Portable multi-channel energy spectrum acquisition instrument, radioactive signal detection system, detection method and model training method

Abstract

The application discloses a portable multi-channel energy spectrum acquisition instrument, a radioactive signal detection system, a detection method and a model training method. The portable multi-channel energy spectrum acquisition instrument comprises a peak arrival timing unit, an input end of the peak arrival timing unit is connected with a pulse signal input end, and a trigger signal is generated according to a peak point of the pulse signal; a filtering unit, an input end of the filtering unit is connected with the pulse signal input end; a sampling unit, a first input end of the sampling unit is connected with an output end of the filtering unit; a single-chip microcomputer, a first input end of the single-chip microcomputer is connected with an output end of the peak arrival timing unit, a second input end of the single-chip microcomputer is connected with an output end of the sampling unit, and a first output end of the single-chip microcomputer is connected with a second input end of the sampling unit. According to the embodiment of the application, the design cost of the multi-channel energy spectrum acquisition instrument is reduced, miniaturization is realized, and the detection efficiency of the environmental radioactivity level is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of radioactive element detection, and particularly relates to a portable multi-channel energy spectrum acquisition instrument, a radioactive signal detection system, a detection method and a model training method. BACKGROUND

[0002] The digital multi-channel energy spectrum acquisition instrument is a device for nuclear radiation energy spectrum analysis, and is widely applied to the fields of nuclear physics experiment, radiation monitoring and nuclear medicine.

[0003] At present, when the radioactivity level is detected in the field, some sampling points are usually set, and then multi-channel energy spectrum acquisition instruments are laid at the positions of the sampling points to collect radioactive signals through the multi-channel energy spectrum acquisition instruments. However, since the existing digital multi-channel energy spectrum acquisition instrument usually uses a preamplifier circuit to amplify the weak current pulse signal output by a detector (such as a NaI scintillation detector) and convert the weak current pulse signal into a voltage signal, and then uses a high-speed analog-to-digital converter to convert the nuclear pulse signal into a digital signal and send the digital signal to a field programmable gate array (FPGA) in real time to realize digital shaping and multi-channel analysis of the nuclear pulse.

[0004] In the related art, since the digital multi-channel energy spectrum acquisition instrument uses a high-speed analog-to-digital converter and an FPGA with high hardware cost, not only is the area occupied in the circuit design larger, which leads to difficulty in miniaturization of the digital multi-channel energy spectrum acquisition instrument, but also the overall hardware cost is high, especially in the application field of radioactivity level detection in the field, which is limited by the cost, and the number of laid digital multi-channel energy spectrum acquisition instruments is small, which leads to low detection efficiency. SUMMARY

[0005] The application embodiment provides a portable multi-channel energy spectrum acquisition instrument, a radioactive signal detection system, a detection method and a model training method, which can reduce the design cost of the multi-channel energy spectrum acquisition instrument, realize miniaturization, and improve the detection efficiency of the environmental radioactivity level.

[0006] In one aspect, the application embodiment provides a portable multi-channel energy spectrum acquisition instrument, comprising:

[0007] a peak reaching timing unit, an input end of the peak reaching timing unit being connected with the pulse signal input end, for detecting a peak point of the pulse signal and generating a trigger signal according to the peak point of the pulse signal;

[0008] a filtering unit, an input end of the filtering unit being connected with the pulse signal input end, for filtering the pulse signal;

[0009] a sampling unit, a first input end of the sampling unit being connected with an output end of the filtering unit;

[0010] a single-chip microcomputer, a first input end of the single-chip microcomputer being connected with an output end of the peak-reach timing unit, a second input end of the single-chip microcomputer being connected with an output end of the sampling unit, and a first output end of the single-chip microcomputer being connected with a second input end of the sampling unit; the single-chip microcomputer being configured to control the sampling unit to collect energy of the filtered pulse signal according to the trigger signal to obtain first energy information.

[0011] Optionally, the peak-reach timing unit comprises:

[0012] an active differential module, an input end of the active differential module being connected with the pulse signal input end, and the active differential module being configured to convert the received pulse signal into a bipolar pulse signal;

[0013] a comparator, a first input end of the comparator being connected with an output end of the active differential module, an output end of the comparator being connected with the first input end of the single-chip microcomputer, and a second input end of the comparator being grounded;

[0014] the comparator being configured to perform zero-crossing comparison on the bipolar pulse signal, and output a trigger signal to the single-chip microcomputer in the case that the bipolar pulse signal occurs level inversion.

[0015] Optionally, the active differential module comprises a first capacitor C1, a first resistor R1 and a first amplifier U1.

[0016] a first end of the first capacitor C1 being connected with the pulse signal input end, a second end of the first capacitor C1 being connected with a first end of the first resistor R1 and an input end of the first amplifier U1 respectively, a second input end of the first amplifier U1 being grounded, and an output end of the first amplifier U1 and a second end of the first resistor R1 being connected with input ends of the comparator.

[0017] Optionally, the filter unit comprises:

[0018] an active low-pass module, an input end of the active low-pass module being connected with the pulse signal input end, and an output end of the active low-pass module being connected with the sampling unit.

[0019] Optionally, the portable multi-channel energy spectrum acquisition instrument further comprises:

[0020] a memory, the memory being connected with the single-chip microcomputer.

[0021] the memory being configured to store the first energy information of the pulse signal output by the single-chip microcomputer.

[0022] Optionally, the portable multi-channel energy spectrum acquisition instrument further comprises:

[0023] A communication module is connected with the single-chip microcomputer, and is configured to enable the single-chip microcomputer to communicate with a host computer.

[0024] In another aspect, the embodiments of the present application provide a radioactive signal detection system, which comprises a plurality of detectors, a processor, and a plurality of portable multi-channel energy spectrum acquisition instruments according to the first aspect.

[0025] The detectors are connected with the portable multi-channel energy spectrum acquisition instruments one by one, and the processor is connected in communication with the plurality of portable multi-channel energy spectrum acquisition instruments.

[0026] In another aspect, the embodiments of the present application provide a radioactive signal detection method, which is applied to a single-chip microcomputer of a portable multi-channel energy spectrum acquisition instrument according to the first aspect, and the method comprises the following steps.

[0027] At least one first pulse signal in a to-be-detected region is acquired.

[0028] The at least one first pulse signal is input into a first preset detection model, and the first preset detection model is used to detect pulse abnormalities of the at least one first pulse signal, so as to obtain at least one second pulse signal, wherein the second pulse signal comprises pulse signals in which abnormal pulses in the first pulse signal are removed.

[0029] The number of peak points and the peak values of the second pulse signal are counted, so as to obtain first energy information of the second pulse signal.

[0030] Based on the first energy information of the second pulse signal, a radioactive signal energy spectrum of the to-be-detected region is generated.

[0031] Optionally, a plurality of portable multi-channel energy spectrum acquisition instruments and a plurality of detectors are arranged in the to-be-detected region, and each of the portable multi-channel energy spectrum acquisition instruments is connected in communication with each other.

[0032] The generating of the radioactive signal energy spectrum of the to-be-detected region based on the first energy information of the second pulse signal comprises the following steps.

[0033] The spatial position information of each detector is acquired.

[0034] The second pulse signal and the spatial position information are input into a second preset detection model, and the second preset detection model outputs a radioactive level of the to-be-detected region according to the second pulse signal and the spatial position information.

[0035] Optionally, the second preset detection model comprises a first convolutional neural network, a second graph neural network, and a third long short-term memory network.

[0036] The second preset detection model outputs a radioactivity level of the to-be-detected region according to the second pulse signal and the spatial position information, and the radioactivity level of the to-be-detected region comprises:

[0037] The first convolutional neural network performs time feature extraction on the first energy information to obtain a time feature vector of the first energy information, and the time feature vector represents time sequence feature data of the first energy information.

[0038] The second graph neural network takes the spatial position information as node positions and takes the time feature vector as node attributes to construct a space-time graph of the second pulse signal.

[0039] The third long short-term memory network predicts energy information output by other portable multi-channel energy spectrum acquisition instruments according to the space-time graph to obtain second energy information.

[0040] According to the second energy information and the first energy information, a radioactivity signal spectrum of the to-be-detected region is determined.

[0041] In still another aspect, an embodiment of the present application provides a model training method, and the method comprises:

[0042] A training data set is obtained, and the training data set comprises pulse signals in multiple environments and corresponding actual signal types of the pulse signals, and the actual signal types comprise normal pulse signals and abnormal pulse signals.

[0043] The pulse signals are taken as inputs of the preset neural network model, the actual signal types are taken as outputs of the preset neural network model, the preset neural network model is trained, and a first preset detection model for performing pulse anomaly detection on at least one first pulse signal to obtain at least one second pulse signal is obtained.

[0044] Optionally, the pulse signals are taken as inputs of the preset neural network model, the actual signal types are taken as outputs of the preset neural network model, the preset neural network model is trained, and the first preset detection model is obtained, and the method comprises:

[0045] The pulse signals are input into a preset neural network model to obtain a first predicted signal type.

[0046] According to the first predicted signal type and the actual signal type, a loss function of the preset neural network model is determined.

[0047] In a case where the loss function does not satisfy the preset convergence condition, the model parameters of the preset neural network model are adjusted, and the inputting of the pulse signal into the preset neural network model to obtain the first predicted signal type is returned until the loss function satisfies the preset convergence condition, and a first preset detection model is obtained.

[0048] Optionally, after the inputting of the pulse signal as the input of the preset neural network model, the actual signal type as the output of the preset neural network model, the training of the preset neural network model to obtain the first preset detection model, the method further comprises:

[0049] Model quantization is performed on the first preset detection model, and the quantized first preset detection model is matched with the single-chip microcomputer.

[0050] The portable multi-channel energy spectrum acquisition instrument, the radioactive signal detection system, the detection method and the model training method provided by the embodiments of the present application can, when acquiring a nuclear pulse signal, first input the pulse signal through a pulse signal input end, then monitor a peak point of the pulse signal by a peak reaching timing unit, and generate a trigger signal when the pulse signal reaches the peak point, and send the trigger signal to a single-chip microcomputer. At this time, the single-chip microcomputer will control a sampling unit to collect the energy of the filtered pulse signal according to the trigger signal to obtain first energy information corresponding to the pulse signal. In the process of energy collection, through the cooperation of the peak reaching timing unit, the sampling unit and the single-chip microcomputer, not only the occupied area in circuit design can be reduced, but also the miniaturization of the portable multi-channel energy spectrum acquisition instrument can be realized, and the cost can be reduced. With the reduction of the cost of the portable multi-channel energy spectrum acquisition instrument, multiple portable multi-channel energy spectrum acquisition instruments can be arranged in a detection area, which provides a good application prospect for network detection, thereby improving the detection efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the embodiments of the present application will be briefly introduced. For those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0052] Figure 1 is a structural block diagram of a portable multi-channel energy spectrum acquisition instrument provided by an embodiment of the present application;

[0053] Figure 2 is a circuit connection diagram of a portable multi-channel energy spectrum acquisition instrument provided by another embodiment of the present application;

[0054] Figure 3 is a waveform diagram for displaying a trigger signal provided by another embodiment of the present application;

[0055] Figure 4 is a flow chart of a radioactive signal detection method provided in another embodiment of the present application;

[0056] Figure 5 is a flow chart of a radioactive signal detection method provided by a specific example of the present application;

[0057] Figure 6 is an example diagram of an energy spectrum diagram provided by another embodiment of the present application;

[0058] Figure 7 is a flowchart of a model training method provided by another embodiment of the present application;

[0059] Figure 8 This is a flowchart of a model training method provided in another embodiment of the present application.

[0060] Description of the drawings: 1. Peak timing unit; 11. Active differential module; 12. Comparator; 2. Filter unit; 3. Sampling unit; 4. Microcontroller; 5. Memory; 6. Communication module. DETAILED DESCRIPTION

[0061] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is merely to provide a better understanding of the present application by illustrating the examples of the present application.

[0062] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, the elements defined by the phrase "comprising..." do not exclude the presence of other identical elements in the process, method, article, or device comprising the elements.

[0063] Currently, when detecting the radioactivity level in the field, some sampling points are usually set, and then multi-channel energy spectrum acquisition instruments are laid at the positions of the sampling points to collect radioactivity signals. However, because the existing digital multi-channel energy spectrum acquisition instrument usually uses a preamplifier circuit to amplify the weak current pulse signal output by a detector (such as a NaI scintillation detector) and convert it into a voltage signal, and then uses a high-speed analog-to-digital converter to convert the nuclear pulse signal into a digital signal and send it to a field programmable gate array (FPGA) in real time to realize functions such as digital shaping of the nuclear pulse and multi-channel analysis.

[0064] In the related art, because the digital multi-channel energy spectrum acquisition instrument uses a high-speed analog-to-digital converter and an FPGA with high hardware cost, not only is the area occupied in the circuit design larger, which leads to difficulties in miniaturization of the digital multi-channel energy spectrum acquisition instrument, but also the overall hardware cost is higher, especially in application fields such as radioactivity level detection in the field, which is limited by cost, and the number of laid digital multi-channel energy spectrum acquisition instruments is smaller, leading to low detection efficiency.

[0065] To solve the problems in the prior art, the embodiments of the present application provide a portable multi-channel energy spectrum acquisition instrument, a radioactivity signal detection system, a detection method and a model training method. In the embodiments of the present application, when collecting a nuclear pulse signal, first, a pulse signal is connected through a pulse signal input end, then a peak point of the pulse signal is monitored by a peak reaching timing unit, and a trigger signal is generated when the pulse signal reaches the peak point, and the trigger signal is sent to a single-chip microcomputer. At this time, the single-chip microcomputer will control a sampling unit to collect the energy of the filtered pulse signal according to the trigger signal, and obtain first energy information corresponding to the pulse signal. In the energy collection process, through the cooperation of the peak reaching timing unit, the sampling unit and the single-chip microcomputer, not only can the area occupied in the circuit design be reduced, but also the miniaturization of the portable multi-channel energy spectrum acquisition instrument can be realized, the cost can be reduced, and with the reduction of the cost of the portable multi-channel energy spectrum acquisition instrument, multiple portable multi-channel energy spectrum acquisition instruments can be set in the detection area, which provides a good application prospect for networked detection, thereby improving the detection efficiency.

[0066] First, the portable multi-channel energy spectrum acquisition instrument provided by the embodiments of the present application will be introduced.

[0067] Figure 1 The structure block diagram of a portable multi-channel energy spectrum acquisition instrument is shown in FIG. 1. Figure 1As shown, the portable multi-channel energy spectrum acquisition instrument can include: a peak reaching timing unit 1, a filtering unit 2, a sampling unit 3 and a single-chip microcomputer 4, wherein the input end of the peak reaching timing unit 1 is connected with the pulse signal input end, the input end of the filtering unit 2 is connected with the pulse signal input end, the first input end of the sampling unit 3 is connected with the output end of the filtering unit 2, the first input end of the single-chip microcomputer 4 is connected with the output end of the peak reaching timing unit 1, the second input end of the single-chip microcomputer 4 is connected with the output end of the sampling unit 3, and the first output end of the single-chip microcomputer 4 is connected with the second input end of the sampling unit 3.

[0068] In the embodiment, the peak reaching timing unit 1 is used for detecting the peak point of the pulse signal and generating a trigger signal according to the peak point of the pulse signal; the filtering unit 2 is used for filtering the pulse signal; and the single-chip microcomputer 4 is used for controlling the sampling unit 3 to collect the energy of the filtered pulse signal according to the trigger signal to obtain first energy information, so as to draw an energy spectrum diagram of the pulse signal.

[0069] It is worth noting that the pulse signal input end is connected with the detector, that is, the detector can detect the radiation signal and then convert the radiation signal into the pulse signal and input the pulse signal into the portable multi-channel energy spectrum acquisition instrument.

[0070] In the embodiment, when the pulse signal is collected to obtain the energy information of the pulse signal, the pulse signal is first input through the pulse signal input end, then the peak reaching timing unit 1 monitors the peak point of the pulse signal and generates a trigger signal when the pulse signal reaches the peak point, and the trigger signal is sent to the single-chip microcomputer 4, at this time, the single-chip microcomputer 4 controls the sampling unit 3 to collect the energy of the filtered pulse signal according to the trigger signal to obtain the first energy information corresponding to the pulse signal. In the process of energy collection, through the cooperation of the peak reaching timing unit 1, the sampling unit 3 and the single-chip microcomputer 4, not only the occupied area in the circuit design can be reduced, but also the miniaturization of the portable multi-channel energy spectrum acquisition instrument can be realized, and the cost can be reduced. With the reduction of the cost of the portable multi-channel energy spectrum acquisition instrument, multiple portable multi-channel energy spectrum acquisition instruments can be arranged in the detection area, which provides a good application prospect for network detection, thereby improving the detection efficiency.

[0071] In some embodiments, referring to Figure 2 , the peak reaching timing unit 1 can include: an active differential module 11 and a comparator 12, wherein the input end of the active differential module 11 is connected with the pulse signal input end, the output end of the active differential module 11 is connected with the first input end of the comparator 12, the second input end of the comparator 12 is grounded, and the output end of the comparator 12 is connected with the first input end of the single-chip microcomputer 4.

[0072] In the embodiment, the active differentiation module 11 is configured to convert the received pulse signal into a bipolar pulse signal, and the comparator 12 is configured to perform zero-crossing comparison on the bipolar pulse signal, and output a trigger signal to the single-chip microcomputer 4 when the bipolar pulse signal undergoes a level inversion.

[0073] In some embodiments, when the peak point energy of the pulse signal is collected, a pulse pile-up event occurs, that is, the detected peak point is a superposition of multiple pulse signal peak points. In order to reduce the influence of the pulse pile-up event on the energy collection, the peak-reach timing circuit in the present application has a differential relationship between the output waveform and the input waveform, so that multiple different waveform inversions are present on the output waveform. These inversions successively generate multiple peak-reach timing levels through the comparator 12. The single-chip microcomputer 4 analyzes the pile-up pulses by the number of trigger levels in a period of time and the time difference of each trigger level, thereby improving the accuracy of energy analysis of the pile-up pulse signal and achieving higher signal pass rate and detection efficiency.

[0074] In some other embodiments, the active differentiation module 11 can include a first capacitor C1, a first resistor R1 and a first amplifier U1. The first end of the first capacitor C1 is connected to the pulse signal input end, the second end of the first capacitor is connected to the first end of the first resistor R1 and the input end of the first amplifier U1 respectively, the second input end of the first amplifier U1 is grounded, and the output end of the first amplifier U1 and the second end of the first resistor R1 are both connected to the input end of the comparator 12.

[0075] In the embodiment, the pulse signal enters the peak-reach timing circuit through the pulse signal input end, the voltage across the first capacitor C1 changes, and a current proportional to the rate of change of the pulse signal is generated. At this time, the current flows into the first amplifier U1 through the first resistor R1, is amplified by the first amplifier U1, and then outputs a changed waveform to form a bipolar pulse signal. Then, the zero-crossing detection of the comparator 12 is performed, and a trigger signal is output from the output end of the comparator 12, that is, when the waveform of the bipolar pulse signal passes through a zero point once, it indicates that the bipolar pulse signal undergoes a level inversion at this time, and the comparator 12 outputs a trigger signal to enable the single-chip microcomputer 4 to sample the pulse signal according to the trigger signal.

[0076] Referring to Figure 3 , Figure 3 The waveform diagram of the filter unit, the peak-reach timing unit and the trigger signal output by the comparator is shown. As can be seen from the waveform diagram, the trigger signal output by the comparator is a falling edge signal converted from a high level to a low level. Of course, this is only an example, and the trigger signal output by the comparator can also be a rising edge signal converted from a low level to a high level, which is not limited herein.

[0077] In some other embodiments, in order to more accurately collect the energy of the pulse signal, the filtering unit 2 can comprise an active low-pass module, an input end of the active low-pass module is connected with the pulse signal input end, and an output end of the active low-pass module is connected with the sampling unit 3.

[0078] In the embodiment, the active low-pass module can perform low-pass filtering on the pulse signal and output the denoised and smoothed pulse signal to be collected, that is, the filtered pulse signal is input into the single-chip microcomputer 4.

[0079] In a specific example, the active low-pass module can comprise a second resistor R2, a second capacitor C2 and a second amplifier U2, a first end of the second resistor R2 is connected with the pulse signal input end, a second end of the second resistor R2 is connected with a first end of the second capacitor C2 and a first input end of the second amplifier U2 respectively, a second end of the second capacitor C2 and a second input end of the second amplifier U2 are both grounded, and an output end of the second amplifier U2 is connected with an input end of the sampling unit 3 and the first input end of the second amplifier U2 respectively.

[0080] In the example, the active low-pass module can be a low-pass filtering circuit.

[0081] In some embodiments, the single-chip microcomputer 4 is provided with collection logic, that is, after the single-chip microcomputer 4 receives the trigger signal output by the peak timing unit 1, the single-point collection of the sampling unit 3 is triggered once, and through software delay and other operations, it is ensured that the single-point collection of the sampling unit 3 can accurately collect the peak point of the filtered pulse unit, and the peak point of the pulse signal can represent the energy information of the pulse signal.

[0082] As an example, the software delay is the hardware delay of the single-chip microcomputer 4, which can be preset to obtain the hardware delay time of the single-chip microcomputer 4, so as to ensure that the peak point of the filtered pulse unit can be accurately collected.

[0083] In the embodiment, in the process of measurement, the peak point of the pulse signal is continuously collected and accumulated in a spectrum with 4096 channels, and then the energy information is counted by counting the peak points.

[0084] In some other embodiments, the sampling unit 3 can be an analog-to-digital converter, which can be integrated in the single-chip microcomputer 4.

[0085] In some embodiments, the portable multi-channel energy spectrum collection instrument can further comprise a memory 5, the memory 5 is connected with the single-chip microcomputer 4, and the memory 5 is used to store the first energy information of the pulse signal output by the single-chip microcomputer 4.

[0086] In the embodiment, the memory 5 can be a direct memory 5, that is, in the case of power-off of the multi-channel energy spectrum acquisition instrument, the direct memory 5 can continue to store the first energy information.

[0087] In some embodiments, the single-chip microcomputer 4 can adopt a single-chip microcomputer of an STM32G431 model, the single-chip microcomputer 4 has a minimum of 32 pins, and an amplifier, a comparator and an analog-to-digital converter are integrated therein, compared with the minimum 144 pins of the FPGA in the prior art, the circuit design is smaller, the cost is lower, and the sampling unit 3 adopts a sampling frequency of 5M, compared with the high-speed analog-to-digital converter in the related art, the same area is smaller, and the power consumption is lower.

[0088] In some other embodiments, the portable multi-channel energy spectrum acquisition instrument further comprises a communication module 6, the communication module 6 is connected with the single-chip microcomputer 4, and is used for communication between the single-chip microcomputer 4 and the host computer.

[0089] In the embodiment, the communication module 6 can be a wired communication module, and can also be a wireless communication module, and correspondingly, in the case of the wired communication module, the communication module 6 can be connected by using a USB line, an RS485 or an RS232 communication line, in the case of the wireless communication module, the communication module 6 can be connected by using a Bluetooth, infrared or wireless network connection mode, which is not limited herein.

[0090] In the embodiment, in the case of connection between the single-chip microcomputer 4 and the host computer, the single-chip microcomputer 4 can directly call the first energy information stored in the direct memory 5, then send the first energy information to the host computer through the communication module 6, and the host computer directly displays the energy spectrum.

[0091] In some other embodiments, based on the cost reduction and the smaller circuit area of the above-mentioned portable multi-channel energy spectrum acquisition instrument, networking type detection application becomes possible, that is, for large-scale field radioactivity detection or environmental radioactivity monitoring, more portable multi-channel energy spectrum acquisition instruments can be laid in a to-be-detected area, and networking detection is formed based on the more portable multi-channel energy spectrum acquisition instruments, so that on-site measurement results of multiple measurement points can be quickly obtained, so that the overall measurement strategy can be more reasonably adjusted according to the measured part of the results, and because the cost is reduced, more intensive measurement points can be laid as much as possible, and the radioactivity level of the to-be-detected area can be restored more accurately.

[0092] Therefore, the embodiment of the application further provides a radioactivity signal detection system, and the detection system can comprise a plurality of detectors, a processor and a plurality of portable multi-channel energy spectrum acquisition instruments.

[0093] In the embodiment, multiple measuring points can be arranged in a to-be-detected area, at least one detector can be placed at each measuring point, each detector is connected to a portable multi-channel energy spectrum acquisition instrument in a one-to-one manner, and the processor is connected to the multiple portable multi-channel energy spectrum acquisition instruments in communication.

[0094] In the embodiment, when detecting the radioactivity level of the to-be-detected area, the detector converts a signal of the to-be-detected area into a pulse signal, converts the radioactivity signal into the pulse signal, and inputs the pulse signal into the portable multi-channel energy spectrum acquisition instrument. Then, the multi-channel energy spectrum acquisition instrument collects first energy information of the pulse signal to detect the radioactivity level of the to-be-detected area.

[0095] In some embodiments, the networking detection can be formed according to random sampling, stratified sampling, systematic sampling, cluster sampling, double sampling, adaptive sampling, and the like, to achieve better measurement effect.

[0096] Reference Figure 4 In some embodiments, based on the above radioactivity signal detection system, the embodiment of the application further provides a radioactivity signal detection method applied to a single-chip microcomputer of the portable multi-channel energy spectrum acquisition instrument. The method can include S401-S404:

[0097] S401, at least one first pulse signal in a to-be-detected area is acquired;

[0098] In some embodiments, in S401, the single-chip microcomputer can acquire at least one first pulse signal through the detector, that is, the single-chip microcomputer can acquire the first pulse signal of the measuring point through the detector placed at the measuring point in the to-be-detected area. That is, the single-chip microcomputer can be connected to the upper computer through the communication module, and can also be connected to the single-chip microcomputers of other portable multi-channel energy spectrum acquisition instruments, that is, each portable multi-channel energy spectrum acquisition instrument is connected to each other, so that each portable multi-channel energy spectrum acquisition instrument can acquire the first energy information of other measuring points in time, thereby realizing the networking detection application of the to-be-detected area.

[0099] S402, at least one first pulse signal is input into a first preset detection model, pulse anomaly detection is performed on the at least one first pulse signal by using the first preset detection model, and at least one second pulse signal is obtained;

[0100] In some embodiments, in S402, the second pulse signal can include a pulse signal in which an abnormal pulse in the first pulse signal is removed. Since the first pulse signal can be abnormally collected in the collection process, in order to better network detect the radioactivity level of the to-be-detected area, the first pulse signal can be input into the first preset detection model to screen abnormal signals and obtain the second pulse signal, and the radioactivity level is detected through the second pulse signal.

[0101] S403, count the number of peak points and the peak value of the second pulse signal to obtain first energy information of the second pulse signal;

[0102] In some embodiments, in S403, the single-chip microcomputer can perform peak collection and multi-channel analysis on the second pulse signal to obtain the first energy information of the second pulse signal.

[0103] S404, generate a radioactivity signal spectrum of the to-be-detected region based on the first energy information of the second pulse signal.

[0104] In the embodiments of the present application, when the radioactivity level of the to-be-detected region is detected in a networking manner, the single-chip microcomputer of each portable multi-channel energy spectrum acquisition instrument can obtain the energy information of the first pulse signal analyzed by other multi-channel energy spectrum acquisition instruments, and screen abnormal signals, so as to detect the radioactivity level according to the first energy information of the second pulse signal, and then generate a radioactivity signal spectrum of the entire to-be-detected region by comprehensively using the first energy information after multi-channel analysis by other single-chip microcomputers, so as to realize networking detection of the to-be-detected region.

[0105] In some embodiments, in order to more clearly illustrate the process of removing abnormal signals by the first preset detection model, a specific example is explained below.

[0106] Reference Figure 5 In a specific example, the first preset detection model can include an input layer, a convolution layer, a batch normalization layer, a full connection layer, a flexible maximization layer and a classification output layer connected in sequence; S402 can specifically include:

[0107] S4021, at least one first pulse signal is input from the input layer of the first preset detection model to the convolution layer;

[0108] S4022, the convolution layer of the first preset detection model performs convolution calculation on the at least one first pulse signal to obtain a plurality of time sequence features;

[0109] S4023, the normalization layer of the first preset detection model normalizes the plurality of time sequence features to obtain a plurality of normalized time sequence features;

[0110] S4024, the full connection layer of the first preset detection model integrates and extracts the plurality of normalized time sequence features to obtain a plurality of deep features;

[0111] S4025, the flexible maximization layer of the first preset detection model converts the plurality of deep features into activation probabilities of abnormal signals and normal signals;

[0112] S4026: The classification output layer of the first preset detection model classifies at least one first pulse signal according to the activation probability of a normal signal and the activation probability of an abnormal signal to obtain a second pulse signal.

[0113] In this example, in S4021, the at least one first pulse signal input may be a timestamp sequence of a differential circuit trigger signal within 10 us;

[0114] In S4022, the convolution layer can perform convolution calculation on the time interval and quantity of the above-mentioned timestamp continuous flow to obtain multiple time series features;

[0115] In S4023, by standardizing multiple time series features, the network convergence can be accelerated, so as to facilitate faster integration and extraction of deep features later.

[0116] In S4024, the fully connected layer can connect and convolve each time series feature to obtain multiple deep features.

[0117] In S4025, an activation function is set in the flexible maximum layer, and the activation probability can be calculated by using the activation function.

[0118] In S4026, classification may be performed using a classifier or other calculation methods.

[0119] In this example, abnormal signal detection is performed on at least one first pulse signal through each layer of the first preset detection model, and the abnormal signal is eliminated, thereby ensuring networked detection of the radioactivity level in the area to be detected.

[0120] In some other embodiments, when multiple portable multi-channel energy spectrum collectors and multiple detectors are provided in the area to be detected, and each portable multi-channel energy spectrum collector is communicatively connected to each other, S404 may include:

[0121] Obtain the spatial position information of each detector;

[0122] The second pulse signal and the spatial position information are input into a second preset detection model, and the second preset detection model obtains the radioactive signal energy spectrum of the area to be detected based on the second pulse signal and the spatial position information.

[0123] In the embodiment, in order to generate the radioactivity signal energy spectrum of the to-be-detected region more quickly, the spatial position information of each detector can be acquired, the spatial position information of each detector is pre-set and can be directly acquired, and then the second preset detection model is trained to obtain the radioactivity energy spectrum of the to-be-detected region according to the second pulse signal and the spatial position information, that is, the second preset detection model can determine the radioactivity level of the to-be-detected region according to the pulse signal converted by each detector and the corresponding position information, so as to more quickly and accurately realize the networking detection of the to-be-detected region, without the need for the upper computer to perform summary analysis on each portable multi-channel energy spectrum acquisition instrument to obtain the radioactivity level of the to-be-detected region.

[0124] Specifically, the second preset detection model can include a first convolutional neural network, a second graph neural network, and a third long short-term memory network, and the second preset detection model outputs the radioactivity level of the to-be-detected region according to the second pulse signal and the spatial position information, including:

[0125] The first convolutional neural network extracts time features of the first energy information to obtain a time feature vector of the first energy information.

[0126] The second graph neural network constructs a space-time graph of the second pulse signal by taking the spatial position information as node positions and taking the time feature vector as node attributes.

[0127] The third long short-term memory network predicts the energy information output by the other portable multi-channel energy spectrum acquisition instrument according to the space-time graph to obtain second energy information.

[0128] According to the second energy information and the first energy information, the radioactivity signal energy spectrum of the to-be-detected region is determined.

[0129] In some embodiments, the first convolutional neural network can extract time features of the first energy information to obtain an event feature vector of the first energy information, and the time feature vector represents time sequence characteristic data of the first energy information, that is, the first energy information is subjected to time sequence feature analysis to understand the time sequence characteristics of the second pulse signal.

[0130] In some other embodiments, a second graph neural network can be used to construct a space-time graph according to the spatial position information and the time sequence feature data of the detector, the space-time graph can convert the radioactivity level of the to-be-detected region into a time and space correlation graph that is convenient for analysis, and the radioactivity of each measurement point is associated, and then a third long short-term memory network is used to predict the radioactivity level change of the to-be-detected region in a period of time, so as to obtain the second energy information of each measurement point at each time. Finally, the radioactivity signal energy spectrum of the to-be-detected region can be output at the output layer of the second preset detection model. Through the second preset detection model, not only the radioactivity level of the to-be-detected region at the current time can be detected in a networked manner, but also the radioactivity level decay degree in a future period of time can be detected, and the radioactivity level can be more accurately determined according to the decay degree of the pulse signal.

[0131] It should be noted that the training method of the second preset detection model is the same as the training method in the related art, and will not be described in detail here.

[0132] Referring to Figure 6 , Figure 6 The energy spectrum diagram provided by the embodiments of the present application, that is, the second preset detection model can output an energy spectrum diagram as shown in Figure 6 , wherein, Figure 6 The abscissa in the energy spectrum diagram represents the channel number, and the ordinate represents the number of peak points.

[0133] Referring to Figure 7 , in some other embodiments, the embodiments of the present application also provide a model training method for training the first preset detection model, specifically, the model training method can include S701-S702:

[0134] S701, obtaining a training data set;

[0135] S702, taking the pulse signal as the input of the preset neural network model, taking the actual signal type as the output of the preset neural network model, training the preset neural network model to obtain the first preset detection model for performing pulse anomaly detection on at least one first pulse signal to obtain at least one second pulse signal as in the first aspect.

[0136] In the present embodiment, when training the first preset detection model, first, a training data set is obtained, wherein the training data set includes pulse signals in multiple environments and corresponding actual signal types, and the actual signal types include normal pulse signals and abnormal pulse signals. Then, the preset neural network model is trained based on the training data set to obtain the first preset detection model. The trained first preset detection model is used to screen abnormal signals to ensure the accuracy of networked detection.

[0137] In this embodiment, the preset neural network model can be a lightweight neural network model suitable for embedded deployment, that is, the preset neural network model can adopt a one-dimensional convolutional neural network to extract time domain features, reduce the size of the convolution kernel, and realize the lightweight of the preset neural network.

[0138] With reference to Figure 8 Specifically, S702 can include:

[0139] S7021, inputting the pulse signal into the preset neural network model to obtain a first predicted signal type;

[0140] S7022, determining a loss function of the preset neural network model according to the first predicted signal type and an actual signal type;

[0141] S7023, in the case where the loss function does not satisfy a preset convergence condition, adjusting model parameters of the preset neural network model, and returning to inputting the pulse signal into the preset neural network model to obtain the first predicted signal type, until the loss function satisfies the preset convergence condition, and obtaining a first preset detection model.

[0142] In this embodiment, the loss function of the preset neural network model can be determined in the above manner to complete the training of the preset neural network model.

[0143] It should be noted that the training manner of the loss function of the preset neural network model is consistent with the determination manner of the loss function in the related art, and will not be described in detail here.

[0144] In addition, after the training is completed, a test set can be obtained, and the trained first preset detection model is tested by using the test set. In the testing process, the model parameters can be adjusted to ensure the generalization ability of the model, and the first preset detection model after the testing can more accurately identify abnormal signals.

[0145] In some other embodiments, after S702, the method can further include:

[0146] quantizing the first preset detection model to match the single-chip microcomputer with the quantized first preset detection model.

[0147] In this embodiment, after the training of the first preset detection model is completed, it is also necessary to ensure that the first preset detection model can be matched to the single-chip microcomputer. Due to the memory limitation and storage format limitation of the single-chip microcomputer, it is necessary to quantize the storage format of the first preset detection model from a floating-point model to an integer model to reduce the calculation amount and memory occupation, and convert the model into a code that can be recognized by the single-chip microcomputer. At the same time, in order to ensure that the first preset detection model is suitable for the memory limitation of the single-chip microcomputer, the model weight can be stored in the direct storage, and the intermediate result can be stored in the RAM.

[0148] In the embodiment, the first preset detection model deployed in the single-chip microcomputer can realize real-time unified processing of abnormal pulse signals, prevent abnormal signals caused by environmental changes, greatly improve the stability of distributed energy spectrum measurement of the portable energy spectrum multichannel acquisition instrument, improve the energy spectrum resolution, reduce the occurrence of energy spectrum abnormal points, and thus more accurately realize networked detection of the radioactivity level of the region to be detected.

[0149] It should be noted that the present application is not limited to the specific configurations and processes described above and shown in the drawings. For the sake of brevity, detailed descriptions of known methods are omitted herein. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present application are not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between steps, after understanding the spirit of the present application.

[0150] The functional blocks shown in the above structural block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a functional card, etc. When implemented in software, the elements of the present application are program or code segments used to perform the required tasks. The program or code segments can be stored in a machine-readable medium or transmitted through a data signal carried in a carrier wave over a transmission medium or communication link. The "machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segments can be downloaded via a computer network such as the Internet, an intranet, etc.

[0151] It should also be noted that the exemplary embodiments mentioned in the present application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiments, or in an order different from the embodiments, or several steps can be performed simultaneously.

[0152] It should be noted that the present application is not limited to the specific configurations and processes described above and shown in the drawings. For the sake of brevity, detailed descriptions of known methods are omitted herein. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present application are not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between steps, after understanding the spirit of the present application.

[0153] The functional blocks shown in the structural block diagrams above can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, and the like. When implemented in software, the elements of the present application are program or code segments that are used to perform the required tasks. The program or code segments can be stored in a machine-readable medium, or transmitted through a data signal carried in a carrier wave over a transmission medium or communication link. A "machine-readable medium" includes any medium that can store or transport information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROMs, flash memory, erasable ROMs (EROMs), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, and the like. The code segments can be downloaded via computer networks such as the Internet, intranets, and the like.

[0154] It is also important to note that the examples mentioned in the present application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the examples, or in an order different from the examples, or several steps can be performed simultaneously.

[0155] The above describes the aspects of the present application with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present disclosure. It should be understood that each block in the flowcharts and / or block diagrams, and the combination of blocks in the flowcharts and / or block diagrams can 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, or other programmable data processing apparatus, to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus enable the implementation of the functions / actions specified in one or more blocks of the flowcharts and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field programmable logic circuit. It can also be understood that each block in the block diagrams and / or flowcharts, and the combination of blocks in the block diagrams and / or flowcharts, can also be implemented by special hardware that performs the specified functions or actions, or can be implemented by a combination of special hardware and computer instructions.

[0156] The above is only a specific embodiment of the present application, and those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, modules and units can refer to the corresponding processes in the foregoing method embodiments, which will not be described here. It should be understood that the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be covered within the protection scope of the present application.

Claims

1. A method for detecting radioactive signals, characterized in that: A single-chip microcomputer applied to a portable multi-channel energy spectrum acquisition instrument, the method comprising: Acquiring at least one first pulse signal within the area to be detected; Inputting at least one of the first pulse signals into a first preset detection model, performing pulse anomaly detection on the at least one first pulse signal using the first preset detection model to obtain at least one second pulse signal, where the second pulse signal includes a pulse signal obtained by removing abnormal pulses from the first pulse signal; Counting the number of peak points and the peak value of the second pulse signal to obtain first energy information of the second pulse signal; generating a radioactive signal energy spectrum of the area to be detected based on the first energy information of the second pulse signal; In the case where a plurality of portable multi-channel energy spectrum collectors and a plurality of detectors are provided in the area to be detected, and each of the portable multi-channel energy spectrum collectors is connected to communicate with each other, Generating the radioactive signal energy spectrum of the area to be detected based on the first energy information of the second pulse signal includes: Obtaining spatial position information of each of the detectors; inputting the second pulse signal and the spatial position information into a second preset detection model, wherein the second preset detection model outputs the radioactivity level of the area to be detected based on the second pulse signal and the spatial position information; The portable multi-channel energy spectrum collector comprises: A peak timing unit (1), the input end of the peak timing unit (1) is connected to the pulse signal input end, and is used to detect the peak point of the pulse signal and generate a trigger signal according to the peak point of the pulse signal; A filtering unit (2), the input end of the filtering unit (2) being connected to the pulse signal input end, and being used for filtering the pulse signal; A sampling unit (3), wherein a first input end of the sampling unit (3) is connected to an output end of the filtering unit (2); A single-chip microcomputer (4), wherein a first input end of the single-chip microcomputer (4) is connected to an output end of the peak timing unit (1), a second input end of the single-chip microcomputer (4) is connected to an output end of the sampling unit (3), and a first output end of the single-chip microcomputer (4) is connected to a second input end of the sampling unit (3); the single-chip microcomputer (4) is used to control the sampling unit (3) to perform energy acquisition on the filtered pulse signal according to the trigger signal to obtain first energy information.

2. The method according to claim 1, characterized in that The second preset detection model includes a first convolutional neural network, a second graph neural network and a third long short-term memory network; The second preset detection model outputs the radioactivity level of the area to be detected according to the second pulse signal and the spatial position information, including: The first convolutional neural network performs time feature extraction on the first energy information to obtain a time feature vector of the first energy information, where the time feature vector represents time series feature data of the first energy information; The second graph neural network uses the spatial position information as the node position and the time feature vector as the node attribute to construct a time-space graph of the second pulse signal; The third long short-term memory network predicts the energy information output by other portable multi-channel energy spectrum collectors based on the space-time graph to obtain second energy information; An energy spectrum of a radioactive signal of the area to be detected is determined according to the second energy information and the first energy information.

3. The method according to claim 1, characterized in that The peak timing unit (1) comprises: An active differential module (11), the input end of the active differential module (11) being connected to the pulse signal input end, and being used for converting the received pulse signal into a bipolar pulse signal; A comparator (12), wherein a first input terminal of the comparator (12) is connected to an output terminal of the active differential module (11), an output terminal of the comparator (12) is connected to a first input terminal of the single chip microcomputer (4), and a second input terminal of the comparator (12) is grounded; The comparator (12) is used to perform zero-crossing comparison on the bipolar pulse signal, and output a trigger signal to the single-chip microcomputer (4) when the bipolar pulse signal undergoes level reversal.

4. The method according to claim 3, characterized in that The active differential module (11) comprises: a first capacitor C1, a first resistor R1 and a first amplifier U1; The first end of the first capacitor C1 is connected to the pulse signal input end, the second end of the first capacitor C1 is connected to the first end of the first resistor R1 and the input end of the first amplifier U1 respectively, the second input end of the first amplifier U1 is grounded, and the output end of the first amplifier U1 and the second end of the first resistor R1 are both connected to the input end of the comparator (12).

5. The method according to any one of claims 1, 3 or 4, characterized in that The filtering unit (2) comprises: An active low-pass module, wherein the input end of the active low-pass module is connected to the pulse signal input end, and the output end of the active low-pass module is connected to the sampling unit (3).

6. The method according to claim 1, characterized in that The portable multi-channel energy spectrum collector also includes: A memory (5), the memory (5) being connected to the single chip microcomputer (4); The memory (5) is used to store first energy information of the pulse signal output by the single chip computer (4).

7. The method according to claim 1 or 6, characterized in that The portable multi-channel energy spectrum collector also includes: A communication module (6), the communication module (6) is connected to the single-chip microcomputer (4) and is used for the single-chip microcomputer (4) to communicate with a host computer.

8. The method according to claim 1, characterized in that The training process of the first preset detection model is: Acquire a training data set, wherein the training data set includes pulse signals under multiple environments and their corresponding actual signal types, wherein the actual signal types include normal pulse signals and abnormal pulse signals; The pulse signal is used as the input of a preset neural network model, the actual signal type is the output of the preset neural network model, and the preset neural network model is trained to obtain a first preset detection model for performing pulse anomaly detection on at least one first pulse signal to obtain at least one second pulse signal as claimed in claim 1.

9. The method according to claim 8, characterized in that The pulse signal is used as the input of the preset neural network model, the actual signal type is the output of the preset neural network model, and the preset neural network model is trained to obtain a first preset detection model, including: Inputting the pulse signal into a preset neural network model to obtain a first prediction signal type; Determining a loss function of the preset neural network model according to the first predicted signal type and the actual signal type; When the loss function does not meet the preset convergence conditions, the model parameters of the preset neural network model are adjusted, and the pulse signal is input into the preset neural network model again to obtain the first prediction signal type, until the loss function meets the preset convergence conditions to obtain the first preset detection model.

10. The method according to claim 8, characterized in that After using the pulse signal as the input of the preset neural network model and the actual signal type as the output of the preset neural network model, and training the preset neural network model to obtain a first preset detection model, the method further includes: The first preset detection model is quantized to match the quantized first preset detection model with the single chip microcomputer.

Images