Gamma-ray energy spectrum generation method based on digital waveform analysis

Abstract

The invention discloses a gamma counter energy spectrum generation method based on digital waveform analysis. The method comprises the following steps: 1) screening out a plurality of candidate peaks from analog pulse signals output by a gamma detector for detecting a target gamma source; 2) for each candidate peak, calculating an average change slope Kavg and an amplitude variation delta V of the candidate peak; 3) respectively comparing the average change slope Kavg and the amplitude variation delta V of the candidate peak with a slope threshold value Kth and an amplitude difference threshold value delta Vth, if Kavg is greater than or equal to Kth and delta V is greater than or equal to delta Vth, judging that the candidate peak is an effective gamma signal peak and transmitting the effective gamma signal peak to a subsequent data processing module, otherwise, judging that the candidate peak is a noise false peak; and 4) the data processing module generates the gamma-ray energy spectrum of the target gamma-ray source according to the received gamma-ray signal peak. According to the invention, the synchronous optimization of the energy spectrum resolution and the measurement sensitivity can be realized, the anti-interference capability is strong, and the robustness is high.

Description

Technical Field

[0001] This invention belongs to the field of nuclear radiation detection and signal processing technology, and relates to a gamma energy spectrum generation method based on digital waveform analysis, which is applied to the low-energy noise false peak filtering scenario of gamma counters. Existing technologies, when using traditional peak finding algorithms, may mistakenly identify low-energy noise false peaks generated by baseline recovery as valid signals, resulting in a decrease in energy spectrum resolution and an increase in background count. Background Technology

[0002] Gamma-ray spectroscopy measurement is a fundamental method in nuclear technology applications, and the conventional signal processing flow of its core equipment, the gamma-ray spectrometer, has become a well-known technology in the field. Specifically, the process involves: the gamma detector outputting an analog pulse signal, which is amplified by a preamplifier, further shaped by a main amplifier, and then sampled and quantized by an analog-to-digital converter (ADC) to convert it into a digital signal stream. This digital signal stream is then fed into digital processing devices such as field-programmable gate arrays (FPGAs) or digital signal processors (DSPs) to sequentially complete subsequent processing steps such as baseline recovery, accumulation rejection, and peak extraction. Finally, a gamma-ray spectrum is constructed based on the effective peak information.

[0003] The core function of baseline restoration is to reduce the impact of baseline drift on signal peak values ​​and ensure the accuracy of subsequent peak extraction. Commonly used baseline restoration algorithms include moving average and polynomial fitting algorithms. In the peak extraction stage, traditional peak finding algorithms are widely used and well-known techniques in this field. Their core logic is to identify signal segments with multiple consecutive sampling points exceeding the threshold and exhibiting a "rising then falling" peak shape by setting a trigger threshold, thus determining them as valid peaks and extracting peak information.

[0004] The closest existing technical solution to this proposal is a gamma counter peak finding and noise suppression method based on a traditional single criterion, and its specific implementation process is as follows: 1. Signal preprocessing: The analog pulse signal output by the detector is quantized into a digital signal by the ADC and then sent to the digital processing device. The baseline recovery algorithm (such as the moving average method) is used to stabilize the signal baseline and eliminate the influence of baseline drift.

[0005] 2. Candidate Peak Screening: The traditional peak finding algorithm is applied to process the signal after the baseline is stabilized. A fixed trigger threshold is set. When multiple consecutive sampling points of the signal exceed the threshold and the waveform meets the peak shape characteristics of "rising first and then falling", the signal segment is identified as a candidate peak, and the starting point and ending point of the candidate peak (i.e. the boundary of the decision interval) are determined.

[0006] 3. Effective peak determination and energy spectrum construction: The selected candidate peaks are directly determined as effective gamma signal peaks, and their peak information is extracted for energy spectrum construction. No additional noise screening step is set, and only some low-amplitude noise is initially suppressed by triggering threshold.

[0007] The disadvantages of existing technologies are as follows: 1. Incomplete noise discrimination logic, leading to a high risk of misjudgment: Existing technologies rely solely on a single criterion of "threshold triggering + peak shape characteristics" for effective peak identification, failing to consider the fundamental differences between noise spurious peaks and real signal peaks in terms of time-domain waveform change rate and overall amplitude change. When processing dense signals or electronic noise, baseline recovery algorithms introduce low-frequency disturbances or peak-like distortions. After subsequent digital shaping, these distortions form a series of spurious signals with peak shape characteristics in the low-energy region (usually corresponding to the region with lower signal amplitude). These noise spurious peaks also satisfy the criterion of "continuously exceeding the threshold + peak shape that rises first and then falls" in the time domain, making it impossible for existing technologies to effectively distinguish them.

[0008] 2. Decreased energy spectrum resolution: When noise false peaks are misjudged as valid signals, they will raise the background level in the low energy region, causing the real small-amplitude signal peaks near the low energy region to be submerged and difficult to be accurately identified and analyzed, directly resulting in a decrease in energy spectrum resolution.

[0009] 3. Increased background count: The false count of noise peaks increases the background count of the entire energy spectrum (especially in the low energy region), reduces the ratio of effective signal to background, and decreases the detection sensitivity of the gamma counter to weak gamma signals. Summary of the Invention

[0010] In view of the problems existing in the prior art, the purpose of this invention is to provide a method for generating gamma energy spectrum based on digital waveform analysis.

[0011] The technical solution of this invention is as follows: A method for generating the energy spectrum of a gamma counter based on digital waveform analysis, comprising the following steps: 1) Select several candidate peaks from the simulated pulse signal output by the gamma detector to the target gamma source; 2) For each candidate peak, calculate the average slope K_avg and amplitude change ΔV of the candidate peak; 3) Compare the average slope K_avg and amplitude change ΔV of the candidate peak with the slope threshold K_th and amplitude difference threshold ΔV_th, respectively. If K_avg ≥ K_th and ΔV ≥ ΔV_th, the candidate peak is determined to be a valid gamma signal peak and transmitted to the subsequent data processing module; otherwise, the candidate peak is determined to be a noise false peak. 4) The data processing module generates the gamma energy spectrum of the target gamma source based on the received gamma signal peak.

[0012] Preferably, the method for calculating the average slope K_avg is as follows: 21) Sample the leading edge of the candidate peak to obtain N consecutive sampling points, calculate the amplitude difference ΔV_i = V (i+1) - V (i) between adjacent sampling points, where V (i) is the amplitude of the i-th sampling point, divide by the sampling interval T to obtain N-1 instantaneous rising slopes and take the average of their absolute values ​​as the rising slope of the leading edge K_rise = (Σ|ΔV_i / T|) / (N-1); 22) Sample the trailing edge of the candidate peak to obtain N consecutive sampling points. Calculate the amplitude difference between adjacent sampling points and divide it by the sampling interval T to obtain N-1 instantaneous falling slopes. Take the average of their absolute values ​​to obtain the trailing edge falling slope K_fall. 23) Calculate the average slope of the candidate peak K_avg = (|K_rise| + |K_fall|) / 2.

[0013] Preferably, the method for calculating the amplitude change ΔV is as follows: traverse all sampling points within the decision interval of the candidate peak to obtain the maximum amplitude V_max and the minimum amplitude V_min within the decision interval; calculate the amplitude change ΔV = V_max - V_min.

[0014] Preferably, the average slope K_avg of the candidate peak and the amplitude change ΔV are synchronously calculated within the FPGA through parallel computing logic.

[0015] Preferably, the method for obtaining the slope threshold K_th and the amplitude difference threshold ΔV_th is as follows: a gamma detector is used to detect a standard gamma source signal with known energy, and then multiple sets of real signal peaks are selected from the simulated pulse signal output by the gamma detector. The average slope K_avg and amplitude change ΔV of these multiple sets of real signal peaks are collected, and 80%-90% of their minimum value is taken as the corresponding slope threshold K_th and amplitude difference threshold ΔV_th.

[0016] Preferably, the method for selecting several candidate peaks is as follows: First, the analog pulse signal is converted into a digital signal stream and sent to the FPGA; the FPGA uses a digital baseline recovery algorithm to process the digital signal stream, tracks the signal baseline in real time and corrects it to obtain a digital signal with a stable baseline; when the digital signal has M consecutive sampling points exceeding a set fixed trigger threshold V_trigger, and subsequent sampling points decrease and fall below the fixed trigger threshold V_trigger within a certain period of time, the corresponding signal segment is determined to be a candidate peak; the sampling point in the signal segment that first exceeds the fixed trigger threshold V_trigger is taken as the starting point of the candidate peak, and the sampling point that falls below the fixed trigger threshold V_trigger again is taken as the ending point of the candidate peak, and the time period formed by all sampling points from the starting point to the ending point of the candidate peak is taken as the decision interval of the candidate peak.

[0017] Preferably, M is greater than or equal to 3.

[0018] Key aspects of this technology include: 1. The study discovered a dual essential difference between noise false peaks and real gamma signal peaks in terms of the average slope of the time-domain waveform change and the magnitude of the change in the decision interval, providing a core basis for dual-criteria identification.

[0019] 2. A precise extraction method for dual feature parameters, including a weighted integration calculation method for the average slope of change and a full-range statistical method for the magnitude of change, to ensure the robustness and discriminative power of the parameters.

[0020] 3. The joint discrimination logic of "slope criterion + amplitude difference criterion" achieves accurate filtering of noise false peaks through "AND" logic, while ensuring that the real signal is not lost.

[0021] The gamma spectrum generation method based on digital waveform analysis of this invention includes a four-step process: signal preprocessing and candidate peak coarse screening, precise extraction of dual feature parameters, joint intelligent identification based on dual criteria, result output and spectrum construction.

[0022] The two characteristic parameters are the average slope of change K_avg and the magnitude change ΔV in the decision interval. K_avg is calculated by averaging the absolute values ​​of the rising slope of the leading edge and the falling slope of the trailing edge, and ΔV is the difference between the maximum and minimum magnitudes in the decision interval.

[0023] The dual-criteria joint discrimination logic is as follows: a candidate peak is determined to be a valid gamma signal peak only when K_avg ≥ the preset slope threshold K_th and ΔV ≥ the preset amplitude difference threshold ΔV_th; otherwise, it is a noise false peak.

[0024] In signal preprocessing, a digital baseline recovery algorithm is used to stabilize the signal baseline, and a traditional peak-finding algorithm is used to identify signal segments that meet the trigger threshold and peak shape characteristics for candidate peak coarse screening.

[0025] The advantages of this invention are as follows: 1. Significantly Improved Noise Identification Accuracy: This proposal employs a dual-feature parameter criterion of "average slope + amplitude change," accurately capturing the essential difference between noise false peaks (small slope, small amplitude difference) and true signal peaks (large slope, large amplitude difference). This solves the problem that existing technologies cannot distinguish between them using a single criterion, significantly reducing the misjudgment rate of low-energy noise false peaks. This advantage stems from the complementary design of the dual feature parameters: the slope parameter reflects the waveform change rate, and the amplitude difference parameter reflects the overall amplitude fluctuation; this dual dimension ensures reliable identification.

[0026] 2. Simultaneous optimization of energy spectrum resolution and measurement sensitivity: While effectively filtering out noise spurious peaks, the threshold setting mechanism in this proposal can accurately match the parameter range of the real signal peaks without causing loss to the real signal. This avoids raising the background in the low-energy region and ensures clear resolution of real small signal peaks in the low-energy region, thus improving both energy spectrum resolution and detector sensitivity to weak signals. This advantage is due to the "AND" logic design of the dual criteria, which retains only signals that simultaneously satisfy two core characteristics.

[0027] 3. Strong anti-interference capability and high robustness: The calculation of the average slope adopts a multi-sampling point mean and weighted integration method, reducing the impact of local noise on parameter extraction; the amplitude change is calculated based on the entire decision interval, unaffected by local waveform distortion. Simultaneously, the threshold can be flexibly set through calibration experiments to adapt to different detector types and application scenarios. Compared with the fixed threshold method in existing technologies, it has stronger anti-interference capability and environmental adaptability. This advantage stems from the robust design of parameter extraction and the flexible calibration mechanism of the threshold. Attached Figure Description

[0028] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0029] The present invention will now be described in further detail with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0030] This invention provides a method and system for filtering low-energy noise spurious peaks with greater accuracy and stronger anti-interference capability, overcoming the limitations of existing technologies and single-criteria discrimination methods. It utilizes the dual differences between noise spurious peaks and real signals in both the rate of change of the time-domain waveform and the overall amplitude change for discrimination. Only when a candidate peak simultaneously meets the conditions of a sufficiently large slope and a sufficiently large amplitude difference is it identified as a valid peak; otherwise, it is considered noise and filtered out. The method flow of this invention is as follows: Figure 1 As shown, the specific steps are as follows.

[0031] (a) Signal preprocessing and coarse screening of candidate peaks 1. Signal Quantization and Baseline Restoration: The analog pulse signal output by the gamma detector to the target gamma source is sampled and quantized by the ADC at a preset sampling frequency (e.g., 100MHz), converted into a digital signal stream, and then sent to the FPGA. Inside the FPGA, a digital baseline restoration algorithm (such as an improved moving average algorithm, with a window length of 50-100 sampling points) is used to process the digital signal stream, track the signal baseline in real time, and perform corrections to ensure signal baseline stability and eliminate the effects of baseline drift.

[0032] 2. Initial Screening of Candidate Peaks: For digital signals after baseline stabilization, a traditional peak-finding algorithm is applied for initial screening. A fixed trigger threshold V_trigger is set (which can be calibrated according to the actual application scenario, such as 10%-20% of the minimum effective signal peak value). When three or more consecutive sampling points of the signal exceed V_trigger, and within a certain time (the width of the effective signal is fixed, i.e., the number of effective signal sampling values ​​greater than V_trigger is fixed), three or more subsequent consecutive sampling points decrease and fall below V_trigger, the signal segment is determined as a candidate peak. At the same time, the starting point (the sampling point that first exceeds V_trigger) and the ending point (the sampling point that falls below V_trigger again) of the candidate peak are recorded, and the decision interval for each candidate peak (i.e., the time period consisting of all sampling points from the starting point to the ending point) is determined.

[0033] (II) Precise Extraction of Dual Feature Parameters For each candidate peak obtained from the initial screening, two key feature parameters are extracted synchronously within the FPGA using parallel computing logic to ensure processing efficiency.

[0034] Parameter 1: Average slope of change K_avg 1. Calculation of the rising slope K_rise: Select N consecutive sampling points (N can be set to 10-20, adjusted according to the sampling frequency) of the candidate peak front (the sampling segment from the starting point to the peak point), calculate the amplitude difference ΔV_i = V (i+1) - V (i) of adjacent sampling points (i is the sampling point number), and then divide by the sampling interval T (determined by the sampling frequency, such as T=10ns for a sampling frequency of 100MHz) to obtain N-1 instantaneous rising slopes; take the average of the absolute values ​​as the rising slope K_rise, that is, K_rise = (Σ|ΔV_i / T|) / (N-1).

[0035] 2. Calculation of the falling slope K_fall: Select N consecutive sampling points at the falling edge of the candidate peak (the sampling segment from the peak point to the end point), and use the same calculation method as the leading edge to obtain the falling slope K_fall, that is, K_fall = (Σ|V (i) - V (i+1)| / T) / (N-1).

[0036] 3. Average slope integration: To improve parameter robustness and avoid the influence of local noise on a single slope, K_avg = (|K_rise| + |K_fall|) / 2 is taken as the final measure of the average slope of the candidate peak. This parameter reflects the speed of change of the candidate peak waveform.

[0037] Parameter 2: Change in amplitude ΔV within the decision interval 1. Traverse all sampling points within the candidate peak decision interval, and record the maximum amplitude V_max and minimum amplitude V_min within the interval in real time through the peak detection logic inside the FPGA.

[0038] 2. Calculate the amplitude change ΔV = V_max - V_min. This parameter directly reflects the overall amplitude fluctuation of the candidate peak throughout the entire duration and is not affected by local fluctuations.

[0039] (III) Joint Intelligent Identification Based on Dual Criteria 1. Threshold Setting: Two preset thresholds are used, where the slope threshold K_th and amplitude threshold ΔV_th need to be determined through calibration experiments. The calibration process is as follows: Input a standard gamma source signal with known energy, collect multiple sets of K_avg and ΔV values ​​of real signal peaks, and take 80%-90% of the minimum value as the corresponding threshold (e.g., if the minimum value of K_avg of the real signal peak is 5V / μs, then K_th is set to 4V / μs; if the minimum value of ΔV is 200mV, then ΔV_th is set to 160mV), to ensure that the real signal peak energy stably meets the criteria.

[0040] 2. Criterion comparison and logical judgment: (1) Slope criterion verification: Compare the extracted K_avg with K_th. If K_avg ≥ K_th, the slope criterion is satisfied; otherwise, it is not satisfied.

[0041] (2) Verification of amplitude difference criterion: Compare the extracted ΔV with ΔV_th. If ΔV ≥ ΔV_th, the amplitude difference criterion is satisfied; otherwise, it is not satisfied.

[0042] (3) Final identification logic: The "AND" logic is used to determine the valid peak. Only when the candidate peak meets both the slope criterion and the amplitude difference criterion is it determined to be a valid gamma signal peak; if either criterion is not met, it is determined to be a noise false peak.

[0043] (iv) Results output and energy spectrum construction The FPGA only outputs the amplitude information of the peaks that are determined to be valid, and transmits it to the subsequent data processing module. The module counts the different amplitude ranges according to conventional energy spectrum construction methods (such as pulse amplitude analyzers) to generate the gamma energy spectrum of the target gamma source. All signal segments that are determined to be noise false peaks are directly filtered out and do not participate in the energy spectrum construction.

[0044] Although specific embodiments of the invention have been disclosed for illustrative purposes to aid in understanding and implementing the invention, those skilled in the art will understand that various substitutions, variations, and modifications are possible without departing from the spirit and scope of the invention and the appended claims. Therefore, the invention should not be limited to the content disclosed in the preferred embodiments, and the scope of protection claimed by the invention is defined by the claims.

Claims

1. A method for generating gamma-ray spectrum based on digital waveform analysis, comprising the steps of: 1) selecting a plurality of candidate peaks from analog pulse signals outputted by a gamma-ray detector in response to a target gamma-ray source; 2) calculating, for each of the candidate peaks, an average variation slope K_avg and a variation amount of amplitude ΔV; 3) comparing the average variation slope K_avg and the variation amount of amplitude ΔV of each of the candidate peaks with a slope threshold K_th and an amplitude difference threshold ΔV_th, respectively, and determining that the candidate peak is a valid gamma-ray signal peak if K_avg ≥ K_th and ΔV ≥ ΔV_th, and transmitting the candidate peak to a subsequent data processing module, otherwise, determining that the candidate peak is a noise false peak; 4) generating a gamma-ray spectrum of the target gamma-ray source by the data processing module according to the received gamma-ray signal peaks.

2. The method of claim 1, wherein, The method for calculating the average variation slope K_avg is: 21) sampling a front edge of the candidate peak to obtain a plurality of continuous sampling points, calculating a difference value ΔV_i = V(i+1) - V(i) of amplitudes of adjacent sampling points, dividing the difference value ΔV_i by a sampling interval T, obtaining a plurality of instantaneous rising slopes, taking an average value of absolute values of the instantaneous rising slopes, and taking the average value as a front edge rising slope K_rise = (Σ|ΔV_i / T|) / (N-1); 22) sampling a trailing edge of the candidate peak to obtain a plurality of continuous sampling points, calculating a difference value of amplitudes of adjacent sampling points, dividing the difference value by the sampling interval T, obtaining a plurality of instantaneous falling slopes, taking an average value of absolute values of the instantaneous falling slopes, and taking the average value as a trailing edge falling slope K_fall; 23) calculating the average variation slope K_avg = (|K_rise| + |K_fall|) / 2 of the candidate peak.

3. The method of claim 1, wherein, The method for calculating the variation amount of amplitude ΔV is: traversing all sampling points in a decision interval of the candidate peak to obtain a maximum amplitude V_max and a minimum amplitude V_min in the decision interval; and calculating the variation amount of amplitude ΔV = V_max - V_min.

4. The method according to claim 1 or 2 or 3, characterized in that, The average variation slope K_avg and the variation amount of amplitude ΔV of the candidate peak are calculated synchronously by parallel computing logic in an FPGA.

5. The method according to claim 1 or 2 or 3, characterized in that, The method for obtaining the slope threshold K_th and the amplitude difference threshold ΔV_th is: detecting a standard gamma-ray source signal of a known energy by a gamma-ray detector, selecting a plurality of real signal peaks from analog pulse signals outputted by the gamma-ray detector, collecting average variation slopes K_avg and variation amounts of amplitudes ΔV of the plurality of real signal peaks, and taking 80%-90% of a minimum value of the average variation slopes K_avg and the variation amounts of amplitudes ΔV as the corresponding slope threshold K_th and amplitude difference threshold ΔV_th.

6. The method according to claim 1 or 2 or 3, characterized in that, The method for screening a plurality of candidate peaks is: firstly, the analog pulse signal is converted into a digital signal stream and sent into an FPGA; the FPGA uses a digital baseline recovery algorithm to process the digital signal stream, tracks and corrects the signal baseline in real time, and obtains a digital signal with a stable baseline; when the digital signal has continuously M sampling points exceeding a set fixed trigger threshold V_trigger, and the subsequent sampling points fall below the fixed trigger threshold V_trigger within a certain time, it is determined that the corresponding signal segment is a candidate peak; the sampling point first exceeding the fixed trigger threshold V_trigger in the signal segment is taken as the starting point of the candidate peak, the sampling point again falling below the fixed trigger threshold V_trigger is taken as the end point of the candidate peak, and the time interval formed by all the sampling points from the starting point to the end point of the candidate peak is taken as the decision interval of the candidate peak.

7. The method of claim 6, wherein, M is greater than or equal to 3.

Images