CN115856987B - Nuclear pulse signal and noise signal discrimination method in complex environment - Google Patents

Nuclear pulse signal and noise signal discrimination method in complex environment Download PDF

Info

Publication number
CN115856987B
CN115856987B CN202310173699.0A CN202310173699A CN115856987B CN 115856987 B CN115856987 B CN 115856987B CN 202310173699 A CN202310173699 A CN 202310173699A CN 115856987 B CN115856987 B CN 115856987B
Authority
CN
China
Prior art keywords
signal
noise
signals
pulse signal
nuclear pulse
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202310173699.0A
Other languages
Chinese (zh)
Other versions
CN115856987A (en
Inventor
张江梅
张草林
赵志豪
刘灏霖
熊芸峰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southwest University of Science and Technology
Original Assignee
Southwest University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Southwest University of Science and Technology filed Critical Southwest University of Science and Technology
Priority to CN202310173699.0A priority Critical patent/CN115856987B/en
Publication of CN115856987A publication Critical patent/CN115856987A/en
Application granted granted Critical
Publication of CN115856987B publication Critical patent/CN115856987B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E30/00—Energy generation of nuclear origin
    • Y02E30/30—Nuclear fission reactors

Abstract

The invention discloses a nuclear pulse signal and noise signal discrimination method in a complex environment, which relates to the field of nuclear radiation detection and comprises the following steps: firstly, performing FIR low-pass filtering on an acquired original signal; then, carrying out time-frequency decomposition on the filtered signals, and mapping the signals from a one-dimensional time sequence to a two-dimensional frequency domain space; extracting signal characteristics in a two-dimensional frequency domain space; inputting the signal characteristics into a trained nuclear pulse signal and noise separation model, outputting, and obtaining a nuclear pulse signal and a noise signal by inverse transformation and combining phases of signals corresponding to the signal characteristics; further preprocessing the obtained nuclear pulse signal; finally, extracting the amplitude of the preprocessed nuclear pulse signal, and generating an energy spectrum according to the amplitude; according to the invention, the kernel pulse and the noise signals can be effectively distinguished by adopting a machine learning-based method, so that the acquisition of the noise signals is reduced.

Description

Nuclear pulse signal and noise signal discrimination method in complex environment
Technical Field
The invention relates to the field of nuclear radiation detection, in particular to a nuclear pulse signal and noise signal discrimination method in a complex environment.
Background
The statements in this section merely provide background information related to the present disclosure and may not constitute prior art.
In nuclear radiation pulse measurement techniques, the useful pulse signal is always superimposed on an unstable baseline voltage, resulting in a pulse amplitude that deviates from the actual value, affecting the energy spectrum measurement; such voltages are called the baseline of the pulse signal, and the phenomenon that the pulse amplitude deviates from the actual value is called the baseline drift of the pulse signal; in the measurement process, environmental noise, leakage current of the detector, temperature drift of electronic components, ripple wave of a power supply and the like can cause baseline drift of the nuclear radiation detector.
Baseline restorer is commonly used in existing analog systems to estimate the baseline size and subtract it, which increases the complexity of the circuit; and new energy resolution loss is introduced due to the influence of uncertain factors; the baseline estimation method in the digital multichannel system can be optimized according to the characteristics of the input signals, but still adopts a fixed threshold comparison mode to distinguish whether the signals are effective pulses or not; when the signal amplitude exceeds a certain fixed threshold value, the signal amplitude is judged to be a valid nuclear pulse signal, and otherwise, the signal amplitude is judged to be a noise signal; in this case, the increase in acquired noise may result in a deterioration of the energy resolution of the detector due to drift of the baseline.
Disclosure of Invention
The invention aims at: aiming at the problems that the acquired noise is increased to cause the energy resolution of the detector to be poor due to the drift of a baseline in the current nuclear radiation pulse measurement, the nuclear pulse signal and noise signal discrimination method under a complex environment is provided, the problems of noise and nuclear pulse signal discrimination, the problems of the acquired noise increase and the energy spectrum resolution to be poor under the condition that the noise baseline drift causes a fixed threshold in the long-time measurement process of the nuclear radiation detector and the like are solved, and the online measurement of gamma energy spectrum is realized.
The technical scheme of the invention is as follows:
a nuclear pulse signal and noise signal discrimination method under a complex environment comprises the following steps:
step S1: performing FIR low-pass filtering on the collected original signals;
step S2: performing time-frequency decomposition on the filtered signals, and mapping the signals from a one-dimensional time sequence to a two-dimensional frequency domain space;
step S3: extracting signal characteristics in a two-dimensional frequency domain space;
step S4: inputting the signal characteristics into a trained nuclear pulse signal and noise separation model, outputting the nuclear pulse signal and noise separation model, and obtaining the nuclear pulse signal and the noise signal by inverse transformation and combining the phases of signals corresponding to the signal characteristics;
step S5: preprocessing the obtained nuclear pulse signal;
step S6: and extracting the amplitude of the preprocessed nuclear pulse signal, and generating an energy spectrum according to the amplitude.
Further, the step S1 includes:
acquiring an original signal by using an ADC, wherein the original signal comprises: a nuclear pulse signal, a noise signal, and a mixed signal of the two;
the original signal is subjected to FIR low-pass filtering, and the filtering formula is as follows:
Figure SMS_1
wherein ：
Figure SMS_2
for the input signal at time n +.>
Figure SMS_3
For the input signal at time n-k, < >>
Figure SMS_4
For FIR filter coefficients>
Figure SMS_5
For the filtered signal, N represents the number of taps of the FIR filter, < >>
Figure SMS_6
For the filter order, k is a positive integer and k +.>
Figure SMS_7
N is a positive integer and is expressed as a certain time.
Further, the step S2 includes:
is provided with
Figure SMS_8
Is a real symmetric window function, +.>
Figure SMS_9
Is a one-dimensional time domain signal->
Figure SMS_10
In->
Figure SMS_11
Time frame, th->
Figure SMS_12
Short-time fourier transform coefficients for the individual frequency bands:
Figure SMS_13
wherein: k is a positive integer, which represents a certain time,
Figure SMS_14
for window function->
Figure SMS_15
At the kth moment, outputting a tth time frame, wherein j is an imaginary unit;
corresponding Fourier energy amplitude spectra
Figure SMS_16
The method comprises the following steps:
Figure SMS_17
wherein ,
Figure SMS_18
representing a modulo operation of the complex domain.
Further, the signal features include: an amplitude spectrum and a power spectrum of the signal.
Further, the training process of the kernel pulse signal and noise separation model is as follows:
step A: time-frequency decomposition; firstly, carrying out time-frequency decomposition on known nuclear pulse signals, noise signals and mixed signals of the known nuclear pulse signals and the noise signals, and mapping the signals from a one-dimensional time sequence to a two-dimensional frequency domain space;
and (B) step (B): extracting features; respectively extracting the characteristics of the nuclear pulse signal, the noise signal and the mixed signal of the nuclear pulse signal and the noise signal in a two-dimensional frequency domain;
step C: training a model; b, training a nuclear pulse signal and noise separation model by using the characteristics obtained in the step B and the time-frequency masking of each signal;
step D: model preservation; finally, a kernel pulse signal and noise separation model capable of separating noise and kernel pulse signals is obtained.
Further, the step S4 includes:
inputting the signal characteristics into a trained nuclear pulse signal and noise separation model, and respectively predicting to obtain separation targets of noise and signals;
and obtaining the waveform signal of the nuclear pulse by the separation target through short-time inverse Fourier transform.
Further, the inverse short-time Fourier transform is performed from
Figure SMS_19
Reconstruction->
Figure SMS_20
Separation of nuclear pulse signals is achieved by estimating the short-time Fourier transform coefficients of the target nuclear pulse, using +.>
Figure SMS_21
To represent the Fourier transform coefficient of the f frequency band of the clean target signal in the mixed signal in the t time frame, then the waveform of the target signal>
Figure SMS_22
Can be calculated by short-time inverse fourier transform:
Figure SMS_23
。
further, the preprocessing includes: signal synthesis, trapezoid forming, stacking judgment and peak extraction.
Compared with the prior art, the invention has the beneficial effects that:
1. the method for discriminating the nuclear pulse signal and the noise signal in the complex environment can effectively discriminate the nuclear pulse signal and the noise signal by adopting a method based on machine learning, thereby reducing the acquisition of the noise signal.
2. The nuclear pulse signal and noise signal discrimination method in complex environment can improve the energy resolution of the detector by reducing the acquisition of the noise signal by the detector; particularly, when the nuclear detector needs long-time measurement, the influence of baseline drift on the detector is reduced, and high-precision energy spectrum measurement is realized.
Drawings
FIG. 1 is a flow chart of a method for discriminating between nuclear pulse signals and noise signals in a complex environment;
FIG. 2 is a flow chart of a training process of a kernel signal and noise separation model;
fig. 3 is a general block diagram of a nuclear detector system.
Detailed Description
It is noted that relational terms such as "first" and "second", and the like, are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The features and capabilities of the present invention are described in further detail below in connection with examples.
Example 1
Referring to fig. 1, a method for discriminating a nuclear pulse signal from a noise signal in a complex environment specifically includes the following steps:
step S1: performing FIR low-pass filtering on the collected original signals;
step S2: performing time-frequency decomposition on the filtered signals, and mapping the signals from a one-dimensional time sequence to a two-dimensional frequency domain space;
step S3: extracting signal characteristics in a two-dimensional frequency domain space; preferably, the signal features include: amplitude spectrum and power spectrum of the signal; the method comprises the steps of extracting a basic unit of a characteristic, wherein the basic unit of the characteristic extraction can be divided into a time-frequency unit level characteristic and a frame level characteristic;
the feature of the time frequency unit level is extracted from one time frequency, the granularity of the extracted feature is finer, and more tiny details can be focused, but the description of the global property and the integrity of the signal is lacking; the characteristics of the frame level are extracted from a frame of signal, the granularity of the characteristics of the level is larger, the space-time structure of the signal can be grasped, and the global property and the integrity are better;
step S4: inputting the signal characteristics into a trained nuclear pulse signal and noise separation model, outputting the nuclear pulse signal and noise separation model, and obtaining the nuclear pulse signal and the noise signal by inverse transformation and combining the phases of signals corresponding to the signal characteristics;
step S5: preprocessing the obtained nuclear pulse signal;
step S6: and extracting the amplitude of the preprocessed nuclear pulse signal, and generating an energy spectrum according to the amplitude.
In this embodiment, specifically, the step S1 includes:
acquiring an original signal by using an ADC, wherein the original signal comprises: a nuclear pulse signal, a noise signal, and a mixed signal of the two;
the original signal is subjected to FIR low-pass filtering, and the filtering formula is as follows:
Figure SMS_24
wherein ：
Figure SMS_25
for the input signal at time n +.>
Figure SMS_26
For the input signal at time n-k, < >>
Figure SMS_27
For FIR filter coefficients>
Figure SMS_28
For the filtered signal, N represents the number of taps of the FIR filter, < >>
Figure SMS_29
For the filter order, k is a positive integer and k +.>
Figure SMS_30
N is a positive integer and is expressed as a certain time.
In this embodiment, specifically, the step S2 includes:
is provided with
Figure SMS_31
Is a real symmetric window function, +.>
Figure SMS_32
Is a one-dimensional time domain signal->
Figure SMS_33
In->
Figure SMS_34
Time frame, th->
Figure SMS_35
Short-time fourier transform coefficients for the individual frequency bands:
Figure SMS_36
wherein: k is a positive integer, which represents a certain time,
Figure SMS_37
for window function->
Figure SMS_38
At the kth moment, outputting a tth time frame, wherein j is an imaginary unit;
corresponding Fourier energy amplitude spectra
Figure SMS_39
The method comprises the following steps:
Figure SMS_40
wherein ,
Figure SMS_41
representing a modulo operation of the complex domain.
Referring to fig. 2, in this embodiment, specifically, the training process of the kernel signal and noise separation model is as follows:
step A: time-frequency decomposition; firstly, carrying out time-frequency decomposition on known nuclear pulse signals, noise signals and mixed signals of the known nuclear pulse signals and the noise signals, and mapping the signals from a one-dimensional time sequence to a two-dimensional frequency domain space;
and (B) step (B): extracting features; respectively extracting the characteristics of the nuclear pulse signal, the noise signal and the mixed signal of the nuclear pulse signal and the noise signal in a two-dimensional frequency domain;
step C: training a model; b, training a nuclear pulse signal and noise separation model by using the characteristics obtained in the step B and the time-frequency masking of each signal; the obtained characteristics and the time-frequency mask of each signal are input into a convolutional neural network;
step D: model preservation; finally, a kernel pulse signal and noise separation model capable of separating noise and kernel pulse signals is obtained; and obtaining a kernel pulse signal and noise separation model from the noisy characteristic to the separation signal.
In this embodiment, specifically, the step S4 includes:
inputting the signal characteristics into a trained nuclear pulse signal and noise separation model, and respectively predicting to obtain separation targets of noise and signals;
and obtaining the waveform signal of the nuclear pulse by the separation target through short-time inverse Fourier transform.
In the present embodiment, in particular, the inverse short-time Fourier transform is performed from
Figure SMS_42
Reconstruction->
Figure SMS_43
Separation of nuclear pulse signals is achieved by estimating the short-time Fourier transform coefficients of the target nuclear pulse, using +.>
Figure SMS_44
To represent the Fourier transform coefficient of the f frequency band of the clean target signal in the mixed signal in the t time frame, then the waveform of the target signal>
Figure SMS_45
Can be calculated by short-time inverse fourier transform:
Figure SMS_46
if the phase is not considered, the noise signal separation process can be converted into the problem of estimating the amplitude spectrum of the target nuclear pulse signal, and according to the amplitude information of the estimated target signal and the phase information of the mixed input signal, the estimated waveform of the target nuclear pulse signal can be obtained through short-time Fourier inverse transformation
Figure SMS_47
。
In this embodiment, specifically, the preprocessing includes: signal synthesis, trapezoid forming, stacking judgment and peak extraction;
wherein, the trapezoidal shaping algorithm is designed as follows:
Figure SMS_48
in the formula ：
Figure SMS_51
is an input signal sequence; />
Figure SMS_54
For outputting a signal sequence; />
Figure SMS_55
，/>
Figure SMS_50
，/>
Figure SMS_53
The length of the inclined side and the flat top of the trapezoid is determined and is adjustable; />
Figure SMS_56
，/>
Figure SMS_57
。/>
Figure SMS_49
The value of (2) is obtained by collecting the actual output signal of the detector for a plurality of times, then performing curve fitting, and finally taking the average value of the plurality of times. />
Figure SMS_52
Is the sampling period of the ADC.
The pile-up judging algorithm is designed as follows:
setting the resolution time of the system to
Figure SMS_58
When the time interval of two pulses is smaller than +.>
Figure SMS_59
When the pulse is to be discarded; when the time interval of two pulses is larger than +.>
Figure SMS_60
When both pulses will be recorded as valid pulses;
the peak extraction algorithm is as follows:
and carrying out summation and averaging on the flat top part of the trapezoidal shaping algorithm, and judging whether the flat top part is an effective amplitude value according to the output of the accumulation judging algorithm.
Example two
The second embodiment is a specific deployment process of the nuclear pulse signal and noise signal discrimination method in the nuclear detector system in the complex environment.
Please refer to fig. 3.
In the first step, a weak nuclear pulse signal is acquired and amplified by a front end detector near the radioactive source.
And secondly, acquiring the amplified signal by using a high-speed ADC.
Thirdly, performing FIR low-pass filtering on the signals acquired by the ADC.
Fourth, the input mixed signal is subjected to time-frequency decomposition, and the signal is mapped from a one-dimensional time sequence to a two-dimensional frequency domain space.
And fifthly, extracting the characteristics of the input signal in a two-dimensional frequency domain.
And sixthly, inputting the characteristics into a trained kernel pulse signal and noise separation model, and respectively predicting to obtain separation targets of noise and signals.
Seventh, the separation target obtained in the previous step is subjected to inverse transformation (inverse fourier transformation) to obtain a waveform signal of the nuclear pulse.
Eighth, designing a trapezoidal forming algorithm.
Figure SMS_61
in the formula ：
Figure SMS_62
is an input signal sequence; />
Figure SMS_63
For outputting a signal sequence. />
Figure SMS_64
，/>
Figure SMS_65
，/>
Figure SMS_66
The length of the hypotenuse and the flat top of the trapezoid is determined and is adjustable.
And ninth, stacking judgment algorithm design.
Setting the resolution time of the system to
Figure SMS_67
When the time interval of two pulses is smaller than +.>
Figure SMS_68
When the pulse is to be discarded; when the time interval of two pulses is larger than +.>
Figure SMS_69
When both pulses are to be recorded as valid pulses.
And tenth, designing a peak value extraction algorithm.
And (3) carrying out summation and average on the flat top part of the trapezoidal shaping algorithm, and judging whether the flat top part is an effective amplitude value according to the output of the ninth step.
And eleventh step, designing an energy spectrum storage algorithm.
The extracted peaks are used to generate a spectrum and stored in a dual-ported ram.
And twelfth, designing a data transmission algorithm.
And reading the energy spectrum data stored in the ram, transmitting the energy spectrum data to the upper computer through the data transmission module, and receiving and displaying the energy spectrum by the upper computer.
The foregoing examples merely represent specific embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the present application. It should be noted that, for those skilled in the art, several variations and modifications can be made without departing from the technical solution of the present application, which fall within the protection scope of the present application.
This background section is provided to generally present the context of the present invention and the work of the presently named inventors, to the extent it is described in this background section, as well as the description of the present section as not otherwise qualify as prior art at the time of filing, are neither expressly nor impliedly admitted as prior art against the present invention.

Claims (3)

1. A method for discriminating a nuclear pulse signal and a noise signal in a complex environment is characterized by comprising the following steps:
step S1: performing FIR low-pass filtering on the collected original signals;
step S2: performing time-frequency decomposition on the filtered signals, and mapping the signals from a one-dimensional time sequence to a two-dimensional frequency domain space;
step S3: extracting signal characteristics in a two-dimensional frequency domain space;
step S4: inputting the signal characteristics into a trained nuclear pulse signal and noise separation model, outputting the nuclear pulse signal and noise separation model, and obtaining the nuclear pulse signal and the noise signal by inverse transformation and combining the phases of signals corresponding to the signal characteristics;
step S5: preprocessing the obtained nuclear pulse signal;
step S6: extracting the amplitude of the preprocessed nuclear pulse signal, and generating an energy spectrum according to the amplitude;
the step S1 includes:
acquiring an original signal by using an ADC, wherein the original signal comprises: a nuclear pulse signal, a noise signal, and a mixed signal of the two;
the original signal is subjected to FIR low-pass filtering, and the filtering formula is as follows:
Figure QLYQS_1
wherein ：
Figure QLYQS_2
for the input signal at time n +.>
Figure QLYQS_3
For the input signal at time n-k, < >>
Figure QLYQS_4
For FIR filter coefficients>
Figure QLYQS_5
For the filtered signal, N represents the number of taps of the FIR filter, < >>
Figure QLYQS_6
For the filter order, k is a positive integer and k +.>
Figure QLYQS_7
N is a positive integer, and is expressed as a certain moment;
the step S2 includes:
is provided with
Figure QLYQS_8
Is a real symmetric window function, +.>
Figure QLYQS_9
Is a one-dimensional time domain signal->
Figure QLYQS_10
In->
Figure QLYQS_11
Time frame, th->
Figure QLYQS_12
Short-time fourier transform coefficients for the individual frequency bands:
Figure QLYQS_13
wherein: k is a positive integer, which represents a certain time,
Figure QLYQS_14
for window function->
Figure QLYQS_15
At the kth moment, outputting a tth time frame, wherein j is an imaginary unit;
corresponding Fourier energy amplitude spectra
Figure QLYQS_16
The method comprises the following steps:
Figure QLYQS_17
wherein ,
Figure QLYQS_18
representing a modulus operation of the complex field;
the signal characteristics include: amplitude spectrum and power spectrum of the signal;
the training process of the kernel pulse signal and noise separation model is as follows:
step A: time-frequency decomposition; firstly, carrying out time-frequency decomposition on known nuclear pulse signals, noise signals and mixed signals of the known nuclear pulse signals and the noise signals, and mapping the signals from a one-dimensional time sequence to a two-dimensional frequency domain space;
and (B) step (B): extracting features; respectively extracting the characteristics of the nuclear pulse signal, the noise signal and the mixed signal of the nuclear pulse signal and the noise signal in a two-dimensional frequency domain;
step C: training a model; b, training a nuclear pulse signal and noise separation model by using the characteristics obtained in the step B and the time-frequency masking of each signal;
step D: model preservation; finally, a kernel pulse signal and noise separation model capable of separating noise and kernel pulse signals is obtained;
the kernel pulse signal and noise separation model is a convolutional neural network.
2. The method for discriminating between nuclear pulse signals and noise signals in a complex environment according to claim 1, wherein said step S4 comprises:
inputting the signal characteristics into a trained nuclear pulse signal and noise separation model, and respectively predicting to obtain separation targets of noise and signals;
and obtaining the waveform signal of the nuclear pulse by the separation target through short-time inverse Fourier transform.
3. The method for discriminating between nuclear pulse signals and noise signals in a complex environment according to claim 1, wherein said preprocessing comprises: trapezoidal forming, stacking and discarding, and peak value extraction.
CN202310173699.0A 2023-02-28 2023-02-28 Nuclear pulse signal and noise signal discrimination method in complex environment Active CN115856987B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310173699.0A CN115856987B (en) 2023-02-28 2023-02-28 Nuclear pulse signal and noise signal discrimination method in complex environment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310173699.0A CN115856987B (en) 2023-02-28 2023-02-28 Nuclear pulse signal and noise signal discrimination method in complex environment

Publications (2)

Publication Number Publication Date
CN115856987A CN115856987A (en) 2023-03-28
CN115856987B true CN115856987B (en) 2023-05-02

Family

ID=85659269

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310173699.0A Active CN115856987B (en) 2023-02-28 2023-02-28 Nuclear pulse signal and noise signal discrimination method in complex environment

Country Status (1)

Country Link
CN (1) CN115856987B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116389214B (en) * 2023-06-05 2023-08-08 四川科冠电子有限公司 Noise reduction method, noise reduction terminal and medium suitable for voltage power line carrier communication
CN117828279A (en) * 2024-03-04 2024-04-05 中国电子产品可靠性与环境试验研究所（（工业和信息化部电子第五研究所）（中国赛宝实验室）） Pulse signal measurement data processing method and device and computer equipment

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3866165A1 (en) * 2020-02-14 2021-08-18 System One Noc & Development Solutions, S.A. Method for enhancing telephone speech signals based on convolutional neural networks
CN113673312A (en) * 2021-07-06 2021-11-19 太原理工大学 Radar signal intra-pulse modulation identification method based on deep learning

Family Cites Families (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4175258A (en) * 1978-07-28 1979-11-20 The United States Of America As Represented By The United States Department Of Energy High level white noise generator
FR2778467B1 (en) * 1998-05-11 2000-06-16 Christian Jeanguillaume IMPROVED HIGH-SENSITIVITY GAMMA CAMERA SYSTEM
IL198502A0 (en) * 2009-05-03 2010-02-17 Hava Bar Noy Collapsible stretchar dragged on wheels
CN103777228B (en) * 2014-02-26 2016-03-16 成都理工大学 Based on the digital core pulse signal Gauss manufacturing process of iir filter
CN106896395B (en) * 2017-04-27 2019-01-11 西南科技大学 A kind of detection device of the faint emission signal based on sparse signal representation
CN109839612B (en) * 2018-08-31 2022-03-01 大象声科（深圳）科技有限公司 Sound source direction estimation method and device based on time-frequency masking and deep neural network
CN111292762A (en) * 2018-12-08 2020-06-16 南京工业大学 Single-channel voice separation method based on deep learning
US11826129B2 (en) * 2019-10-07 2023-11-28 Owlet Baby Care, Inc. Heart rate prediction from a photoplethysmogram
CN111402395B (en) * 2020-02-17 2023-07-04 西安电子科技大学 CNN correction-based passive polarization three-dimensional reconstruction method
CN112764082B (en) * 2020-12-08 2023-05-23 武汉第二船舶设计研究所(中国船舶重工集团公司第七一九研究所) FPGA-based nuclear pulse digital forming sampling method
CN115310472A (en) * 2021-05-07 2022-11-08 四川轻化工大学 Nuclear pulse peak sequence-based one-dimensional convolution neural network nuclide identification method
CN113970420B (en) * 2021-10-13 2022-04-05 中国科学院力学研究所 Deep learning-based shock tunnel force measurement signal frequency domain analysis method
CN115034254A (en) * 2022-03-22 2022-09-09 四川轻化工大学 Nuclide identification method based on HHT (Hilbert-Huang transform) frequency band energy features and convolutional neural network

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3866165A1 (en) * 2020-02-14 2021-08-18 System One Noc & Development Solutions, S.A. Method for enhancing telephone speech signals based on convolutional neural networks
CN113673312A (en) * 2021-07-06 2021-11-19 太原理工大学 Radar signal intra-pulse modulation identification method based on deep learning

Also Published As

Publication number Publication date
CN115856987A (en) 2023-03-28

Similar Documents

Publication Publication Date Title
CN115856987B (en) Nuclear pulse signal and noise signal discrimination method in complex environment
CN101882964A (en) De-noising method of transient electromagnetic detecting echo signal
Demir et al. Hyperspectral image classification using denoising of intrinsic mode functions
Reddy et al. Footstep detection and denoising using a single triaxial geophone
CN116773894A (en) Collector power failure detection system and method thereof
Liu et al. Maximum correntropy criterion-based blind deconvolution and its application for bearing fault detection
Yang et al. Radar emitter signal recognition based on time-frequency analysis
US6772182B1 (en) Signal processing method for improving the signal-to-noise ratio of a noise-dominated channel and a matched-phase noise filter for implementing the same
Conru et al. Time-frequency detection using Gabor filter bank and Viterbi based grouping algorithm
CN110287853B (en) Transient signal denoising method based on wavelet decomposition
CN111474581B (en) Transient weak signal detection method based on nonlinear time extrusion time-frequency transformation
CN110542927B (en) Variable window weighted seismic data spike noise suppression method
CN112285793B (en) Magnetotelluric denoising method and system
Sattar et al. On detection using filter banks and higher order statistics
CN102799757A (en) Weak signal extraction method for removing interferences of strong trend term and transient-state pulse
Hussain et al. A novel wavelet thresholding method for adaptive image denoising
Tibuleac et al. Automatic determination of secondary seismic phase arrival times using wavelet transforms
Dwivedi et al. A robust energy features estimation for detection and classification of power quality disturbances
CN111008356A (en) WTSVD algorithm-based background-subtracted gamma energy spectrum set analysis method
CN113238206B (en) Signal detection method and system based on decision statistic design
CN113158797B (en) Micro-seismic data denoising method, system, electronic equipment and storage medium
CN116599606B (en) Spread spectrum signal receiving method and system based on channelized weighted cross-correlation processing
Zhang et al. The study of time domain denoising for the time-frequency electromagnetic method prospecting data
CN111835495B (en) Method and system for detecting reference signal, readable storage medium and electronic device
Li et al. Robust unsupervised Tursiops aduncus whistle enhancement based on complete ensembled empirical optimal envelope local mean decomposition with adaptive noise

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant