CN-121723157-B - Real-time nuclear signal identification processing method and system

CN121723157BCN 121723157 BCN121723157 BCN 121723157BCN-121723157-B

Abstract

The invention discloses a real-time nuclear signal identification processing method and a real-time nuclear signal identification processing system, which belong to the technical field of signal processing, wherein continuous nuclear pulse analog signals conforming to double-exponential characteristics are converted into discrete nuclear pulse signals, corresponding attenuation parameters required by trapezoidal shaping are calculated in real time according to the discrete nuclear pulse signals, and the discrete nuclear pulse signals are converted into trapezoidal pulse signals according to the attenuation parameters calculated in real time. The invention can obviously improve the energy resolution and the counting accuracy of the nuclear detection system.

Inventors

ZHU DEJUN
HU QINGSHENG
LI LIANMING
WU XU

Assignees

东南大学

Dates

Publication Date: 20260512
Application Date: 20260226

Claims (8)

1. A real-time nuclear signal identification processing method is characterized by comprising the following steps: Converting a continuous nuclear pulse analog signal conforming to the double-exponential characteristic into a discrete nuclear pulse signal; calculating corresponding attenuation parameters required by trapezoidal shaping in real time according to the discrete nuclear pulse signals; According to the attenuation parameters calculated in real time, converting the discrete nuclear pulse signals into trapezoidal pulse signals; Order the Representing the sampling period, the discrete nuclear pulse signal is represented by the following formula : ; Wherein N represents an N-th sampling point, n=0, 1,..; And In order to be amplitude-value, And Respectively the attenuation constants of the signals, wherein N represents the total number of sampling points, and N sampling points form a sampling point set; Order the , ; Discrete nuclear pulse signal Expressed as: ; the second order linear recurrence relation satisfied by the discrete nuclear pulse signal is: ; Wherein, the And As the coefficient of the light-emitting diode, , ; Extracting k sampling points around the peak value of the nuclear pulse in sequence in the sampling point set, wherein the k sampling points span the peak value part of the nuclear pulse; Establishing a Hankel matrix equation, wherein an autoregressive model of the Hankel matrix equation is as follows: ; Wherein the method comprises the steps of Is the first of the set of sampling points Sampling points, and solving an autoregressive model by adopting a least square method so as to estimate And : , ; Wherein, the 、、、、、 Are all the parameters in the middle of the method, , , , , , , And (3) , , , And The following relationship is satisfied: 。
2. A real-time nuclear signal identification processing method according to claim 1, characterized in that the method also comprises the following steps And Calculating the amplitude of the nuclear pulse analog signal: Will be And Substituting the second order linear difference equation to obtain And Based on the value of (2) And Is to establish the following matrix: ; the matrix is subjected to least square solution, so that amplitude is obtained And : 。
3. The method for identifying and processing a real-time nuclear signal according to claim 1, wherein the discrete nuclear pulse signal is converted into a trapezoidal pulse signal, specifically: ; Wherein, the Is a trapezoidal signal at time t, wherein And Respectively a trapezoid rising edge, a trapezoid flat top, a trapezoid falling edge and a return-to-zero function, And The expressions of (2) are respectively as follows: ; Wherein, the 、 And The time parameters of the rising edge, the flat top and the falling edge of the trapezoid respectively, , ; H is the height of the flat top; , And Are all forming parameters, and ; Is a unit step function, will The corresponding transfer function formula is obtained by performing Z transformation after the comparison with the continuous nuclear pulse analog signal: ; Wherein, the As a complex variable of the Z-transform, And Is an attenuation parameter, and , ; Finally to Performing inverse Z transformation to obtain a time domain signal of double-index pulse, wherein the time domain signal is as follows: ; wherein Z -1 represents the inverse transform operation, Represented as a trapezoidal time domain signal corresponding to the nth sample, Representing the discrete nuclear pulse signal corresponding to the nth sample.
4. The method for identifying and processing the real-time nuclear signal according to claim 1, further comprising the steps of data storage of signals after trapezoidal shaping and parameter calibration, peak value information extraction, energy spectrum construction, and signal display and data analysis through an interactive interface.
5. The method of claim 4, further comprising sending the stored data to a host computer, and processing and displaying the data by the host computer.
6. A system for application to a method of real-time nuclear signal identification processing as claimed in claim 1, comprising: the discrete module is used for converting the continuous nuclear pulse analog signals into discrete nuclear pulse signals; the preprocessing module is used for denoising the original signal; The self-adaptive regression estimation module is used for carrying out self-adaptive regression calculation on the discrete nuclear pulse signals to obtain the amplitude values of the nuclear pulse signals and attenuation parameters required by trapezoid forming; the trapezoid forming module is used for converting the nuclear pulse signal with double-index characteristics into trapezoid output according to the preprocessed nuclear pulse signal and attenuation parameters required by trapezoid forming.
7. The system of claim 6, wherein the adaptive regression estimation module comprises first and second delay modules, first through eleventh multipliers, first through fifth accumulators, first through third adders, and first and second dividers; The input signal is input to a first delay module, a second delay module, a first multiplier and a fifth multiplier, the output of the first delay module is input to the first multiplier, the second multiplier and the third multiplier respectively, the output of the second delay module is input to the third multiplier, the fourth multiplier and the fifth multiplier respectively, the output of the first multiplier is input to the second input end of the first accumulator, the output of the second multiplier is input to the second input end of the second accumulator, the output of the third multiplier is input to the second input end of the third accumulator, the output of the fourth multiplier is input to the second input end of the fourth multiplier, the output of the fifth multiplier is input to the second input end of the fifth accumulator, the first input end of the first to fifth accumulator is connected with the output end, the output of the first accumulator is input to the sixth multiplier and the eighth multiplier, the output of the second accumulator is input to the ninth multiplier and the seventh multiplier, the output of the third multiplier is input to the sixth multiplier is input to the second input end of the eighth multiplier, the output of the eighth multiplier is connected to the eighth multiplier is input to the eighth multiplier is connected to the eighth multiplier, the output of the eighth multiplier is connected to the output end of the eighth multiplier, the output end of the third adder is connected with the second divider, the output of the first divider is used as the first output of the adaptive regression estimation module, and the output of the second divider is used as the second output of the adaptive regression estimation module.
8. The system of claim 7, wherein the trapezoidal shaping module comprises a third to sixth delay module, a fifth to eighth adder, a sixth to seventh accumulator, a twelfth to thirteenth multiplier, and a logarithmic amplifier, wherein the input signal is input to the non-inverting input terminal of the third delay module and the fifth adder, the output terminal of the third delay module is input to the inverting input terminal of the fifth adder, the output terminal of the fifth adder is connected to the fourth delay module, the non-inverting input terminal of the fifth delay module and the seventh adder, the output of the fourth delay module and the fifth delay module are respectively input to the input terminal of the twelfth multiplier and the input terminal of the thirteenth multiplier, the first output of the adaptive regression estimation module is input to the twelfth multiplier, the output of the twelfth multiplier is input to the non-inverting input terminal of the sixth adder, the output terminal of the thirteenth multiplier is input to the inverting input terminal of the sixth adder, the output terminal of the seventh adder is connected to the inverting input terminal of the seventh adder, the output terminal of the seventh adder is connected to the output terminal of the eighth adder, the output terminal of the eighth multiplier is connected to the inverting input terminal of the eighth multiplier is input to the eighth multiplier.

Description

Real-time nuclear signal identification processing method and system Technical Field The invention belongs to the technical field of signal processing, and particularly relates to a method and a system for identifying and processing a real-time nuclear signal. Background In the field of nuclear radiation detection, the nuclear pulse signal output by the detector typically has a negative exponential waveform characteristic of fast rising edges and slow decay. In order to achieve accurate measurement of radiation energy, improve system energy resolution, reduce count loss, and enhance anti-interference capability, an effective shaping process must be performed on the original nuclear pulse signal. Traditional analog forming techniques (such as CR-RC forming (Capacitor-Resistor Differentiator/Resistor-Capacitor Integrator Shaping), pseudo Gao Sicheng and the like) perform well in low count rate environments, but are difficult to adapt to complex multi-pulse stacking application scenes due to fixed forming parameters. For example, in the high-throughput nuclear event detection, the signal interval is shortened, pulse pile-up and even pile-up frequency discovery image occur, and the detection effect is poor and the detection performance is reduced due to the problems of energy measurement errors, pile-up signal omission or misjudgment and the like in the traditional fixed parameter forming method. In recent years, with the rapid development of digital signal processing technology, digital forming technology based on DSP and FPGA has gradually replaced traditional analog circuits, and has become the mainstream of research. At present, the digital trapezoidal shaping method mainly comprises two major types, namely a Z-domain transformation method and a functional convolution method, wherein the former is used for obtaining a time domain expression of a signal through Z-domain transformation, and the latter is used for realizing shaping by adopting convolution operation of an input signal and a trapezoidal function. Although the prior art has made many improvements to the trapezoidal shaping method, such as iterative optimization of the attenuation parameters by calculating the flat-top slope of the trapezoidal shaping, the method requires multiple resources and has a long shaping period, and real-time adjustment of the attenuation parameters cannot be realized when a plurality of nuclear signals with different parameters are input. In order to solve the problems, an attempt is made to introduce intelligent algorithms such as adaptive filtering and neural networks to process signals and obtain a certain effect. At present, the method also has the following problems that the forming parameter adjusting process depends on manual setting and modeling is complex, the algorithm is insufficient in instantaneity, the real-time data processing requirement is difficult to meet, more computing resources are occupied, and the realization of the method on an embedded platform with limited resources is limited. In summary, the existing iteration scheme and deep learning method cannot process the input nuclear signal in real time, and the method is complex and occupies more resources, so that forming cannot be completed under the condition of high counting rate. Therefore, a real-time shaping algorithm that is simple, has low resource occupancy, and can be implemented in an FPGA is needed. Disclosure of Invention The invention aims to solve the problems in the prior art, and provides a real-time nuclear signal identification processing method and a real-time nuclear signal identification processing system. The invention provides a real-time nuclear signal identification processing method, which specifically comprises the following steps: Converting a continuous nuclear pulse analog signal conforming to the double-exponential characteristic into a discrete nuclear pulse signal; calculating corresponding attenuation parameters required by trapezoidal shaping in real time according to the discrete nuclear pulse signals; and converting the discrete nuclear pulse signals into trapezoidal pulse signals according to attenuation parameters calculated in real time. Further, let theRepresenting the sampling period, the discrete nuclear pulse signal is represented by the following formula: ; Wherein N represents an N-th sampling point, n=0, 1,..; And In order to be amplitude-value,AndThe attenuation constant of the signal is distinguished, N represents the total number of sampling points, and N sampling points form a sampling point set. Further, the method comprises the steps of, Order the,; Discrete nuclear pulse signalExpressed as: ; the second order linear recurrence relation satisfied by the discrete nuclear pulse signal is: ; Wherein, the AndAs the coefficient of the light-emitting diode,,; Extracting k sampling points around the peak value of the nuclear pulse in sequence in the sampling point set, wherein the k sampling points span