteristics, and the energy Bayesian factor and the time Bayesian factor are determined based on the energy and time interval information in the measured nuclear detection event sequence. By combining the two factors, the nuclear detection event sequence can achieve effective identification and distinction of the Compton plateau. Under the same identification conditions and when the same confidence level is set, the nuclide identification method provided by the present invention can effectively identify the existence and type of radioactive nuclides faster than the energy spectrum decomposition analysis-characteristic peak matching method; compared with fuzzy mathematics and neural network nuclide identification methods, the nuclide identification method provided by the present invention is more universal.

[0026] In addition, compared with the existing sequential nuclide identification method based on Bayesian theory and sequential probability ratio test, the nuclide identification method provided by the present invention can significantly reduce the false alarm rate and the missed alarm rate, and can very effectively identify Compton particles from high-energy rays appearing in the low-energy energy window.

[0027] Furthermore, the present invention provides a nuclide identification system, which is used to implement the nuclide identification method provided above; the system comprises:

[0028] A ray region determination module, used for acquiring time-energy information of the ray, and determining whether the ray belongs to a certain ROI region based on energy information in the time-energy information of the ray;

[0029] A nuclear detection event sequence description module is used to describe all nuclear detection event sequences in a certain ROI area when the ray belongs to the ROI area;

[0030] An energy Bayesian factor determination module is used to determine an energy Bayesian factor based on energy information in the described nuclear detection event sequence;

[0031] An energy decision function determination module, used to determine an energy decision function based on the energy Bayes factor;

[0032] A ROI region time interval determination module is used to determine the time interval within the ROI region based on the time information in the described nuclear detection event sequence;

[0033] A temporal Bayesian factor determination module, configured to determine a temporal Bayesian factor based on the time interval;

[0034] A time decision function determination module, used to determine a time decision function based on the time Bayes factor;

[0035] A decision function joint module, used for jointly combining the energy decision function and the time decision function to obtain a joint decision function of the ROI area;

[0036] The nuclide joint decision function determination module is used to retrieve the potential nuclides corresponding to the ROI area based on the characteristic gamma rays corresponding to the ROI area, and to obtain the nuclide joint decision function by combining the joint decision functions of the ROI areas corresponding to the retrieved nuclides;

[0037] The nuclide identification module completes the determination and identification of nuclides based on the nuclide joint decision function.

[0038] Furthermore, the present invention also provides an electronic device, the electronic device comprising:

[0039] Memory for storing computer programs;

[0040] A processor is connected to the memory and is used to retrieve and execute the computer program to implement the nuclide identification method provided above.

[0041] Since the technical effects achieved by the system and electronic device provided by the present invention are the same as the technical effects achieved by the nuclide identification method provided above, they will not be described in detail here. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1The overall flow chart of the nuclide identification method provided by the present invention;

[0044] Figure 2 A flowchart of the implementation of the nuclide identification method provided by the present invention;

[0045] Figure 3 A graph showing the change of the first 59keV gamma ray time and energy decision function provided by the present invention;

[0046] Figure 4 A graph showing the change of the first 661keV gamma-ray time and energy decision function provided by the present invention;

[0047] Figure 5 The present invention provides 241 Variation diagram of the first nuclide joint decision function of Am

[0048] Figure 6 The present invention provides 137 Change diagram of the first nuclide joint decision function of Cs;

[0049] Figure 7 A second 59keV gamma ray time and energy decision function change diagram provided by the present invention;

[0050] Figure 8 A second 661keV gamma ray time and energy decision function change diagram provided by the present invention;

[0051] Fig. 9 The present invention provides 241 Variation diagram of the joint decision function of the second nuclide of Am;

[0052] Fig.10 The present invention provides 137 Variation diagram of the second nuclide joint decision function of Cs. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] The purpose of the present invention is to provide a radionuclide identification method, system and electronic equipment, which can distinguish Compton Plateau well, can more quickly and effectively identify the existence and type of radionuclides, and are more universal.

[0055] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0056] When gamma rays are incident on the sensitive volume of the gamma spectrometer measurement system, the measurement system will output the energy and time data pair (ε, t) of the rays to the host computer. The host computer calculates the decision function and completes the statistical inference based on the energy and time data pair (ε, t) of the incident rays according to the rapid nuclide identification algorithm, and makes effective inferences on the presence and type of radioactive nuclides. Based on this, Figure 1 and Figure 2 As shown, the nuclide identification method provided by the present invention includes:

[0057] Step 1: ROI region identification.

[0058] When the detector detects a ray, it outputs the time-energy information of the ray ξ (0) =(ε (0) ,τ (0) ), according to the energy information ε (0) Determine whether the ray belongs to a certain ROI area. If not, discard the ray. If the ray belongs to a certain ROI area, for example, it belongs to ROI i area, then the ROI i The entire sequence of nuclear detection events in the area It can be described as:

[0059]

[0060] ε i,j =ε (0) , τ i,j =τ (0) .

[0061] In the formula, ξ i,j Representing ROI i The time and energy information of the jth ray in the region, where ROI i represents the i-th region of interest, expressed as:

[0062] ROI i =(E i -k·σ i ,E i +k·σ i ).

[0063] Among them, E i represents the energy of the characteristic gamma ray corresponding to the i-th ROI, k is the inclusion factor, σ i is the standard deviation of the full energy peak of the characteristic γ-ray corresponding to the i-th ROI.

[0064] Step 2: Calculation of energy Bayes factor.

[0065]

[0066] in It is the ROI calculated based on the energy information of the current ray i The j-th energy Bayes factor of the region, and are the energy probability density functions under the null hypothesis H0 (no full energy peak exists) and the alternative hypothesis H1 (there is a full energy peak), and are the standard deviation of energy under the H0 and H1 assumptions, respectively, is the prior probability under the H1 hypothesis. And we have:

[0067]

[0068] Among them, Δ Ei =2·k·σ i,1 ROI i Energy width of the region, π i (σ 2 ) is the ROI i Regional energy prior probability density function, π i，0 and π i,1 are the energy prior probabilities under the H0 and H1 assumptions, respectively.

[0069] Step 3: Calculation of energy decision function.

[0070]

[0071] in, and They are the ROIs calculated based on the energy of the current ray under the assumptions of H0 and H1. i The energy posterior probability of the region is also the decision function of the gamma-ray decision.

[0072] Step 4: Update of energy prior probability.

[0073]

[0074] Step 5: ROI i Time interval calculation within the area.

[0075] Use the ROI from step 1 i Regional Kernel Detection Event Sequence Calculating ROI i The time interval within the region is:

[0076] Δt i,j =τi,j -τ i,j-1 .

[0077] Among them, τ i,j is the measurement time of the current ray, τ i,j-1 ROI i The measurement time of the last ray in the region.

[0078] Step 6: ROI i Calculation of intra-regional temporal Bayes factors.

[0079]

[0080] In the formula, is the temporal Bayes factor, g i,0 (Δt i,j |τ i,0 ) is the probability density function of the time interval under the null hypothesis M0 (no radionuclide exists), g i,1 (Δt i,j |τ i,1 ) is the probability density function of the time interval under the alternative hypothesis M1 (the presence of radionuclides), τ i,0 is the mathematical expectation of the time interval under the null hypothesis M0, τ i,1 is the mathematical expectation of the time interval under the alternative hypothesis M1, Δt i,j is the time interval, h i,1 (τ i,1 ) is the probability density function under the alternative hypothesis M1. And:

[0081]

[0082]

[0083] In the formula, g0(Δt i,j |τ0) and g1(Δt i,j |τ1) are the probability density functions of the time interval under the null hypothesis M0 (no radionuclide) and the alternative hypothesis M1 (the existence of radionuclide), τ0 and τ1 are the mathematical expectations of the time interval under the M0 and M1 assumptions, respectively, and h1(τ1) is the probability density function under the M1 assumption. and τ bkg are the background count rate and the mean of the time interval, which are reciprocals of each other. min It determines the lower limit of detection sensitivity under the M1 assumption η min Parameters. Among them, the detection sensitivity lower limit η minIt is defined as the minimum signal-to-noise ratio η required to make effective judgments under active conditions and meet certain detection performance. The signal-to-noise ratio is the ratio of the full-spectrum net count rate to the background count rate. θ(τ) is the prior probability density function of the full-spectrum time interval. θ0 and θ1 are the prior probabilities under the assumptions of M0 and M1, respectively, and θ0+θ1=1.

[0084] Step 7: Time decision function calculation.

[0085]

[0086]

[0087] in, and They are the time posterior probabilities calculated according to the time interval of the current ray under the assumptions of M0 and M1, and are also the decision functions for the γ-ray decision.

[0088] Step 8: Update of temporal prior probability.

[0089]

[0090] Step 9: Gamma ray decision making.

[0091] Combined ROI i The decision function of time and energy in the region is used to make a decision on the existence of gamma rays, specifically:

[0092]

[0093] in, and are the upper and lower thresholds of the energy decision function, and are the upper and lower thresholds of the temporal decision function respectively.

[0094]

[0095] Among them, α π , β π are the probabilities of making the first and second type errors when making energy threshold decisions, respectively, where α θ , β θ The probabilities of making type I and type II errors when making decisions at the time threshold are respectively.

[0096] Step 10: Radionuclide I x Calculation of joint decision functions.

[0097]

[0098] Among them, P XIndicates nuclide I X The decision function, m represents the nuclide I X Characteristic gamma ray number, λ m represents the weight coefficient of the corresponding γ-ray, and Respectively represent the decision functions of the energy and time corresponding to the ray, η m and κ m are the branching ratio and intrinsic detection efficiency of γ-rays, respectively, and N is the nuclear I x The total number of gamma rays.

[0099] Step 11: Radionuclide decision making.

[0100] The decision made from the above is:

[0101]

[0102] Among them, P X,up_th and P X,low_th are the upper and lower thresholds respectively.

[0103] Based on the above description, compared with the prior art, the present invention has the following characteristics:

[0104] 1) The probability density function construction of the null hypothesis and the alternative hypothesis is one of the key points of the present invention. In the present invention, the hypothesis of whether there is a radionuclide is converted into whether there is a total energy peak in the energy window based on the monoenergetic decomposition theory and certain transformations.

[0105] 2) The method for constructing the probability density function of energy-related statistics is one of the key points of the present invention. In the actual process, the probability density function of energy is actually a linear superposition of Gaussian distribution, uniform distribution and triangular distribution according to their respective count rate shares, rather than being limited to the Gaussian distribution and uniform distribution in the examples of the present invention. The specific probability density distribution can be determined according to the actual use scenario.

[0106] 3) The construction of the prior probability of energy-related statistics and time-related statistics is one of the key points of the present invention. In the present invention, the prior probability density function can select the non-information prior probability density function or the conjugate prior probability density function, which is not limited to the non-information prior probability density function in the example of the present invention. At the same time, the initial value of the prior probability can be determined according to the actual situation, including but not limited to the parameters (0.5, 0.5) in the example of the present invention.

[0107] 4) The construction structure of the decision function of energy-related statistics and time-related statistics is one of the key points of the present invention. In the present invention, the decision function can have two construction paths: a) the decision function is obtained according to the prior probability and the sample probability density function; b) during the construction process, the decision function can also first calculate the Bayes factor according to the sample probability density function, and then jointly construct it according to the aforementioned prior probability and Bayes factor. Regardless of the path, the decision function structure constructed according to the aforementioned key information is the key point to be protected in this case. [The Bayes factor is an intermediate parameter and may not participate in the actual construction process]

[0108] 5) Solving the decision function is one of the key points of the present invention, whether it is calculated based on the prior probability and the Bayes factor, or directly calculated based on the prior information and sample information. In the process of calculating based on the prior probability and the Bayes factor, the Bayes factor is an intermediate parameter and may not be involved in the calculation in the actual calculation process.

[0109] 6) The construction method of finally performing threshold discrimination when performing statistical inference of energy-related statistics and time-related statistics is one of the key points of the present invention. In the conventional Bayesian method, the relative size of the posterior probability of the null hypothesis and the alternative hypothesis is generally selected (and the upper and lower thresholds are both 0.5), that is, the probability ratio test. When statistical decision-making is performed, when the posterior probabilities of the null hypothesis and the alternative hypothesis are relatively close, there is a high probability of making an incorrect decision. In the present invention, the decision functions of energy and time interval are set with upper and lower thresholds of different values ​​(such as in this example, the judgment thresholds for the energy window are 0.8 and 0.2, which are examples for example, and do not represent the specific values ​​of the protection points of the present invention), which greatly improves the accuracy of the decision results. At the same time, the introduction of decision thresholds for two types of decisions (supporting the null hypothesis and supporting the alternative hypothesis) of the decision function is the protection point of the present invention. As for the selection of the specific value of the threshold, it can be determined according to the actual use scenario.

[0110] 7) One of the key points of the present invention is to use the joint discrimination result based on the energy-related statistics and time-related statistics of the incoming gamma ray as the discrimination result of the gamma ray. Combined with the content in 6), although there are also methods that use the combination of time and energy to identify nuclides, the probability density function and decision function used are different from those used in the present invention.

[0111] 8) Constructing a decision function for a nuclide based on the product of all gamma-ray decision functions of the radioactive nuclide and the constructed weight coefficient is one of the key points of the present invention. The prior art does not disclose the same or similar method.

[0112] 9) In the statistical decision of the decision function, the upper and lower thresholds are set to different values, which greatly improves the accuracy of the decision result. At the same time, the introduction of the decision thresholds for the two types of decisions (supporting the null hypothesis and supporting the alternative hypothesis) of the decision function is one of the key points of the present invention. The selection of the specific value of the threshold can be determined according to the actual use scenario.

[0113] 10) In constructing the weight coefficient of the decision function, the normalized result of the product of the ray branching ratio and the intrinsic detection efficiency in the corresponding nuclide is used as the weight coefficient in the nuclide.

[0114] Below is the background and 137 Cs is taken as an example to illustrate the specific implementation process of the nuclide identification method provided by the present invention. Among them, a lanthanum bromide γ spectrometer is used as an example, and any γ spectrometer can be selected for the actual operation process.

[0115] Example 1

[0116] In this example, a 1.5-inch lanthanum bromide detection system is used, and the energy scale and half-height width scale are:

[0117] E=0.07525·ch+0.9285(R 2 =1).

[0118]

[0119] A nuclide library was constructed containing 22 nuclides and a total of 46 characteristic gamma rays. Accordingly, 46 ROIs were delineated in the full spectrum. The ROI was taken as one-tenth of the height-width interval FWTM. At this time, the inclusion factor k = 2.146, the confidence level is about 96.8%, and the ROI i The energy width of the region ΔE i =4.296*σ i . And collect background data, calibrate the background count rate of the whole spectrum and background count rates in 46 ROI areas

[0120] Since the prior probability is unbiased, the initial value of the prior probability is set to (0.5, 0.5). i The detection sensitivity lower limit of the region is 30%, and there is τ i,min =50%·τ i,bkg , set α π =β π =0.2, α θ =β θ =0.2, the upper and lower thresholds of decision making are P X,up_th=0.64, P X,low_th =0.04.

[0121] Based on this, background identification is carried out.

[0122] The detector detects the first ray, and the time and energy information are (0.045657025, 44) / (time / second, energy / channel address).

[0123] Based on step 1 provided above in the present invention, it is obtained that the ray belongs to ROI1 (33, 51) corresponding to the 32keV ray and ROI2 (38, 58) corresponding to the 35keV ray.

[0124] The energy Bayes factors of ROI1 and ROI2 are calculated by step 2 respectively, which are:

[0125] The energy decision functions of ROI1 and ROI2 are calculated by step 3 respectively, which are:

[0126] Update the energy prior probability of ROI1 and ROI2 respectively by step 4: π 1,0 =0.3909, π 2,0 =0.4464.

[0127] The time interval, time Bayes factor, and time decision function of ROI1 and ROI2 are calculated by steps 5 to 8 respectively, as follows:

[0128] Δt1=0.0457、

[0129] Δt2=0.0457、

[0130] The decision on the 32keV characteristic gamma ray corresponding to ROI1 and the 35keV characteristic gamma ray corresponding to ROI2 is made in step 9 as follows:

[0131] because and ROI1 does not make a decision and waits for the next ray. and ROI2 does not make a decision and waits for the next ray.

[0132] Assume that step 9 determines that the 32keV characteristic gamma ray corresponding to ROI1 exists or that ROI1 is entirely background signal. 133 Ba, then it is necessary to examine the radionuclides 133 Does Ba exist?

[0133] From step 10, use ROI1, ROI5, ROI corresponding to 32keV, 81keV, 276keV, 302keV, 356keV and 383ke characteristic gamma rays 17 ROI 18 ROI 22 and ROI 23 Time and energy decision functions to calculate nuclides 133 Ba joint decision function.

[0134] From step 11, according to 133 The joint decision function of Ba is used to identify the radionuclides.

[0135] When the next ray arrives, repeat steps 1 to 8 according to the time and energy information of the ray to calculate the time and energy decision function of the corresponding ROI. Then repeat step 9 to identify the characteristic gamma ray. Finally, repeat step 10 to identify the radioactive nuclide.

[0136] In this example, a total of 10 4 The total time for the ray particles was 166.25 seconds, and the full spectrum count rate was 60.15s -1 .

[0137] To explain the nuclide identification process in detail, the 3424th ray detected by the detector is taken as an example (i.e., the first time that the 661keV characteristic γ ray and the corresponding nuclide are not present in the following 137 Cs decision time), in order to observe the changes of the γ-ray decision function and the nuclide joint decision function before and after the ray.

[0138] The detector detects the 3424th ray, and the time and energy information is (57.85114355, 881) / (time / second, energy / channel address). At this time, the prior probability of energy and time has been updated to π 31,0 =0.999975,θ 31,0 =0.6936;

[0139] From step 1, we can see that this ray belongs to the ROI corresponding to the 661keV ray. 31 (857,900).

[0140] Calculate ROIs from step 2 31 The energy Bayes factor of is:

[0141] Calculate ROI from step 3 31 The energy decision function is:

[0142] Update ROIs separately from step 4 31The energy prior probability is: π 31,0 =0.999960.

[0143] Calculate ROI from steps 5 to 8 31 The time interval, time Bayes factor, and time decision function are:

[0144] Δt 31 =0.0042,

[0145] From step 9, ROI 31 The corresponding 661keV characteristic gamma ray makes a decision: and Making ROI 31 All signals come from the judgment of background. 31 The characteristic energy of the signal is 661keV, and the corresponding characteristic gamma ray belongs to the nuclide 137 Cs, then the nuclides need to be calculated 137 The joint decision function of Cs for nuclides 137 Whether Cs exists is determined and updated.

[0146] Calculate the nuclide from step 10 137 The joint decision function of Cs is:

[0147] According to step 11, 137 Make a decision whether Cs exists: Make nuclide 137 Decision judgment that Cs does not exist (due to nuclide 137 Cs has only one 661keV characteristic gamma ray, so make a ROI in step 8 31 When the signal comes entirely from the background, the nuclide 137 Decision judgment that Cs does not exist, here the nuclide is calculated 137 The joint decision function of Cs is only used to illustrate the detailed process of nuclide identification.

[0148] For nuclides 241 Am and its 59keV characteristic gamma ray were both judged as non-existent at 14.9575 seconds (881 total samples, 17 ROI4 samples). 137 Cs has a 661keV characteristic γ-ray of 57.8611 seconds (3424 total samples, 28 ROIs 31 The two γ-ray decision functions change with time as shown in Figure 3 and Figure 4 shown. 241 Am and 137The change of the nuclide joint decision function of Cs over time is as follows Figure 5 and Figure 6 As shown. Among them, Figure 3 The final recognition result is: there is no 59.54keV γ ray, the time is 14.9757s, the number of full spectrum rays is 881, and the number of ROI area rays is 17. Figure 4 The final recognition result is: there is no 661keV γ-ray. The time is 57.8511s, the number of full spectrum rays is 3424, and the number of ROI area rays is 28. Figure 5 In the final identification result, the nuclide Am-241 does not exist. Figure 6 The final identification result is: the nuclide Cs-137 is not present.

[0149] In addition, under background conditions, no radionuclide false alarms occurred (false alarms are defined as radionuclides that do not exist but are judged to exist by the decision). When the above two characteristic gamma rays are judged to be non-existent and the corresponding signals in the ROI are all from the background, the time required is too long because the rate of collecting samples in the ROI corresponding to the characteristic gamma rays under background conditions is too low. When making an effective judgment, it can be found that a small amount of samples collected in the ROI can make an effective identification.

[0150] Example 2

[0151] The components and parameters used in this embodiment are the same as those in the above-mentioned embodiment 1. 137 Cs(9.22*10 3 Bq) (source-detector front face distance 35cm) (equivalent dose rate is about 5.52nGy / h) identification result.

[0152] The detector detects the first ray, and the time and energy information are (0.030928838, 161) / (time / second, energy / channel address).

[0153] From step 1, it can be obtained that the ray belongs to ROI8 (150, 171) corresponding to the 122keV ray.

[0154] Calculate the energy Bayes factor of ROI8 region by step 2:

[0155] The energy decision function of ROI8 is calculated by step 3:

[0156] Update the energy prior probability of ROI8 area by step 4: π 1,0 =0.3849.

[0157] The time interval, time Bayes factor, and time decision function of ROI8 are calculated from steps 5 to 8:

[0158] Δt1=0.0309、

[0159] Step 9 makes a decision on the 122keV ray corresponding to ROI8: and ROI8 does not make a decision and waits for the next ray.

[0160] If step 8 determines that gamma rays exist, assuming that 122keV rays corresponding to ROI8 exist, then the radionuclides need to be examined separately. 57 Co and 152 Does Eu exist? Specific:

[0161] From step 10, use 122keV radiation and 136keV radiation to correspond to ROI8 and ROI 10 Time and energy decision functions for regions, calculation of nuclides 57 Co joint decision function. Use ROI8, ROI corresponding to 122keV, 244keV, 344keV, 779keV, 964keV and 1408keV 16 ROI 21 ROI 32 ROI 36 and ROI 42 Time and energy decision functions for regions, calculation of nuclides 152 Eu joint decision function.

[0162] From step 11, according to 57 Co and 152 Eu's joint decision function makes decisions on the hypothesis test of the existence of two radioactive nuclides.

[0163] When the next ray arrives, repeat steps 1-8 according to the time and energy information of the ray to calculate the time and energy decision function of the corresponding ROI area. Then repeat step 9 to identify the gamma ray. Finally, repeat steps 10-11 to identify the radioactive nuclide.

[0164] In this example, a total of 10 4 The total time for the ray particles was 153.70 seconds, and the full spectrum count rate was 64.14s -1 .

[0165] To illustrate the nuclide identification process in detail, the 281st ray detected by the detector is taken as an example (i.e., the first occurrence of the 661keV characteristic γ-ray and nuclide in the following 137 Cs decision time), in order to observe the changes of the γ-ray decision function and the nuclide joint decision function before and after the ray.

[0166] The detector detects the 281st ray, and the time and energy information is (4.260600813, 865) / (time / second, energy / channel address). At this time, the prior probability of energy and time has been updated to π 31,0 =0.1818,θ 31,0 =0.2220.

[0167] From step 1, we get the ROI corresponding to the 661keV ray. 31 (857,900).

[0168] Calculate ROIs from step 2 31 The energy Bayes factor of is:

[0169] Calculate ROI from step 3 31 Energy decision function:

[0170] Update ROIs from step 4 31 Energy prior probability: π 31,0 =0.1818.

[0171] Calculate ROI from steps 5 to 8 31 Time interval, time Bayes factor, time decision function:

[0172] Δt 31 =0.0075,

[0173] From step 9, ROI 31 Make a decision corresponding to the 661keV characteristic gamma ray: and The judgment is made that there is a 661keV characteristic gamma ray. 31 The 661keV characteristic gamma ray corresponding to the signal belongs to the nuclide 137 Cs, then the nuclides need to be calculated 137 The joint decision function of Cs for nuclides 137 Whether Cs exists is determined and updated.

[0174] Calculate the nuclide from step 10 137 The joint decision function of Cs:

[0175] According to step 11, 137 Make a decision whether Cs exists: Make nuclide 137 The decision to determine the presence of Cs. 137Cs has only one 661keV characteristic gamma ray, so make a ROI in step 11 31 When the signal comes entirely from the background, the nuclide 137 Decision judgment that Cs does not exist, here the nuclide is calculated 137 The joint decision function of Cs is only used to illustrate the detailed process of nuclide identification.

[0176] For nuclides 241 Am and its 59keV characteristic gamma ray were both judged as non-existent at 72.6919 seconds (4613 total samples, 98 ROI4 samples). 137 Cs and its 661keV characteristic gamma ray, at 4.2606 seconds (281 total samples, 7 ROIs 31 The two γ-ray decision functions change with time as shown in Figure 7 and Figure 8 shown. 241 Am and 137 The change of the nuclide joint decision function of Cs over time is as follows Fig. 9 and Fig.10 As shown. Among them, Figure 7 The final recognition result is: there is no 59.54keV γ-ray. The time is 72.6919s, the number of full spectrum rays is 4613, and the number of ROI area rays is 98. Figure 8 The final identification result is: there is 661.7keV γ-ray. It takes 4.2606s, 281 full spectrum rays, and 7 ROI area rays. Fig. 9 The final identification result: No nuclide exists 241 Am. Fig.10 The final recognition result: exists 137 Cs.

[0177] By comparing the nuclide identification process under the above background conditions (i.e., Example 1 and Example 2), it can be found that:

[0178] In the background and 137 Under the condition of Cs radiation source, 241 The characteristic gamma ray decision function and the nuclide joint decision function of Am are both close to 1, making the characteristic gamma ray and 241 The judgment of Am nuclide. 137 Cs radiation source conditions, no false alarms occurred. 137 The 661keV rays of Cs 241 The Compton count of ROI4 corresponding to 59keV of Am only affects its convergence speed to 1, but does not affect its accuracy (the results of other characteristic gamma rays and corresponding nuclides not corresponding to 661keV are similar to 241The judgment and recognition results of Am are the same).

[0179] Under background conditions, 137 The characteristic gamma ray decision function of Cs quickly approaches 1, making a decision of no characteristic gamma ray and making no 137 Nuclide determination of Cs nuclide. 137 Under the Cs nuclide condition, the characteristic gamma ray decision function and the nuclide joint decision function both quickly approach 0, making a characteristic gamma ray decision and a 137 The nuclide judgment of Cs nuclide. It can be seen that this example realizes 137 Rapid identification of Cs nuclides (identification time of 4.2606 seconds, 281 total samples, equivalent dose rate of approximately 5.52 nGy / h).

[0180] The nuclide identification method of the present invention is based on the Bayesian principle and sequential posterior probability. By replacing the test model of the nuclide identification method, the value range of the time interval is set on the prior probability, rather than directly giving a preset fixed value for the time interval parameter, so that the universality of the method is greatly improved. The background probability density function and the Compton probability density function in the ROI area are determined according to statistical characteristics, and the energy Bayes factor and the time Bayes factor are determined based on the energy and time interval information in the measured nuclear detection event sequence. Combining the two factors, effective identification and differentiation of the Compton plateau can be achieved. Under the same identification conditions, when the same confidence level is set, the method of this study can effectively identify the existence and type of radioactive nuclides faster than the energy spectrum decomposition analysis-characteristic peak matching method. Compared with fuzzy mathematics and neural network nuclide identification methods, the method of this study is more universal.

[0181] In addition, compared with the sequential nuclide identification method based on Bayesian theory and sequential probability ratio test published in the research, in order to solve the problems of inability to distinguish Compton plateau and poor universality in previous studies, the sequential Bayesian nuclide identification method based on Bayesian method and sequential posterior probability in the present invention can be found, according to the above examples, that the false alarm rate and the missed alarm rate are very low, and it is shown that Compton particles from high-energy rays appearing in the low-energy energy window can be very effectively identified.

[0182] Furthermore, the present invention provides a nuclide identification system, which is used to implement the nuclide identification method provided above. The system includes: a ray region determination module, a nuclear detection event sequence description module, an energy Bayesian factor determination module, an energy decision function determination module, a ROI region time interval determination module, a time Bayesian factor determination module, a time decision function determination module, and a decision function joint module. 、 Nuclide joint decision function determination module and nuclide identification module.

[0183] The ray region determination module is used to obtain the time-energy information of the ray and determine whether the ray belongs to a certain ROI region based on the energy information in the time-energy information of the ray.

[0184] The nuclear detection event sequence description module is used to describe all nuclear detection event sequences in a certain ROI area when the ray belongs to the ROI area.

[0185] The energy Bayes factor determination module is used to determine the energy Bayes factor based on the energy information in the described nuclear detection event sequence.

[0186] The energy decision function determination module is used to determine the energy decision function based on the energy Bayes factor.

[0187] The ROI region time interval determination module is used to determine the time interval within the ROI region based on the time information in the described nuclear detection event sequence.

[0188] The temporal Bayesian factor determination module is used to determine the temporal Bayesian factor based on the time interval.

[0189] The time decision function determination module is used to determine the time decision function based on the time Bayes factor.

[0190] The decision function joint module is used to jointly combine the energy decision function and the time decision function to obtain the joint decision function of the ROI area.

[0191] The nuclide joint decision function determination module is used to retrieve the potential nuclides corresponding to the ROI area based on the characteristic gamma rays corresponding to the ROI area, and to obtain the nuclide joint decision function by combining the joint decision functions of each ROI area corresponding to the retrieved nuclides.

[0192] The nuclide identification module completes the judgment and identification of nuclides based on the nuclide joint decision function.

[0193] Furthermore, the present invention also provides an electronic device, which includes: a memory and a processor.

[0194] Memory, used to store computer programs.

[0195] The processor is connected to the memory and is used to retrieve and execute the computer program to implement the nuclide identification method provided above.

[0196] Among them, when the computer program in the above-mentioned memory is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk.

[0197] Further, based on the above description, (1) the nuclide identification method provided by the present invention can be used in conjunction with various types of detectors, including but not limited to scintillator detectors, semiconductor detectors and other detectors with energy resolution capabilities. (2) The number and type of nuclides in the nuclide standard library, the number and importance of gamma rays selected for each nuclide and other parameters are not limited to the form in the example of the present invention. (3) The probability density of energy and time interval in the nuclide identification process is not limited to the Gaussian distribution and uniform distribution in the example of the present invention, and the specific probability density distribution can be determined according to the actual use scenario. (4) The initial values ​​of the prior probability of the null hypothesis and the alternative hypothesis in the nuclide identification process can be determined according to actual conditions, including but not limited to the parameters (0.5, 0.5) in the example of the present invention. (5) The prior probability density function can select an uninformative prior probability density function or a conjugate prior probability density function, and is not limited to the uninformative prior probability density function in the example of the present invention. (6) The selection of the time interval value range in the prior probability is related to the background conditions and the detector type, including but not limited to the parameters in the example of the present invention. (7) The upper and lower thresholds in the nuclide identification process are related to the false alarm rate and the missed alarm rate, and their settings can be set according to the actual use scenario and requirements, including but not limited to the parameters in the examples of the present invention. (8) For the division of the energy window in the full energy peak region, the size of the k value can be set according to the specific situation, including but not limited to the parameters in the examples of the present invention. (9) The decision function in the present invention is solved based on the prior probability and the Bayesian factor. The Bayesian factor is an intermediate parameter and may not be involved in the calculation during actual use.

[0198] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0199] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for nuclide identification, characterized in that: include: Acquiring time-energy information of a ray, and determining whether the ray belongs to a certain ROI region based on energy information in the time-energy information of the ray; When the ray belongs to a certain ROI area, describe all nuclear detection event sequences in this ROI area; The energy Bayes factor is determined based on the energy information in the described nuclear detection event sequence; the energy Bayes factor is: In the formula, is the jth energy Bayes factor of the i-th ROI region, ε i,j is the energy information of the current ray, is the energy probability density function under the null hypothesis H0, is the energy probability density function under the alternative hypothesis H1, is the standard deviation of energy under the null hypothesis H0, is the standard deviation of energy under the alternative hypothesis H1, For the alternative hypothesis H1 The probability density function of determining an energy decision function based on the energy Bayes factor; Determine the time interval within the ROI region based on the time information in the described nuclear detection event sequence; A temporal Bayes factor is determined based on the time interval; the temporal Bayes factor is: In the formula, is the temporal Bayes factor, g i,0 (Δt i,j |τ i,0 ) is the probability density function of the time interval under the null hypothesis M0, g i,1 (Δt i,j |τ i,1 ) is the probability density function of the time interval under the alternative hypothesis M1, τ i,0 is the mathematical expectation of the time interval under the null hypothesis M0, τ i,1 is the mathematical expectation of the time interval under the alternative hypothesis M1, h i,1 (τ i,1 ) is the probability density function under the alternative hypothesis M1, Δt i,j is the time interval; determining a temporal decision function based on the temporal Bayes factor; Combining the energy decision function and the time decision function to obtain a joint decision function of the ROI area; Based on the characteristic gamma rays corresponding to the ROI area, potential nuclides corresponding to the ROI area are retrieved, and joint decision functions of the ROI areas corresponding to the retrieved nuclides are combined to obtain a nuclide joint decision function; The determination and identification of nuclides are completed based on the nuclide joint decision function.

2. The nuclide identification method according to claim 1, characterized in that: The energy decision function of the ROI area is: In the formula, is the energy posterior probability of the i-th ROI region under the original hypothesis H0, is the energy posterior probability of the i-th ROI region under the alternative hypothesis H1, π i,0 is the energy prior probability of the i-th ROI region under the null hypothesis H0.

3. The nuclide identification method according to claim 2, characterized in that: After determining the energy decision function based on the energy Bayes factor, the method further includes: The energy decision function is used to update the energy prior probability.

4. The nuclide identification method according to claim 1, characterized in that: The time decision function is: In the formula, is the temporal posterior probability of the i-th ROI region under the null hypothesis M0, is the temporal posterior probability of the i-th ROI region under the alternative hypothesis M1, θ i,0 is the time prior probability of the i-th ROI region under the null hypothesis M0.

5. The nuclide identification method according to claim 4, characterized in that: After determining the time decision function based on the time Bayes factor, the method further includes: The temporal prior probability is updated based on the temporal decision function.

6. The nuclide identification method according to claim 1, characterized in that: Before the energy decision function and the time decision function are combined to obtain a joint decision function, the method further includes: The ray signal recognition result in the ROI area is obtained based on the relationship between the calculation result of the energy decision function and the upper and lower thresholds of the energy decision function and the relationship between the calculation result of the time decision function and the upper and lower thresholds of the time decision function.

7. A nuclide identification system, characterized in that: The system is used to implement the nuclide identification method according to any one of claims 1 to 6; the system comprises: A ray region determination module, used for acquiring time-energy information of the ray, and determining whether the ray belongs to a certain ROI region based on energy information in the time-energy information of the ray; A nuclear detection event sequence description module is used to describe all nuclear detection event sequences in a certain ROI area when the ray belongs to the ROI area; The energy Bayes factor determination module is used to determine the energy Bayes factor based on the energy information in the described nuclear detection event sequence; the energy Bayes factor is: In the formula, is the jth energy Bayes factor of the i-th ROI region, ε i,j is the energy information of the current ray, is the energy probability density function under the null hypothesis H0, is the energy probability density function under the alternative hypothesis H1, is the standard deviation of energy under the null hypothesis H0, is the standard deviation of energy under the alternative hypothesis H1, For the alternative hypothesis H1 The probability density function of An energy decision function determination module, used to determine an energy decision function based on the energy Bayes factor; A ROI region time interval determination module is used to determine the time interval within the ROI region based on the time information in the described nuclear detection event sequence; A time Bayes factor determination module is used to determine a time Bayes factor based on the time interval; the time Bayes factor is: In the formula, is the temporal Bayes factor, g i,0 (Δt i,j |τ i,0 ) is the probability density function of the time interval under the null hypothesis M0, g i,1 (Δt i,j |τ i,1 ) is the probability density function of the time interval under the alternative hypothesis M1, τ i,0 is the mathematical expectation of the time interval under the null hypothesis M0, τ i,1 is the mathematical expectation of the time interval under the alternative hypothesis M1, h i,1 (τ i,1 ) is the probability density function under the alternative hypothesis M1, Δt i,j is the time interval; A time decision function determination module, used to determine a time decision function based on the time Bayes factor; A decision function joint module, used for jointly combining the energy decision function and the time decision function to obtain a joint decision function of the ROI area; The nuclide joint decision function determination module is used to retrieve the potential nuclides corresponding to the ROI area based on the characteristic gamma rays corresponding to the ROI area, and to obtain the nuclide joint decision function by combining the joint decision functions of the ROI areas corresponding to the retrieved nuclides; The nuclide identification module completes the determination and identification of nuclides based on the nuclide joint decision function.

8. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, connected to the memory, is used to call and execute the computer program to implement the nuclide identification method according to any one of claims 1 to 6.

Images