Frequency-domain Detection Method and System for Signals in Colored Noise Environment Based on Spectrum Envelope Extraction

Abstract

This application belongs to the field of radio electronics and signal processing technology, and specifically relates to a method and system for detecting signals in the frequency domain in a colored noise environment based on spectral envelope extraction, including a signal acquisition unit, a signal conversion unit, a signal judgment unit, and a signal output unit; The present invention is a research summary of an algorithm for detecting signals in the frequency domain in a colored noise environment based on spectral envelope extraction. This algorithm is used for blind signal detection under broadband reception with an uneven noise spectrum and has good robustness. It can be used in low signal-to-noise ratio scenarios and has a high signal detection accuracy.

Description

Technical Field

[0001] This application belongs to the technical fields of radio electronics and signal processing, and particularly relates to a frequency-domain detection method and system for signals in a colored noise environment based on spectral envelope extraction. Background Art

[0002] Signal frequency-domain detection algorithms are a class of techniques that analyze signals by converting them from the time domain to the frequency domain to detect, identify, or extract specific frequency components. Their core purpose is to use frequency-domain characteristics (such as spectrum, power spectrum, phase spectrum, etc.) to reveal hidden information in signals, which may be difficult to directly observe in the time domain, and achieve purposes such as frequency component identification, narrowband or broadband signal separation, and modulation analysis.

[0003] Traditional signal detection techniques mainly include energy detection, cyclostationary feature detection, matched filtering, eigenvalue detection, singular value detection, etc. These detection methods have their own advantages and disadvantages in different scenarios: The energy detection method detects signals by comparing the difference between the noise energy and the signal energy in the environment. Due to its advantages of not requiring any prior information and low computational complexity, it is widely used and is a blind signal detection method. However, the algorithm performance is limited by energy information, sensitive to low signal-to-noise ratio signals, and the noise uncertainty leads to a decrease in the detection probability. The cyclostationary feature detection method can effectively extract the periodic or cyclic characteristics in signals by analyzing the cyclostationary characteristics of signals, thereby distinguishing modulated signals, interference signals, and stationary noise, and distinguishing signals from noise through the spectral correlation function, with strong anti-interference ability. However, the algorithm has high complexity and poor timeliness. Matched filtering detection is used for theoretically optimal sensing performance, but it requires prior knowledge of the specific information of the signal to be detected and is not a blind detection method. Eigenvalue-based detection does not require any prior information and is less affected by noise instability, but has a large computational complexity and requires a large number of sampling points. Summary of the Invention

[0004] The object of the present invention is to solve the above problems existing in the prior art, and propose a signal detection method applicable to different frequency band widths in a colored noise environment, and having advantages such as robustness and high detection accuracy. Its technical solution is as follows:

[0005] A frequency-domain detection method for signals in a colored noise environment based on spectral envelope extraction, comprising the following steps:

[0006] S1. Perform DFT on the obtained sampled signal to convert it from the time domain to the frequency domain to obtain the spectrum of the signal , and then extract the envelope of the spectrum to obtain the envelope data of the spectrum ;

[0007] S2. Traverse the signal spectrum envelope obtained in step S1 to find the extreme points and update the corresponding maximum value set , minimum value set , and perform step S3 or step S4 processing according to the type of extreme point;

[0008] S3. If the current extreme point is a minimum value point, it is necessary to compare with the spectrum amplitude of the previous maximum value point, calculate the spectrum amplitude difference between the two and the set threshold to determine the decision flag value;

[0009] S4. If the current extreme point is a maximum value point, it is necessary to compare with the spectrum amplitude of the previous maximum value point, calculate the spectrum amplitude difference between the two and the set threshold ; and then combine the decision flag value in step S3 to decide whether to jump out of the current step and continue to traverse the envelope to find extreme points, or continue to judge the amplitude difference between the current maximum value point and the previous minimum value point and the threshold ;

[0010] S5. For the signal frequency points that meet the threshold decision requirements in step S4 , update them to the set of the start and end positions of the signal frequency band , and finally until the traversal of the signal spectrum envelope data is completed, output the detection result and the detection ends.

[0011] Preferably, the specific steps of step S1 are as follows:

[0012] S11. Perform DFT on the obtained discrete sampling signal to convert it from the time domain to the frequency domain to obtain the spectrum of the signal , and the expression is:

[0013] ;

[0014] where represents the th frequency point of the spectrum, is the number of spectrum points;

[0015] S12. Extract the envelope of the spectrum , set the window size to , and the extraction idea is to take the maximum value of the spectrum amplitude for each window, and slide and extract all the spectrum data in turn. The spectrum envelope data of the th point is calculated as:

[0016] ;

[0017] Among them, the value range of is

[0018] Preferably, the decision formula for the extreme point in step S2 is as follows:

[0019] ;

[0020] Among them, represents the maximum point of the envelope, represents the minimum point of the envelope, and belongs to the corresponding set: , 。

[0021] Preferably, step S3 includes:

[0022] S3.1. Set the extreme point sets , , if the th extreme point found currently is the minimum point , then it is necessary to compare with the spectral amplitude of the previous maximum point , and calculate the amplitude difference The expression for

[0023] is:

[0024] Among them, represents taking the absolute value;

[0025] S3.2. Compare the amplitude difference with the set threshold to determine the value of the decision flag , and then continue to traverse the envelope data to find extreme points; the expression for the decision flag is:

[0026] 。

[0027] Preferably, in step S4, set the extreme point sets: ;

[0028] ;

[0029] If the th extreme point found currently is the maximum point , then it is necessary to compare with the spectral amplitude of the previous maximum point , and calculate the amplitude difference The expression for

[0030] is:

[0031] Preferably, in step S4, the amplitude difference is compared with a set threshold as follows:

[0032] If , it is considered that the current maximum point may be the signal frequency point, then step S4 is exited and step S2 is entered to continue traversing the envelope data to find the extreme points;

[0033] If , then according to the value of the decision flag , there are different processing steps:

[0034] If the decision flag , then step S5 is performed;

[0035] If the decision flag , the current th maximum point is compared with the spectrum amplitude of the previous minimum point to calculate the amplitude difference , and the expression of the amplitude difference

[0036] is:

[0037] The amplitude difference is compared with the set threshold . If , then this step is exited and step S2 is entered to continue traversing the envelope data to find the extreme points; if , then step S5 is performed.

[0038] A frequency-domain detection system for signals in a colored noise environment based on spectrum envelope extraction includes a signal acquisition unit, a signal conversion unit, a signal judgment unit, and a signal output unit;

[0039] Signal acquisition unit: acquires the sampling signal to be processed;

[0040] Signal conversion unit: converts the sampling signal from the time domain to the frequency domain to obtain the spectrum of the signal ;

[0041] Signal judgment unit: extracts the envelope of the spectrum to obtain the envelope data of the spectrum , traverses the signal spectrum envelope to find the extreme points and updates the corresponding maximum set , minimum set ; introduces a dual-threshold joint decision on the signal frequency band range;

[0042] Signal output unit: visually outputs the result.

[0043] Preferably, the signal judgment unit: if the current extreme point is a minimum point, it is necessary to compare it with the spectral amplitude of the previous maximum point, calculate the difference in spectral amplitude between the two and the set threshold to determine the size relationship, so as to determine the decision flag value;

[0044] For the signal frequency points that meet the threshold decision requirements , update them to the set of the start and end positions of the signal frequency band , and finally until the traversal of the signal spectrum envelope data is completed;

[0045] If the current extreme point is a maximum point, it is necessary to compare it with the spectral amplitude of the previous maximum point, calculate the difference in spectral amplitude between the two and the set threshold to determine the size relationship; combined with the decision flag value, decide whether to jump out of the current step and continue to traverse the envelope to find extreme points, or continue to judge the amplitude difference and threshold between the current maximum point and the previous minimum point size relationship.

[0046] Preferably, the applicable range is communication signals with modulation methods of BPSK, QPSK, 8PSK, GMSK, 16QAM and the noise floor fluctuates randomly, and the signal-to-noise ratio is greater than or equal to 3dB.

[0047] Compared with the prior art, the beneficial effects of this application are as follows:

[0048] 1. The method for detecting the frequency domain of signals in a colored noise environment based on spectral envelope extraction of the present invention is based on the spectral amplitude spectrum of the signal, extracts the upper envelope of the spectrum within the receiver bandwidth, and according to the differences in the signal and noise amplitudes and the noise and noise amplitudes in the upper envelope, it is used for distinction according to the set threshold. The frequency band range of the signal can be obtained in one traversal, and the overall processing steps of the algorithm are few, the complexity is low, and the timeliness is good.

[0049] 2. The method for detecting the frequency domain of signals in a colored noise environment based on spectral envelope extraction of the present invention introduces the step of extracting the spectral envelope in the case of lack of prior information and a colored noise background with uneven frequency domain noise, so that only the relative change amount of the signal and noise spectra at adjacent positions needs to be compared during subsequent frequency domain detection to determine the signal frequency band range, eliminating problems such as signal missed detection and false alarms caused by unreasonable threshold decisions of traditional frequency domain detection algorithms in a colored noise environment, eliminating the influence of colored noise, and at the same time being equally applicable to Gaussian white noise with a flat frequency domain, and having strong robustness.

[0050] 3. The frequency-domain detection method of the colored noise environment signal based on spectrum envelope extraction according to the present invention introduces a dual-threshold joint decision on the signal frequency band range. When both broadband signals and narrowband signals appear within the receiver range, the spectral amplitude values within the bandwidth of the broadband signal in the colored noise environment fluctuate significantly. Traditional algorithms may detect and segment the broadband signal, resulting in incorrect estimation of the number of detected signals, frequency points, etc. The algorithm proposed in the present invention performs a secondary threshold decision on the amplitudes of adjacent signal frequency points, improving the above problems. Simulation experiments show that this method is applicable to the detection of signals with different modulation types and different bandwidths within the receiver bandwidth range.

[0051] 4. The signal detection method of the colored noise environment based on spectrum envelope extraction according to the present invention demonstrates excellent detection performance under the conditions of low signal-to-noise ratio and colored noise background. When the signal-to-noise ratio is as low as 3 dB, the traditional frequency-domain detection algorithm is greatly affected by noise due to the statistical characteristics of the adaptive threshold based on the local window, resulting in a significant decline in detection performance. For communication signals with modulation methods of BPSK, QPSK, 8PSK, GMSK, and 16QAM and random fluctuations in the noise floor, when the signal-to-noise ratio is greater than or equal to 3 dB, the detection accuracy rate reaches 97.9%, the miss detection rate is 2.1%, and the false alarm rate is 0, which is much higher than that of the traditional frequency-domain detection algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is the algorithm flowchart of the present invention;

[0053] Figure 2 is the received signal spectrogram in the algorithm of the present invention;

[0054] Figure 3 is the spectrum extraction envelope diagram of the signal in the algorithm of the present invention;

[0055] Figure 4 is the signal detection result diagram of the algorithm of the present invention under the Gaussian white noise background;

[0056] Figure 5 is the signal detection result diagram of the algorithm of the present invention under the colored noise background;

[0057] Figure 6 is the signal detection result diagram of the algorithm of the present invention under the broadband spectrum. DETAILED DESCRIPTION OF THE INVENTION

[0058] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following provides a detailed description of a frequency-domain detection method and system for colored noise environment signals based on spectrum envelope extraction according to the present invention in combination with the accompanying drawings and specific embodiments.

[0059] The foregoing and other technical contents, features, and effects of the present invention will be clearly presented in the following detailed description in conjunction with the accompanying drawings. Through the description of the specific embodiments, a more in-depth and specific understanding of the technical means and effects adopted by the present invention to achieve the predetermined purpose can be obtained. However, the attached drawings are only for reference and illustration, and are not used to limit the technical solution of the present invention.

[0060] Embodiment 1:

[0061] As Figure 1 shown, the frequency-domain detection method of the colored noise environment signal based on spectral envelope extraction in this embodiment includes:

[0062] Step 1: Perform DFT on the obtained discrete sampling signal to convert it from the time domain to the frequency domain, obtaining the spectrum of the signal. Then extract the envelope of the spectrum to obtain the envelope data of the spectrum.

[0063] Specifically, Step 1 includes:

[0064] Step 1.1: Perform DFT on the obtained discrete sampling signal to convert it from the time domain to the frequency domain, obtaining the spectrum of the signal, and the expression is:

[0065] ;

[0066] where represents the th frequency point of the spectrum, and is the number of spectrum points.

[0067] Step 1.2: Extract the envelope of the spectrum , set the window size to , and the extraction idea is to take the maximum value of the spectrum amplitude for each window, and slide and extract all the spectrum data in turn. The calculation formula for the spectrum envelope data of the th point is:

[0068] ;

[0069] where has a value range of .

[0070] Specifically, in this embodiment, converting the signal from the time domain to the frequency domain and detecting the spectrum of the signal can be applicable to the scenario of low signal-to-noise ratio. Extracting the envelope further improves the situation of large fluctuations in the noise floor. Please refer to Figure 2 , Figure 3, Figure 2 is the received signal spectrogram in the algorithm of the present invention, Figure 3 is the envelope diagram of the signal spectrum extraction in the algorithm of the present invention. It can be seen from the comparison that after the spectrum extraction envelope, the amplitude smoothing degree of the noise part is improved, and the spectrum amplitude difference between the signal and the noise still exists. In the background of colored noise, the algorithm pays more attention to the amplitude difference between adjacent spectrum points, reducing the interference of the noise base thickness on the judgment of the signal frequency point amplitude.

[0071] Step 2. Traverse the signal spectrum envelope obtained in Step 1 to find the extreme points and update the corresponding maximum value set and minimum value set , and perform Step 3 or Step 4 processing according to the type of extreme point. The decision formula for the extreme point is as follows:

[0072] ;

[0073] wherein, represents the maximum point of the envelope, represents the minimum point of the envelope, belonging to the corresponding sets: , .

[0074] Specifically, in this embodiment, since there is an obvious difference in the spectrum amplitude between the frequency point position of the signal and the amplitudes of other adjacent spectra, it is necessary to traverse the envelope to find the extreme points, and no special processing is performed on the monotonic positions of the envelope. The minimum and maximum points of the envelope appear alternately, and the processing flow of Step 3 or Step 4 is selected according to the type of extreme point.

[0075] Step 3. If the current extreme point is a minimum point, it is necessary to compare the spectrum amplitude with that of the previous maximum point, calculate the spectrum amplitude difference between the two and the set threshold to determine the value of the decision flag .

[0076] Specifically, Step 3 includes:

[0077] Step 3.1. Set the extreme point sets , , if the th extreme point found currently is a minimum point , then it is necessary to compare the spectrum amplitude with that of the previous maximum point , and the expression for calculating the amplitude difference is:

[0078] ;

[0079] wherein, Indicates taking the absolute value.

[0080] Step 3.2: Compare the amplitude difference with the set threshold and, which can be adjusted according to the signal quality received by the receiver. Generally, it can be set to to determine the decision flag . Then continue to traverse the envelope data to find the extreme points. The expression of the decision flag is:

[0081] ;

[0082] Specifically, in this embodiment, the set threshold represents the minimum amplitude difference between the spectral amplitude of the signal frequency point position and the noise floor, and it cannot be set too low, otherwise the false alarm probability of detection is high.

[0083] Step 4: If the current extreme point is a maximum point, it is necessary to compare it with the spectral amplitude of the previous maximum point, calculate the spectral amplitude difference between the two and the set threshold . Then, combined with the value of the decision flag in Step 3, decide whether to jump out of the current step and continue to traverse the envelope to find the extreme point, or continue to judge the amplitude difference between the current maximum point and the previous minimum point and the threshold .

[0084] Specifically, Step 4 includes:

[0085] Step 4.1: Set the extreme point set , . If the th extreme point found currently is a maximum point , it is necessary to compare it with the spectral amplitude of the previous maximum point . The expression for calculating the amplitude difference is:

[0086] ;

[0087] Step 4.2: Compare the amplitude difference with the set threshold and, which can be adjusted according to the signal quality received by the receiver. Generally, it can be set to .

[0088] If , jump out of Step 4 and go to Step 2 to continue traversing the envelope data to find the extreme point; if , different processing steps are taken according to the value of the decision flag .

[0089] Step 4.3: If the decision flag is satisfied, then proceed to Step 5; if the decision flag is satisfied, take the currently found th maximum point and compare its spectral amplitude with that of the previous minimum point to calculate the amplitude difference whose expression is:

[0090] ;

[0091] Compare the amplitude difference with the set threshold . If , then jump out of this step and proceed to Step 2 to continue traversing the envelope data to find extreme points; if , then proceed to Step 5.

[0092] Specifically, in this embodiment, the set threshold represents the spectral amplitude difference between adjacent possible signal frequency point positions. The judgment logic here can prevent the broadband signal from being segmented during the detection process.

[0093] Step 5: For the signal frequency points that meet the threshold decision requirements in Step 4 , update them to the set of the start and end positions of the signal frequency band. Finally, until the traversal of the signal spectrum envelope data is completed, output the detection result and the detection ends.

[0094] The above method is based on the extraction of the signal spectrum envelope. While reducing the fluctuation of the noise floor, it retains the spectral amplitude difference between the signal frequency point positions and the noise, effectively improving the interference of the noise floor thickness on signal detection in the low signal-to-noise ratio scenario; and traversing the signal spectrum envelope to find extreme points, paying more attention to the spectral amplitude difference between adjacent frequency points, without setting a unified threshold for all data, having strong applicability, and still having good results in the frequency domain detection of signals with large noise floor fluctuations in the colored noise environment.

[0095] Embodiment 2:

[0096] This embodiment explains the performance of the signal frequency domain detection method based on spectral envelope extraction in Embodiment 1 through simulation experiments.

[0097] Please refer to Figures 4-6 , Figure 4 for the simulation results of the algorithm performance. Figure 5 is the signal detection result graph of the algorithm of the present invention under the Gaussian white noise background, Figure 6This is the signal detection result graph of the algorithm of the present invention under broadband spectrum. The abscissa is the number of frequency index points, and the ordinate is the spectrum amplitude value, with the unit of dB.

[0098] Figure 4 Simulation parameter settings: The signal-to-noise ratio is 3 dB, the modulation method is BPSK, the signal carrier frequency is 1.8 GHz, the number of sampling points is 4096, the noise spectrum is flat, and it is Gaussian white noise; Figure 5 Simulation parameter settings: The signal-to-noise ratio is 3 dB, the modulation method is BPSK, the signal frequency point is 1.76 GHz, the number of sampling points is 4096, the noise spectrum fluctuates, and it is the uneven colored noise characteristic presented after the signal generator passes through the actual channel simulator; Figure 6 This is the blind signal spectrum of broadband real-time collected data. The number of sampling points is 10001, the receiver bandwidth is 1.2 GHz, the frequency resolution is 120 KHz, the number of statistical signals is 2, there is one broadband signal and one narrowband signal.

[0099] Such as Figure 4 、 Figure 5 The detection situations of signals under Gaussian white noise background and colored noise background are compared. The two ends of the spectrum data are the signal spectrum attenuation caused by the band-pass filter, accounting for about 200 frequency points at the front and back respectively. When the spectrum is flat, the algorithm correctly detects the signal frequency band range at the signal frequency point; when the spectrum is uneven, the signal spectrum amplitude value significantly increases at the frequency index 1460, and there is random noise frequency point interference at the frequency index 1776. The algorithm correctly detects the signal frequency band range. The present invention is applicable to both colored noise and Gaussian white noise environments and has strong robustness.

[0100] To verify the feasibility of the present invention in actual engineering, the following uses the algorithm proposed by the present invention to perform spectrum sensing on a certain real-time collected data. Figure 6 This is the signal detection result graph of the algorithm of the present invention under broadband spectrum. There are both broadband signals and narrowband signals in the spectrum graph, and the spectrum amplitude fluctuates greatly within the broadband signal interval. At about the frequency index 5000, the spectrum amplitude is extremely low, approximately equal to the noise spectrum amplitude value. The detection results show that the algorithm proposed by the present invention can effectively prevent the phenomenon of broadband signal detection segmentation, and the narrowband signals can also be correctly detected. The algorithm is applicable to signals with different frequency band widths and different spectrum amplitudes.

[0101] To verify the reliability of the algorithm proposed by the present invention, the signal detection algorithms of five different modulation methods (BPSK, QPSK, 8PSK, GMSK, 16QAM) such as phase modulation and amplitude modulation are verified. 1200 packets of data with different code rates and signal-to-noise ratios are simulated, and the correct detection probability is statistically calculated. When the signal-to-noise ratio is greater than or equal to 3 dB, the detection correct rate reaches 97.9%, the missed detection rate is 2.1%, and the false alarm rate is 0, which is much higher than the traditional frequency domain detection algorithm.

[0102] A frequency-domain detection system for signals in a colored noise environment based on spectral envelope extraction, comprising a signal acquisition unit, a signal conversion unit, a signal judgment unit and a signal output unit;

[0103] Signal acquisition unit: acquire the sampled signal to be processed;

[0104] Signal conversion unit: convert the sampled signal from the time domain to the frequency domain to obtain the spectrum of the signal ;

[0105] Signal judgment unit: extract the envelope of the spectrum to obtain the envelope data of the spectrum , traverse the signal spectrum envelope to find the extreme points and update the corresponding maximum set and minimum set ; introduce double-threshold joint decision for the signal frequency band range;

[0106] If the current extreme point is a minimum point, it is necessary to compare it with the spectral amplitude of the previous maximum point, calculate the difference between their spectral amplitudes and the set threshold to determine the decision flag value;

[0107] For the signal frequency points that meet the threshold decision requirements , update them to the set of the start and end positions of the signal frequency band , and finally until the traversal of the signal spectrum envelope data is completed;

[0108] If the current extreme point is a maximum point, it is necessary to compare it with the spectral amplitude of the previous maximum point, calculate the difference between their spectral amplitudes and the set threshold ; and then combine the value of the decision flag to decide whether to jump out of the current step and continue to traverse the envelope to find extreme points, or to continue to judge the amplitude difference between the current maximum point and the previous minimum point and the threshold relationship.

[0109] Signal output unit: visually output the results.

[0110] For communication signals with modulation methods of BPSK, QPSK, 8PSK, GMSK, 16QAM and random fluctuations of the noise floor, when the signal-to-noise ratio is greater than or equal to 3 dB, the detection accuracy rate reaches 97.9%, the missed detection rate is 2.1%, and the false alarm rate is 0, which is much higher than the traditional frequency-domain detection algorithm.

[0111] The above are only embodiments of the present invention, and thus do not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall similarly be included within the patent protection scope of the present invention.

Claims

1. A frequency-domain detection method for signals in a colored noise environment based on spectral envelope extraction, characterized in that It includes the following steps: S1. For the acquired sampling signal perform DFT to transform it from the time domain to the frequency domain, obtaining the spectrum of the signal , and then extract the envelope of the spectrum to obtain the envelope data of the spectrum ; S2. Traverse the signal spectrum envelope obtained in step S1 to find the extreme points and update the corresponding maximum value set and minimum value set , and perform step S3 or step S4 processing according to the type of extreme point; S3. If the current extreme point is a minimum point, it is necessary to compare it with the spectral amplitude of the previous maximum point, calculate the difference in spectral amplitude between the two, and set a threshold to determine the relationship in magnitude, so as to determine the value of the decision flag ; S3.

1. Set the extreme point set , , if the th extreme point found currently is a minimum extreme point , it is necessary to compare with the spectral amplitude of the previous maximum extreme point and calculate the amplitude difference . The expression is as follows: ; Among them, represents taking the absolute value; S3.

2. Compare the amplitude difference with the set threshold to determine the value of the decision flag , and then continue to traverse the envelope data to find the extreme points. The expression of the decision flag is: ； S4. If the current extreme point is a maximum point, it is necessary to compare it with the spectral amplitude of the previous maximum point, calculate the difference in spectral amplitude between the two and the set threshold to determine their magnitude relationship; then, combined with the decision flag in step S3, decide whether to jump out of the current step and continue traversing the envelope to find extreme points, or continue to judge the magnitude relationship between the amplitude of the current maximum point and the previous minimum point and the threshold ; Set the set of extreme points: ; ； If the current extreme point found is a maximum extreme point , it is necessary to compare with the spectral amplitude of the previous maximum extreme point and calculate the amplitude difference . The expression is as follows: ； Compare the amplitude difference with the set threshold as follows: If , it is considered that the current maximum point may be the signal frequency point, then step S4 is skipped and step S2 is carried out to continue traversing the envelope data to find the extreme points; If , then according to the value of the judgment flag , there are different processing steps: If the judgment flag is true, then proceed to step S5; If the decision flag When, the currently found th maximum point is compared with the previous minimum point in terms of spectral amplitude, and the amplitude difference is expressed as: ; Compare the amplitude difference with the set threshold and if , then jump out of this step and proceed to step S2 to continue traversing the envelope data to find the extreme points; if , then proceed to step S5; S5. For the signal frequency points that meet the threshold decision requirements in step S4 , update them to the set of the start and end positions of the signal frequency band . Finally, until the traversal of the signal spectrum envelope data is completed, output the detection result and end the detection.

2. The frequency-domain detection method for color noise environment signals based on spectral envelope extraction according to claim 1, characterized in that The specific steps of step S1 are as follows: S11. Perform DFT on the obtained discrete sampling signal to convert it from the time domain to the frequency domain and obtain the spectrum of the signal , and the expression is as follows: ; Among them, represents the th frequency point of the spectrum, is the number of spectrum points; S12. For the spectrum Extract the envelope, and set the window size to . The extraction idea is to take the maximum value of the spectrum amplitude for each window, and slide and extract all the spectrum data in sequence. The spectrum envelope data for the th point is calculated as follows: ; Among them, has a value range of .

3. The frequency-domain detection method for a colored noise environment signal based on spectral envelope extraction according to claim 1, wherein The decision formula for the extreme point in step S2 is as follows: ； Among them, represents the maximum point of the envelope, represents the minimum point of the envelope, which belongs to the corresponding set: and .

4. A frequency-domain detection system for color noise environment signals based on spectral envelope extraction, which adopts the frequency-domain detection method for color noise environment signals based on spectral envelope extraction as described in any one of claims 1-3, is characterized in that, It includes a signal acquisition unit, a signal conversion unit, a signal judgment unit and a signal output unit; Signal acquisition unit: Acquire the sampling signal to be processed; Signal conversion unit: Convert the sampled signal from the time domain to the frequency domain to obtain the signal spectrum ; Signal judgment unit: extract the envelope of the spectrum to obtain the envelope data of the spectrum , for the signal spectrum envelope traverse to find the extreme points and update the corresponding maximum value set , minimum value set ; introduce double-threshold joint judgment of the signal frequency band range; Signal output unit: Visually output the result.

5. The frequency-domain detection system for color noise environment signals based on spectral envelope extraction according to claim 4, wherein Signal judgment unit: If the current extreme point is a minimum point, it is necessary to compare it with the spectral amplitude of the previous maximum point, calculate the difference between the two spectral amplitudes and the set threshold to determine the relationship in size, so as to determine the value of the decision flag ; Signal frequency points that meet the threshold judgment requirements , update to the set of start and end positions of the signal frequency band , until the traversal of the signal spectrum envelope data is finally completed; If the current extreme point is a maximum point, it is necessary to compare it with the spectral amplitude of the previous maximum point, calculate the difference in spectral amplitude between the two and set a threshold to determine their magnitude relationship; then combine the decision flag value to decide whether to jump out of the current step and continue traversing the envelope to find extreme points, or to continue judging the magnitude relationship between the current maximum point and the previous minimum point and the threshold value.

6. The frequency domain detection system for chromatic noise environment signals based on spectral envelope extraction according to claim 4, wherein The applicable range is communication signals with modulation methods of BPSK, QPSK, 8PSK, GMSK, 16QAM, and the noise floor fluctuates randomly, and the signal-to-noise ratio is greater than or equal to 3 dB.

Images