Multi-signal detection method based on spectrum utilization rate estimation

Through spectrum utilization estimation and statistical histogram analysis, the signal detection threshold is quickly determined, which solves the problem of degradation in the signal detection performance of the prior art in complex electromagnetic environments, and realizes efficient signal detection in large bandwidth and multi-signal scenarios.

CN119011046APending Publication Date: 2024-11-22NO 30 INST OF CHINA ELECTRONIC TECH GRP CORP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410957218.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-17
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

The existing signal detection methods are difficult to adapt to dynamic changes when facing complex electromagnetic environments and spectrum utilization changes, resulting in a degradation of detection performance, especially in large bandwidth and multi-signal scenarios, which detect false alarms with high effect on weak signals. .

Method used

By smoothing the spectrum data and quantizing the equal spectrum interval, amplitude/power statistical histogram is generated, spectrum utilization is calculated, noise floor/signal detection threshold is determined, search slope inflection point is detected using threshold method, and signal bandwidth is quickly estimated and Frequency realizes detection and determination of multi-signal areas.

Benefits of technology

This method can measure spectrum utilization in real time, dynamically adjust signal detection threshold, improve signal detection performance, reduce false alarms, adapt to large bandwidth and multi-signal scenarios, and improve detection accuracy and robustness.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119011046A_ABST
    Figure CN119011046A_ABST
Patent Text Reader

Abstract

The invention discloses a multi-signal detection method based on frequency spectrum utilization rate estimation, which comprises the following steps of: performing smoothing preprocessing on frequency spectrum data, and generating an amplitude / power statistical histogram from the frequency spectrum data subjected to the smoothing preprocessing by adopting an equal frequency spectrum interval quantization method; carrying out accumulation and summation according to histogram intervals by utilizing the amplitude / power statistical histogram to obtain an accumulation sum curve, detecting and searching slope inflection points by adopting a threshold value method, calculating to obtain a frequency spectrum utilization rate, and further determining a bottom noise / signal detection threshold value; based on the determined ground noise / signal detection threshold value, the multiple signal areas are judged, and a detection signal set is obtained; and carrying out rapid bandwidth and frequency estimation on the detection signal set, and outputting detection results of a plurality of signals. The detection method is simple and efficient, and the accuracy of multi-signal detection is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of signal detection, and in particular to a multi-signal detection method based on spectrum utilization estimation. Background Art

[0002] Signal detection plays an important role in the field of communication and communication countermeasures, mainly to extract useful signals in the background of noise and interference. In engineering applications facing complex electromagnetic environments, after broadband sampling and digitization, the communication signal receiving system first needs to complete the signal detection task under the conditions of multiple signals, different signal-to-noise ratios and different signal bandwidths, and then realize the extraction and separation of multiple signals.

[0003] There are many theories and methods for signal detection. Power spectrum analysis is a commonly used signal detection method based on spectrum analysis. In this type of method, the detection threshold becomes the key to signal detection. The main processing methods for determining the detection threshold include the constant method and the adaptive threshold method. Among them, the constant method relies on expert experience and knowledge, and defines a single detection threshold by observing the noise distribution. It does not have the ability to dynamically adjust the detection threshold value and cannot adapt to the dynamically changing noise environment; the adaptive threshold method includes single threshold and double threshold detection methods, etc. The two methods are similar in principle. The histogram method is used to numerically process the spectrum. Based on the measurement of the statistical noise layer mean, the signal detection threshold is dynamically adjusted. It has the ability to adapt to complex environments, but the logarithmic statistical method uses a single threshold to divide the noise area and the signal area, resulting in a high false alarm detection. The double threshold detection method uses a double threshold judgment based on the logarithmic statistical method, introduces a pending area, and avoids strong interference. The detection effect of weak signals is not strong.

[0004] Among the above signal detection methods, the adaptive threshold method has been used more. Although it has a certain ability to adjust the dynamic adaptive threshold, it is still necessary to multiply the relevant signal and noise coefficients on the basis of the spectrum histogram statistical measurement based on the division of signal area and noise area in the detection method to cope with large bandwidth monitoring, multi-signal detection and other scenarios. The correlation coefficient is an empirical value. When the signal spectrum utilization rate changes, it has a great impact on the detection performance of this type of method, and it still cannot adapt well to the dynamic changes of wide spectrum and multiple signals. Under the conditions of large bandwidth and multi-signal detection, obtaining and utilizing spectrum utilization change information, better determining the detection threshold based on the principle of the adaptive threshold method, ensuring signal detection performance, and realizing detection automation and adaptation as much as possible have become important research ideas for signal detection methods based on spectrum analysis. Summary of the invention

[0005] In view of this, the present invention provides a multi-signal detection method based on spectrum utilization estimation.

[0006] The present invention discloses a multi-signal detection method based on spectrum utilization estimation, which comprises:

[0007] Step 1: Smooth preprocessing of the spectrum data, using the equally divided spectrum interval quantization method, and generating an amplitude / power statistical histogram of the smooth preprocessed spectrum data;

[0008] Step 2: Using the amplitude / power statistical histogram, perform cumulative summation according to the histogram interval to obtain a cumulative sum curve, use the threshold method to detect the search slope inflection point, calculate the spectrum utilization rate, and then determine the noise floor / signal detection threshold;

[0009] Step 3: Based on the determined noise floor / signal detection threshold, multiple signal areas are judged to obtain a detection signal set;

[0010] Step 4: Rapidly estimate the bandwidth and frequency of the detection signal set and output the detection results of multiple signals.

[0011] Furthermore, the step 1 comprises:

[0012] Assume that the spectrum data of the input sampled broadband signal is S(k), where k is the frequency spectrum harmonic order, k=0, 1, ..., N-1, and N is the number of Fourier transform points;

[0013] The moving average method is used to implement mean smoothing filtering on S(k) to obtain the smoothed spectrum data S sm (j):

[0014]

[0015] Where L0 = (L-1) / 2, L is the span of the moving average, S(k) and S sm (j) The number of points is the same, j is the number of smoothing filters.

[0016] Furthermore, the step 1 comprises:

[0017] Step 11: quantize the spectrum in equal parts and measure the maximum and minimum values ​​of the spectrum data;

[0018] Step 12: Calculate the point set between each equal partition and count the frequency points in the interval. All frequency points in the spectrum correspond to their corresponding intervals, and the frequency-power distribution is counted, which is the statistical histogram.

[0019] Furthermore, the step 11 comprises:

[0020] S max =max(S sm ) (2)

[0021] S min =min(Ssm ) (3)

[0022] The effective interval is S max -S min , the interval is divided into D equal parts, then the quantization height of each equal part is:

[0023] h=(S max -S min ) / D (4)

[0024] Where h is the quantized height of each equal part.

[0025] Further, the step 12 comprises:

[0026] Calculate the point set between each equal partition, and count the frequency points of the interval, and arrange them from small to large intervals. Each interval is represented by (S min +(i-1)*h,S m i n +i*h], where 0<i≤D;

[0027] Frequency points in each interval H i The initial value is 0. The spectrum is searched. If there is a frequency point spectrum value belonging to interval i, the number of points in the interval is increased by 1 until the entire spectrum is traversed. At this time, all frequency points in the spectrum correspond to their corresponding intervals. The intervals are arranged from small to large, and the frequency-power distribution is statistically calculated, which is a statistical histogram.

[0028] Furthermore, the step 2 comprises:

[0029] Step 21: Perform cumulative summation on the statistical histogram to obtain the cumulative sum sequence M i :

[0030]

[0031] Coordinate unification processing is performed to make the ratio of the vertical and horizontal axis lengths of the statistical histogram cumulative sum curve equal to 1, and the processed cumulative sum sequence M′ is obtained. i :

[0032] M′ i =D / N*M i (6)

[0033] Step 22: Process the accumulated sum sequence M′ i Calculate the slope and convert it into phase angle θ after equidistant differentiation n :

[0034] Let the differential distance be K, and K is a positive integer, then:

[0035] θ n =tan -1 ((M′n+k -M′ n ) / K) (7)

[0036] Among them, M′ n+k is the cumulative sum sequence after the n+kth processing, M′ n is the cumulative sum sequence after the nth processing, n is the number of differences, n = 0, 1, ..., NK;

[0037] Step 23: Based on the phase angle θ n , calculate the spectrum utilization F u .

[0038] Further, the step 23 comprises:

[0039] First, determine the bottom noise layer inflection point decision value θ th According to the same process in step 21, the inflection point appears, indicating that the cumulative sum increment becomes smaller, θ n The value becomes smaller, and θ is increased from small to large according to n. n Value judgment, comparison θ n and θ th , when θ appears for the first time n <θ th , the corresponding n value is recorded as I n , which is the inflection point of the background noise;

[0040] Thus, the spectrum utilization factor F is obtained u :

[0041]

[0042] At the same time, calculate the noise floor / signal detection threshold P n :

[0043] P n =(1-F u )*h+S min (9)

[0044] P n It can be used directly as the signal detection threshold, or it can be multiplied by a fixed detection coefficient and used as the signal detection threshold.

[0045] Furthermore, the step 3 comprises:

[0046] A single threshold or double threshold judgment method is used to divide the noise area and signal area by the background noise / signal detection threshold, and the smoothed spectrum data S sm Perform traversal search, and for non-overlapping multiple signals, assign the continuous frequency points divided into the signal area to the same detection signal set G m :

[0047] G m={S sm (n)|S sm (n)>P n} (10)

[0048] Wherein, m is the serial number of the detected signal set, m=1, ..., M, and M is the total number of detected signal sets.

[0049] Furthermore, the step 4 comprises:

[0050] Step 41: Obtain detection signal set G m The maximum value S of the spectrum data sm (n0), according to the detection bandwidth, take the bandwidth detection threshold P w :

[0051] P w =S sm (n0)-Q

[0052] Among them, Q is the detection bandwidth, n0 is the maximum position of the spectrum data;

[0053] Search from position to n0 spectrum around to obtain less than P w The signal bandwidth can be estimated by using the spectrum data positions n1 and n2:

[0054] W s =(n2-n1)*f s / N (11)

[0055] Among them, f s is the sampling frequency;

[0056] Step 42: Estimate the signal carrier frequency f using the centroid method c :

[0057]

[0058] Step 43: Repeat steps 41 and 42 for all detection signal sets G m Estimate bandwidth and carrier frequency to output detection results for multiple signals.

[0059] Due to the adoption of the above technical solution, the present invention has the following advantages:

[0060] 1. Compared with the traditional multi-signal detection method based on spectrum analysis, in terms of signal detection adaptive threshold acquisition, the present invention introduces a new method for spectrum utilization estimation. On the basis of generating a spectrum statistical histogram, by using the relationship and law between the histogram accumulation and the curve and the spectrum utilization, the inflection point of the curve is quickly found to calculate the spectrum utilization, and then the noise floor detection line and the signal detection threshold are determined. The signal detection method is simple, efficient, and versatile. It does not require pre-setting the signal spectrum utilization conditions, and can measure and estimate the current spectrum utilization in real time. It solves the problem that many current signal detection methods lead to a decrease in detection performance when facing actual spectrum utilization changes, and can be used in combination with other signal detection methods, which is beneficial to improving signal detection performance.

[0061] 2. The present invention also provides implementation steps such as multi-signal area detection and determination, as well as rapid parameter estimation, which are practical, fast, and easy to implement in engineering. The method can adapt to the task scenarios of large bandwidth monitoring and multi-signal detection, and automatically completes the signal detection task through real-time measurement and estimation, dynamically adjusting the signal detection threshold, and embodies good signal detection performance and robustness, and can be used as a choice of signal detection method. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0063] Figure 1 A schematic flow chart of a multi-signal carrier detection method under broadband receiving conditions according to an embodiment of the present invention;

[0064] Figure 2 Schematic diagram of comparison of signal spectrum before and after preprocessing according to an embodiment of the present invention;

[0065] Figure 3 A schematic diagram of a spectrum utilization rate estimation process according to an embodiment of the present invention;

[0066] Figure 4 Schematic diagram of statistical histogram accumulation and curve inflection point detection according to an embodiment of the present invention. DETAILED DESCRIPTION

[0067] The present invention is further described in conjunction with the accompanying drawings and embodiments, and the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field should fall within the scope of protection of the embodiments of the present invention.

[0068] The present invention mainly provides a multi-signal carrier detection method under broadband reception conditions. In terms of the division of noise areas and signal areas, it introduces statistical measurement of spectrum utilization, which can automatically determine the bottom noise area, solves the problem that other signal detection methods of this type cannot adapt well to the electromagnetic background with large detection bandwidth and dynamic changes of multiple signals, improves the detection accuracy of multiple signals, weak signals, etc., and quickly provides information such as the number of signals, carriers, bandwidth, etc.

[0069] See also Figure 1 The present invention provides an embodiment of a multi-signal detection method based on spectrum utilization estimation, which includes: spectrum smoothing preprocessing, spectrum graph digitization, spectrum utilization estimation, multi-signal area detection and judgment, and signal parameter rapid estimation.

[0070] (1) Spectrum smoothing preprocessing:

[0071] The digital signal sampled by the intermediate frequency broadband is transformed by Fourier transform to obtain the spectrum data. At this time, the spectrum data has certain noise and interference signals, and there are peaks and burrs on the spectrum, which affect the subsequent signal detection. The mean filter is used to pre-process the spectrum data to obtain smoothed spectrum data for subsequent processing. The comparison of the spectrum data before and after smoothing is shown as follows: Figure 2 As shown in (a) and (b), the smoothed spectrum data has less fluctuation and effectively removes background glitches and spike signals.

[0072] (2) Spectrum graphics digitization:

[0073] The digitization of spectrum graphics adopts the statistical histogram method which is more commonly used in image graphics processing. The statistical histogram refers to finding the grayscale range and calculating the grayscale difference after a given image histogram, dividing the groups according to the extreme difference and calculating the mean of each group, replacing the grayscale data of each group, and the resulting histogram is the statistical histogram. The spectrum data is processed to generate an amplitude / power statistical histogram, which, combined with the aforementioned smoothing filtering effect, can adapt well to changes in the spectrum. When the background noise changes, the amplitude / power statistical histogram weakens the impact of external noise changes on the statistics of the spectrum data, and can adaptively perform the statistical effect of the energy in the noise area, laying the foundation for subsequent signal detection.

[0074] (3) Spectrum utilization estimation:

[0075] This link is a key step for the present invention to improve the accuracy of signal detection. Other spectrum analysis methods based on statistical histograms often perform noise floor statistics through statistical histograms, and multiply this by experience to obtain the noise coefficient as the signal detection threshold. When the background electromagnetic signal changes significantly, the noise floor statistics are affected by the spectrum utilization, resulting in inaccurate signal detection thresholds, which ultimately leads to a decrease in detection performance. Starting from the problem, the present invention introduces statistics on spectrum utilization when determining the signal detection threshold to adapt to the impact of dynamic changes in background electromagnetic signals and enhance the stability of detection.

[0076] The spectrum utilization estimation process is as follows: Figure 3 As shown in the figure, the amplitude / power statistical histogram is used to accumulate and sum according to the histogram interval to obtain the cumulative sum curve. The curve is subjected to two-dimensional coordinate homogeneity, and the gradient / slope is calculated segmentally. The threshold method is used to detect the inflection point of the search slope. The detection diagram is shown in the figure. Figure 4 As shown, the spectrum utilization is calculated, and then the background noise detection line and the signal detection threshold are determined.

[0077] (4) Multi-signal area detection and judgment:

[0078] The signal detection threshold obtained by spectrum utilization is based on the actual measured noise floor inflection point, which can be used as a single threshold signal detection line to detect the signal spectrum. The mean amplitude point above the signal detection line is the signal frequency point, corresponding to the signal area; the mean spectrum frequency point below the signal detection line is the noise frequency point, corresponding to the noise area. When there are multiple signals in the detection frequency band, multiple signals can be detected at the same time. In order to take into account weak signal detection and avoid strong noise interference, a dual threshold detection method can also be used to improve performance.

[0079] (5) Rapid estimation of signal parameters:

[0080] The bandwidth and carrier frequency are estimated one by one for the multiple signal areas obtained by the above S detection. The bandwidth is estimated by dropping 3dB or 6dB from the highest peak of the current signal area, which is suitable for spectral symmetry and suppressed carrier signals; the carrier frequency is estimated from the 3dB or 6dB signal area obtained by bandwidth estimation, and the centroid method is used for rapid carrier frequency estimation. Finally, when the signal detection is completed, the signal quantity, bandwidth, carrier frequency and other information are output at the same time.

[0081] For ease of understanding, the present invention provides a more specific embodiment:

[0082] A multi-signal carrier detection method based on spectrum utilization estimation comprises the following steps:

[0083] S1: Assume that the spectrum data of the input sampled broadband signal is S(k), where k is the frequency spectrum harmonic order, k=0, 1, ..., N-1, and N is the number of Fourier transform points.

[0084] The moving average method is used to implement mean smoothing filtering on S(k) to obtain the smoothed spectrum data S sm (j).

[0085]

[0086] Where L0 = (L-1) / 2, L is the span of the moving average, which is usually an odd number. To ensure the corresponding relationship between the spectrum data of subsequent detection, S(k) and S sm (j) The points are the same, for S sm (j) The endpoint performs the following processing. For example, when L=5, the initial data is:

[0087] S sm (0) = S(0)

[0088] S sm (1) = (S(0) + S(1) + S(2)) / 3

[0089] S sm (2)=(S(0)+S(1)+S(2)+S(3)+S(4)) / 5

[0090] …

[0091] Similarly, the ending data is processed in the same way.

[0092] S2: Use the equally divided spectrum interval quantization method to generate the amplitude / power statistical histogram, and calculate S as follows:

[0093] S2.1: Quantify the effective interval of the spectrum equally and measure the maximum and minimum values ​​of the spectrum data:

[0094] S max =max(S sm ) (2)

[0095] S min =min(S sm ) (3)

[0096] The effective interval is S max -S min , the interval is divided into D equal parts, then the quantization height h of each equal part is:

[0097] h=(S max -S min ) / D (4)

[0098] S2.2: Calculate the point set between each equal partition and count the frequency points of the interval. Arrange the intervals from small to large, and each interval can be expressed as (S min +(i-1)*h,Smin +i*h], where 0 <i≤D。

[0099] Frequency points in each interval H i The initial value of is 0, and the spectrum is searched. If there is a frequency point spectrum value belonging to interval i, the number of points in the interval is increased by 1 until the entire spectrum is traversed. At this time, all frequency points in the spectrum correspond to their corresponding intervals, and the intervals are arranged from small to large, and the frequency-power distribution is statistically calculated, which is the statistical histogram data, in order to facilitate the calculation and determination of the subsequent signal detection threshold.

[0100] S3: Calculate the spectrum utilization rate based on the characteristics of the statistical histogram accumulation curve. Calculate S as follows:

[0101] S3.1: Accumulate and sum the statistical histogram to obtain the cumulative sum sequence M i .

[0102]

[0103] Perform coordinate homogeneity processing to make the ratio of the vertical and horizontal axis lengths of the statistical histogram cumulative sum curve equal to 1, and obtain the processed cumulative sum sequence M′ i .

[0104] M′ i =D / N*M i (6)

[0105] S3.2: After processing, the cumulative sum sequence M′ i Calculate the slope and convert it into phase angle θ after equidistant differentiation n .

[0106] Let the differential distance be K, where K can be a positive integer in the interval [D / 10, D / 20]. Then:

[0107] θ n =tan -1 ((M′ n+k -M′ n ) / K) (7)

[0108] Among them, M′ n+k is the cumulative sum sequence after the n+kth processing, M′ n is the cumulative sum sequence after the nth processing, n is the number of differences, n = 0, 1, ..., NK;

[0109] S3.3: Calculate the spectrum utilization factor F u First, determine the bottom noise layer inflection point decision value θ th According to the unified processing of S3.1, the inflection point indicates that the accumulation and increment become smaller, θ n The value becomes smaller. Here, we usually take θ th<π / 4.

[0110] According to n from small to large θ n Value judgment, comparison θ n and θ th , when θ appears for the first time n <θ th , the corresponding n value is recorded as I n , which is the inflection point of the background noise.

[0111] Thus, the spectrum utilization rate is obtained

[0112]

[0113] At the same time, calculate the noise floor / signal detection threshold

[0114] P n =(1-F u )*h+S min (9)

[0115] Here, P n It can be used directly as the subsequent signal detection threshold, or it can be multiplied by a fixed detection coefficient, for example, adding 3dB as the signal detection threshold, that is, P n =2P n .

[0116] S4: Determine multiple signal areas. This S is similar to the traditional spectrum analysis method and can adopt a single threshold or double threshold determination method.

[0117] Taking the single threshold decision method as an example, the signal detection threshold P obtained by S3 n As the dividing line between the noise and signal areas, the frequency points above the dividing line constitute the signal area, and the frequency points below the noise line constitute the noise area.

[0118] For spectrum data S sm Perform traversal search. For non-overlapping multiple signals, the continuous frequency points exceeding the dividing line belong to the same detection signal set G. m ,

[0119] G m ={S sm (n)|S sm (n)>P n} (10)

[0120] Here, m is the number of the detected signal set, m=1, ..., M, and M is the total number of detected signal sets.

[0121] S5: The detected signal set G m Perform a quick estimate of bandwidth and frequency. Process S as follows:

[0122] S5.1: Get G m The maximum value S of the spectrum data in the set sm (n0), according to the bandwidth detection condition, take the bandwidth detection threshold P w , for example, the detection bandwidth is 3dB:

[0123] P w =S sm (n0)-3

[0124] Here, Q is the detection bandwidth, and n0 is the maximum position of the spectrum data;

[0125] Search from the n0 position to the left and right of the spectrum horizontal axis to obtain w The signal bandwidth can be estimated by using the spectrum data positions n1 and n2:

[0126] W s =(n2-n1)*f s / N (11)

[0127] Among them, f s is the sampling frequency.

[0128] S5.2: Estimation of signal carrier frequency f using the centroid method c .

[0129]

[0130] S5.3: Repeat S5.1 and S5.2 for all detection signal sets G m Estimate bandwidth and carrier frequency to output detection results for multiple signals.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A multi-signal detection method based on spectrum utilization estimation, characterized in that: include: Step 1: Smooth preprocessing of the spectrum data, using the equally divided spectrum interval quantization method, and generating an amplitude / power statistical histogram of the smooth preprocessed spectrum data; Step 2: Using the amplitude / power statistical histogram, perform cumulative summation according to the histogram interval to obtain a cumulative sum curve, use the threshold method to detect the search slope inflection point, calculate the spectrum utilization rate, and then determine the noise floor / signal detection threshold; Step 3: Based on the determined noise floor / signal detection threshold, multiple signal areas are judged to obtain a detection signal set; Step 4: Rapidly estimate the bandwidth and frequency of the detection signal set and output the detection results of multiple signals.

2. The method according to claim 1, characterized in that The step 1 comprises: Assume that the spectrum data of the input sampled broadband signal is S(k), where k is the frequency spectrum harmonic order, k=0, 1, ..., N-1, and N is the number of Fourier transform points; The moving average method is used to implement mean smoothing filtering on S(k) to obtain the smoothed spectrum data S sm (j): Where L0 = (L-1) / 2, L is the span of the moving average, S(k) and S sm (j) The number of points is the same, j is the number of smoothing filters.

3. The method according to claim 2, characterized in that The step 1 comprises: Step 11: quantize the spectrum in equal parts and measure the maximum and minimum values ​​of the spectrum data; Step 12: Calculate the point set between each equal partition and count the frequency points in the interval. All frequency points in the spectrum correspond to their corresponding intervals, and the frequency-power distribution is counted, which is the statistical histogram.

4. The method according to claim 3, characterized in that The step 11 comprises: S max =max(S sm ) (2) S min =min(S sm ) (3) The effective interval is S max -S min , the interval is divided into D equal parts, and the quantization height of each equal part is: h=(S max -S min ) / D (4) Where h is the quantized height of each equal part.

5. The method according to claim 4, characterized in that The step 12 comprises: Calculate the point set between each equal partition, and count the frequency points of the interval, and arrange them from small to large intervals. Each interval is represented by (S min +(i-1)*h,S min +i*h], where 0<i≤D; Frequency points in each interval H i The initial value is 0. The spectrum is searched. If there is a frequency point spectrum value belonging to interval i, the number of points in the interval is increased by 1 until the entire spectrum is traversed. At this time, all frequency points in the spectrum correspond to their corresponding intervals. The intervals are arranged from small to large, and the frequency-power distribution is statistically calculated, which is a statistical histogram.

6. The method according to claim 1, characterized in that The step 2 comprises: Step 21: Perform cumulative summation on the statistical histogram to obtain the cumulative sum sequence M i : Coordinate unification processing is performed to make the ratio of the vertical and horizontal axis lengths of the statistical histogram cumulative sum curve equal to 1, and the processed cumulative sum sequence M′ is obtained. i : M′ i =D / N*M i (6) Step 22: Process the accumulated sum sequence M′ i Calculate the slope and convert it into phase angle θ after equidistant differentiation n : Let the differential distance be K, and K is a positive integer, then: θ n =then -1 ((M′ n+k -M′ n ) / K) (7) Among them, M′ n+k is the cumulative sum sequence after the n+kth processing, M′ n is the cumulative sum sequence after the nth processing, n is the number of differences, n = 0, 1, ..., NK; Step 23: Based on the phase angle θ n , calculate the spectrum utilization F u .

7. The method according to claim 6, characterized in that The step 23 comprises: First, determine the bottom noise layer inflection point decision value θ th According to the same process in step 21, the inflection point appears, indicating that the cumulative sum increment becomes smaller, θ n The value becomes smaller, and θ is increased from small to large according to n. n Value judgment, comparison θ n and θ th , when θ appears for the first time n <θ th , the corresponding n value is recorded as I n , which is the inflection point of the background noise; Thus, the spectrum utilization factor F is obtained u : At the same time, calculate the noise floor / signal detection threshold P n : P n =(1-F u )*h+S min (9) P n It can be used directly as the signal detection threshold, or it can be multiplied by a fixed detection coefficient and used as the signal detection threshold.

8. The method according to claim 1, characterized in that The step 3 comprises: A single threshold or double threshold judgment method is used to divide the noise area and signal area by the background noise / signal detection threshold, and the smoothed spectrum data S sm Perform traversal search, and for non-overlapping multiple signals, assign the continuous frequency points divided into the signal area to the same detection signal set G m : G m ={S sm (n)|S sm (n)>P n } (10) Wherein, m is the serial number of the detected signal set, m=1, ..., M, and M is the total number of detected signal sets.

9. The method according to claim 1, characterized in that: The step 4 comprises: Step 41: Obtain detection signal set G m The maximum value S of the spectrum data sm (n0), according to the detection bandwidth, take the bandwidth detection threshold P w : P w =S sm (n0)-Q Among them, Q is the detection bandwidth, n0 is the maximum position of the spectrum data; Search from position to n0 spectrum around to obtain less than P w The signal bandwidth can be estimated by using the spectrum data positions n1 and n2: W s =(n2-n1)*f s / N (11) Among them, f s is the sampling frequency; Step 42: Estimate the signal carrier frequency f using the centroid method c : Step 43: Repeat steps 41 and 42 for all detection signal sets G m Estimate bandwidth and carrier frequency to output detection results for multiple signals.