Communication spectrum signal detection method and system based on noise floor adaptive extraction

Abstract

The invention discloses a communication spectrum signal detection method and system based on noise floor adaptive extraction. The method comprises the following steps: obtaining a detected communication time domain signal and performing Fourier transform to obtain communication signal spectrum data; counting the frequency spectrum points of the communication signal frequency spectrum data in each frequency spectrum power value interval to obtain power value-frequency spectrum point distribution data; determining a power demarcation point value between the spectrum noise and the communication signal spectrum according to the power value-spectrum point distribution data; extracting spectrum noise floor data from the communication signal spectrum data according to the power demarcation point value; and fitting the extracted spectrum noise floor data to form a noise floor fitting curve, and detecting a final communication spectrum signal from the original communication signal by taking the noise floor fitting curve as an adaptive detection threshold. The method has the advantages of being simple in implementation method, low in complexity, high in detection precision, high in robustness and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of communication signal detection, and particularly relates to a communication spectrum signal detection method and system based on noise floor adaptive extraction. Background Art

[0002] Due to the unpredictability of the distribution of ground communication, interference signals, and other noises, during the detection and estimation of spectrum signal parameters, since the noise floor in the spectrum to be detected often fluctuates greatly, it will increase the difficulty of detecting and estimating spectrum signal parameters. At the same time, since the noise floor of frequency-using devices also changes dynamically over time, resulting in dynamic fluctuations in the noise floor of monitoring data, it will also increase the difficulty of subsequent processing of the monitored spectrum data.

[0003] The most direct method for detecting signals on the spectrum is to use a straight line on the spectrum as the signal detection threshold. If the spectrum data is higher than the threshold, the signal is considered to exist. However, this method is only applicable to the case where the spectrum noise floor is flat. In a complex and changeable electromagnetic environment, using a fixed straight line as the signal detection threshold for the entire spectrum will not be applicable, and it is difficult to effectively detect signals on the spectrum. Take Figure 2 the situation shown in (a) below as an example. The noise floor of the signal spectrum is not flat, and there is an obvious bulge near 900 MHz. If the signal on the entire spectrum is detected according to a fixed straight line at this time, it is difficult to avoid the influence caused by the uneven noise floor.

[0004] To solve the problem of signal detection in the case of an uneven noise floor, the prior art usually adopts the method of segmenting the spectrum to detect signals to eliminate the influence of the uneven noise floor. However, in the actual environment, the electromagnetic environment is complex and changeable. The method of spectrum segmentation requires relying on human-computer interaction software to set the analyzed frequency band. The degree of signal detection automation is not high, and the implementation operation is complex. When the electromagnetic environment changes, the detection parameters need to be reset. Summary of the Invention

[0005] The technical problem to be solved by the present invention: Aiming at the above problems of the prior art, a communication spectrum signal detection method and system based on noise floor adaptive extraction with a simple implementation method, low complexity, high detection accuracy, and strong robustness are provided.

[0006] To solve the above technical problems, the technical solution adopted by the present invention is as follows: A communication spectrum signal detection method based on noise floor adaptive extraction, comprising: Obtaining the measured communication signal and performing Fourier transform to obtain communication signal spectrum data; Counting the number of spectrum points in each spectrum power value interval of the communication signal spectrum data to obtain power value-spectrum point distribution data of the change of the number of spectrum points with the power value; Determine the power demarcation point value between the spectral noise and the communication signal spectrum according to the power value - spectral point number distribution data; Extract the spectral noise floor data from the communication signal spectrum data according to the power demarcation point value; Fit the extracted spectral noise floor data to form a noise floor fitting curve, and use the noise floor fitting curve as an adaptive detection threshold to detect the final communication spectrum signal from the original communication signal.

[0007] Further, the determining the power demarcation point value between the spectral noise and the communication signal spectrum according to the power value - spectral point number distribution data includes: Perform cumulative summation on the power value - spectral point number distribution data to obtain cumulative summation distribution data of the cumulative sum of spectral point numbers changing with the spectral power value; Judge the change trend of the cumulative summation distribution data, find the first power value corresponding to the time when the growth trend starts to weaken, and use the found first power value as the power demarcation point value between the spectral noise and the communication signal spectrum.

[0008] Further, the determining the power demarcation point value between the spectral noise and the communication signal spectrum according to the power value - spectral point number distribution state includes: Find the maximum value of the spectral point number statistical data and the corresponding second power value from the power value - spectral point number distribution data; Traverse in the direction greater than the second power value until the spectral point number corresponding to the found third power value is less than a specified proportion of the maximum statistical value, and use the found third power value as the power demarcation point value between the spectral noise and the communication signal spectrum.

[0009] Further, when traversing in the direction greater than the second power value, if no third power value meeting the preset conditions is found after the traversal is completed, use the last power value in the power value - spectral point number distribution data as the power demarcation point value between the spectral noise and the communication signal spectrum.

[0010] Further, after statistically obtaining the power value - spectral point number distribution data of the spectral point numbers of the communication signal spectrum data in each spectral power value interval, it further includes performing mean filtering on the communication signal spectrum data.

[0011] Further, the obtaining the communication signal spectrum data by acquiring the measured communication time - domain data and performing Fourier transform includes: According to the number of sampling points of the time-domain data of the communication signal, perform a fast Fourier transform on the measured communication signal with a step of a preset number of points to obtain short-time Fourier transform time-frequency data, and then perform median filtering on multiple points before and after the short-time Fourier transform time-frequency data in the time dimension to obtain the filtered communication signal spectrum data.

[0012] Further, extracting spectrum noise floor data from the communication signal spectrum data according to the power demarcation point value includes: Remove the spectrum data corresponding to the power value greater than the demarcation point value in the communication signal spectrum data, and use the noise spectrum power value of the point adjacent to the frequency edge of the communication signal to fill the spectrum data corresponding to the power value greater than the demarcation point value to obtain the noise floor data.

[0013] Further, the steps of fitting the extracted spectrum noise floor data to form a noise floor fitting curve include: Perform exponential weighted filtering and smoothing on the spectrum noise floor data to obtain the first smoothed noise floor data; Flip the spectrum noise floor data left and right, then perform exponential weighted filtering and then flip it left and right again to obtain the second smoothed noise floor data; Sum and average the second smoothed noise floor data and the first smoothed noise floor data to obtain the final noise floor fitting curve.

[0014] The present invention also provides a communication spectrum signal detection system based on noise floor adaptive extraction, including a microprocessor and a memory connected to each other, and the microprocessor is programmed or configured to execute the above-mentioned communication spectrum signal detection method based on noise floor adaptive extraction.

[0015] The present invention also provides a computer-readable storage medium, in which a computer program is stored, and the computer program is used to be programmed or configured by a microprocessor to execute the above-mentioned communication spectrum signal detection method based on noise floor adaptive extraction.

[0016] Compared with the prior art, the advantages of the present invention are: The present invention obtains the power value - spectrum point number distribution data by statistically counting the spectrum point numbers of communication signal spectrum data in different spectrum power value intervals, determines the power demarcation point value between the spectrum noise and the communication signal spectrum based on the power value - spectrum point number distribution data, and then extracts the spectrum noise floor data by using the power demarcation point value. The spectrum noise floor data can be effectively extracted. Furthermore, by fitting the extracted spectrum noise floor data to form a noise floor fitting curve and using the noise floor fitting curve as an adaptive detection threshold, the detection thresholds of different communication signals can be adaptively determined according to the spectrum noise floor data, so as to accurately detect the communication spectrum signal from the original communication signal. Compared with the traditional method of using a fixed straight line as the detection threshold, the accuracy of spectrum signal detection can be greatly improved, the influence on signal detection caused by the unevenness of the spectrum noise floor data can be avoided, and the implementation complexity and cost can also be reduced. It can be applied to the detection of communication signals in various complex communication scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 FIG. is a schematic flow chart of a communication spectrum signal detection method based on noise floor adaptive extraction provided by an embodiment of the present invention.

[0018] Figure 2 FIG. is a spectrum diagram of communication signal spectrum data before and after filtering, where (a) is the unfiltered spectrum diagram of communication signal spectrum data before filtering, and (b) is the spectrum diagram of communication signal spectrum data after filtering.

[0019] Figure 3 FIG. is a statistical distribution data diagram of the spectrum point number statistical data value distribution with respect to the power value in this embodiment.

[0020] Figure 4 In this embodiment, for Figure 3 FIG. is the first statistical distribution data diagram obtained by filtering the statistical distribution data in FIG.

[0021] Figure 5 In this embodiment, for Figure 4 FIG. is the second statistical distribution data diagram of the cumulative sum of spectrum point numbers with respect to the power value obtained by performing cumulative summation on the first statistical distribution data diagram in FIG.

[0022] Figure 6 FIG. is a schematic diagram for determining the noise - signal power demarcation point in this embodiment.

[0023] Figure 7 FIG. is a noise floor data diagram extracted according to the determined demarcation point in this embodiment.

[0024] Figure 8 FIG. is a schematic flow chart for smoothing and fitting the noise floor data in this embodiment.

[0025] Figure 9This is the curve graph of the fitted smooth noise floor after smooth fitting in this embodiment.

[0026] Figure 10 This is the curve graph of the adaptive signal detection threshold in this embodiment.

[0027] Figure 11 This is the curve graph of the signal detection threshold in the actual electromagnetic environment obtained by applying the communication spectrum signal detection method based on noise floor adaptive extraction of this embodiment. Detailed implementation manners

[0028] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings of the specification and specific implementation manners.

[0029] Use the signal threshold determination method based on statistics to determine the threshold for rough processing of signal noise, then fit and complete the noise floor, and finally subtract the fitted noise floor signal from the original signal to correct the original signal. This type of method can better solve the influence of uneven noise floor on signal detection, but it needs to rely on accurately extracting the noise floor signal. For the extraction of the spectrum noise floor, there are the following several methods for extracting the spectrum noise floor of communication signals currently: 1. Adopt the method of threshold sorting. By statistically analyzing the spectrum data, sorting the spectrum data and then multiplying by a proportionality coefficient as the threshold of the noise floor, and sorting out the noise floor data according to the threshold. However, this type of method is only applicable to the case where there are signals in the spectrum. When there are no signals in the spectrum and all are noise floors, the finally obtained signal detection threshold will detect a large amount of noise as signals.

[0030] 2. Adopt the method of using high-order Fourier series to fit the spectrum noise floor data. Since the Fourier series fitting algorithm is a method of decomposing the representation of a periodic function into a trigonometric function series, it has certain limitations and is only applicable to fitting periodic cyclic signals, and the fitting effect for non-periodic data is not good. The spectrum noise floor data is obviously not a type of data that appears periodically in a cycle. For a specific section of spectrum noise floor data, although a relatively ideal result can be achieved by debugging different Fourier fitting series, the randomness of the change of spectrum noise floor data determines that fitting the noise floor by Fourier series does not have universality. In addition, the convergence speed of Fourier series fitting is very slow, and a large number of series terms are required to achieve sufficient accuracy.

[0031] 3. Adopt the method of using high-order polynomials to fit the spectrum noise floor data. Since the high-order polynomial fitting algorithm is difficult to establish an effective model for non-linear data, the calculation is very time-consuming when the data volume is large, and it is difficult to express complex random data. And the selection of the fitting order is also very crucial. For data with different spectrum noise floor characteristics, the order of polynomial fitting needs to be continuously adjusted, and it is not very adaptable to continuously changing spectrum data.

[0032] Considering that the noise floor accounts for the vast majority of the distribution across the entire spectrum. For example, in a wideband spectrum, noise occupies the vast majority of frequency bands. The present invention obtains the power value - spectrum point number distribution data of the spectrum point number of the communication signal spectrum data varying with the power value in each spectrum power value interval, performs statistical analysis on the power value to determine the power demarcation point value between the spectrum noise and the communication signal spectrum, then extracts the spectrum noise floor data from the communication signal spectrum data through the power demarcation point value, and finally fits the extracted spectrum noise floor data to form a noise floor fitting curve. Using this noise floor fitting curve as an adaptive detection threshold to achieve signal detection, it is possible to adaptively determine the detection thresholds of different communication signals according to the spectrum noise floor data, thereby accurately detecting the communication spectrum signal from the original communication signal. Compared with the traditional method of using a fixed straight line as the detection threshold, it can greatly improve the accuracy of spectrum signal detection, avoid the influence on signal detection caused by the unevenness of the spectrum noise floor data, and can also reduce the implementation complexity and cost, and is applicable to the detection of communication signals in various complex communication scenarios.

[0033] As Figure 1 shown, the steps of the communication spectrum signal detection method based on noise floor adaptive extraction according to the embodiments of the present invention include: Step S1: Obtain the measured communication time-domain signal and perform Fourier transform to obtain the communication signal spectrum data; Step S2: Count the spectrum point numbers of the communication signal spectrum data in each spectrum power value interval to obtain the power value - spectrum point number distribution data of the spectrum point number varying with the power value; Step S3: Determine the power demarcation point value between the spectrum noise and the communication signal spectrum according to the power value - spectrum point number distribution data; Step S4: Extract the spectrum noise floor data from the communication signal spectrum data according to the power demarcation point value; Step S5: Fit the extracted spectrum noise floor data to form a noise floor fitting curve, and use the noise floor fitting curve as an adaptive detection threshold to detect the final communication spectrum signal from the original communication signal.

[0034] It can be understood that in this embodiment, by utilizing the characteristic that the noise floor power has an obvious aggregation phenomenon, and obtaining the power value-spectrum point number distribution data of the number of spectrum points in each spectrum power value interval of the communication signal spectrum data changing with the power value, the power demarcation point value between the spectrum noise and the communication signal spectrum can be effectively and accurately determined according to the power value-spectrum point number distribution data, improving the extraction accuracy of the subsequent spectrum noise floor data; by extracting the spectrum noise floor data from the communication signal spectrum data through the power demarcation point value, different noise floor detection thresholds for different communication signals can be adaptively determined according to the extracted spectrum noise floor data, making full use of the fluctuating state of the noise floor data to adaptively change the detection threshold, so as to effectively detect the spectrum signal from the original communication signal and avoid the influence on signal detection caused by the uneven spectrum noise floor data. Compared with the existing traditional method of directly using a fixed straight line to extract the spectrum noise floor data, the extraction accuracy of the spectrum noise floor data can be improved; compared with the traditional method of using spectrum segment detection signals to extract the spectrum noise floor data, the implementation complexity and cost can also be reduced; compared with the traditional method of using the difference between the maximum value and the minimum value and then multiplying by a proportionality coefficient as the noise floor extraction threshold method, the problem of detecting noise signals when there is no signal in the spectrum can be avoided.

[0035] In this embodiment, step S1 includes: according to the number of sampling points of the communication signal time-domain data, performing a fast Fourier transform on the measured communication signal with a preset step size of points to obtain short-time Fourier transform time-frequency data, and then performing median filtering on multiple points before and after in the time dimension of the short-time Fourier transform time-frequency data to obtain the filtered communication signal spectrum data.

[0036] In a specific application embodiment, a fast Fourier transform with a step of 2048 points and 16384 points can be performed on the collected communication signal time-domain data to obtain the short-time Fourier transform time-frequency data as shown in Figure 2 (a) below, and then median filtering is performed on 64 points before and after in the time dimension of the time-frequency data to obtain the filtered spectrum data as shown in Figure 2 (b) below.

[0037] In a specific application embodiment, the number of spectrum points in each power value interval of the filtered spectrum data is counted to obtain the statistical data value of the number of spectrum points in each power interval as shown in Figure 3 below, where the abscissa is the power value of the spectrum data and the ordinate is the statistical result (statistical data value) of the number of spectrum points. It can be seen from Figure 3 that the noise floor power obviously has an aggregation phenomenon, so the noise floor data can be extracted using this characteristic.

[0038] For the convenience of subsequent step calculations, after step S2 in this embodiment, mean filtering is also included for the communication signal spectrum data. For example, Figure 3The data is subjected to mean filtering of 32 points to obtain the result as shown in Figure 4 so as to facilitate more accurately determining the power demarcation point between the noise power and the signal in the subsequent process.

[0039] In this embodiment, step S3 includes: Step S31: Cumulatively sum the power value - spectrum point number distribution data to obtain cumulative sum distribution data of the cumulative sum of spectrum point numbers changing with the spectrum power value; Step S32: Judge the change trend of the cumulative sum distribution data, find the first power value corresponding to the weakening of the change when starting from the growth trend, and use the found first power value as the power demarcation point value between the spectrum noise and the communication signal spectrum.

[0040] The power value - spectrum point number distribution data can be regarded as the spectrum data probability density distribution function, and this density distribution function is the comprehensive result of the noise probability density distribution function and the signal probability density distribution function. If the noise probability density distribution function is known, the power boundary between the noise and the signal can be determined through further processing. However, due to the complexity of the actual electromagnetic environment, it is difficult to model the actual situation with a standard mathematical probability density distribution function (such as Gaussian distribution, Rayleigh distribution, etc.), and the actual environment is complex and changeable. Even if the noise statistics model for the current situation is very fitting, the previously established model may no longer be applicable over time, so it is difficult to accurately model.

[0041] The distribution characteristics of the noise power value and the signal power value are different. In the spectrum data, the noise power value distribution has an aggregation characteristic (as shown in Figure 3 it will pile up together), while the signal power value distribution has independent randomness (determined by the randomness of different users, distances, etc.) and a wider distribution range. In this embodiment, by cumulatively summing the power value - spectrum point number distribution data and judging the change trend of the cumulative sum distribution data, the power value corresponding to the weakening of the change when starting from the growth trend in the cumulative sum distribution data is used as the power demarcation point value, so that the power demarcation point value can be determined quickly and accurately. Thus, it is possible to stably and effectively distinguish noise and signal in a complex electromagnetic environment, without the need for precise and complex derivation and modeling in mathematics, and the differences in the noise statistical characteristics and the signal statistical characteristics can be used to distinguish the two intuitively and simply.

[0042] In a specific application embodiment, in order to determine the power demarcation point between the noise and the signal, the obtained spectrum power value - spectrum point number statistical data is cumulatively summed (that is, the integral of the spectrum data probability density distribution function is calculated); when the power value integral passes through the noise distribution area, the sum integral value will rise sharply (determined by the aggregation characteristic of the noise distribution), and the cumulative sum distribution data of the cumulative sum of spectrum point numbers changing with the spectrum power value (that is, the integral sum curve) is as shown in Figure 5As shown, the abscissa is the power value and the ordinate is the integral of the cumulative number of spectrum points before this power value. When the statistical growth trend change of the noise power value begins to weaken (i.e., Figure 5 when the derivative value of the curve starts to decrease from the maximum value), it can be considered as the power demarcation point between the noise and the signal.

[0043] In step S3, in addition to determining the power demarcation point value in the above-mentioned manner of steps S31 and S32, it can also be determined by another method, and its implementation steps are as follows: Step A31, find the maximum value of the spectrum point statistical data and the corresponding second power value from the power value - spectrum point distribution data; Step A32, traverse in the direction greater than the second power value until the spectrum points corresponding to the found third power value are less than the maximum statistical value of the specified ratio, and take the found third power value as the power demarcation point value between the spectrum noise and the communication signal spectrum.

[0044] Among them, in step A32, when traversing in the direction greater than the second power value, if the third power value that meets the preset conditions is not found after the traversal is completed, then take the last power value in the power value - spectrum point distribution data as the power demarcation point value between the spectrum noise and the communication signal spectrum.

[0045] In a specific application embodiment, the detailed steps for determining the power demarcation point value according to steps A31 and A32 are as follows: For Figure 4 find the maximum value and the corresponding second power value from the power value - spectrum point distribution data, and its expression is as follows: (1) In the above formula, is the power value - spectrum point distribution statistical data value, is to find the maximum statistical value and the corresponding index position (power value), is the maximum statistical value, is the power value corresponding to the maximum statistical value.

[0046] As Figure 6 shown, search sequentially to the right for the statistical values greater than the corresponding power value. When the situation where the statistical value is less than the maximum statistical value of the specified ratio appears for the first time (the ratio coefficient is preferably 0.01 here, and the corresponding statistical value is ), then the corresponding power value (the third power value) is considered as the power demarcation point between the noise and the signal. That is, Figure 6 the abscissa of the solid dot shown in , the corresponding vertical coordinate power value is the noise floor judgment threshold. If no power value satisfying less than is found until the end of the traversed data, the last power value is used as the demarcation point between the noise and the signal power.

[0047] It should be noted that in the above two different methods for determining the power demarcation point value between the spectral noise and the communication signal spectrum in steps S31 - S32 and steps A31 - A32, in steps S31 - S32, the power value - spectral point number distribution data is first cumulatively summed, differentiated and then searched. Its computational complexity is moderate, the accuracy is high, and the robustness is stronger; in steps A31 - A32, the power value - spectral point number distribution data is directly searched, and its computational complexity is smaller, which is suitable for application scenarios with high requirements for power consumption and size such as portable and airborne platforms. Any one of the above two methods (steps S31 - S32 and steps A31 - A32) can be selectively adopted according to the actual situation (such as the performance of the hardware device, the signal characteristics of the measured communication signal, etc.) to determine the power demarcation point value, and which one to specifically select is not limited in the present invention.

[0048] It can be understood that in this embodiment, the extraction algorithm for the noise floor of the spectral data specifically adopts statistical analysis of the power value, and determines the noise floor extraction threshold according to the distribution characteristics of the spectral power value, which can avoid the problem of detecting noise signals when there is no signal in the spectrum caused by the traditional method of multiplying the difference between the maximum value and the minimum value by a proportional coefficient as the noise floor extraction threshold, and can also solve the problem of low automation degree of traditional segmented spectral signal detection, and at the same time improve the applicability flexibility of the detection.

[0049] In this embodiment, step S4 includes: removing the spectral data corresponding to the power values greater than the demarcation point value in the communication signal spectral data, and filling it with the noise floor spectral data within a neighboring preset range to obtain the noise floor data.

[0050] In a specific application embodiment, after determining the noise floor threshold, the extraction of the spectral noise floor can be carried out. The spectral values greater than the noise floor threshold are signals. Remove the spectral power values within the signal frequency range, and use the noise spectral power values of the adjacent points at the signal frequency edge to fill the removed spectral power values within this section of the signal frequency range, and then the noise floor data is obtained. As Figure 7 shown is the noise floor data extracted in a specific application embodiment.

[0051] In this embodiment, step S5 includes: Performing exponentially weighted filtering and smoothing on the spectral noise floor data to obtain the first smoothed noise floor data; Flipping the spectral noise floor data left and right, then performing exponentially weighted filtering and then flipping left and right again to obtain the second smoothed noise floor data; Sum and average the second smoothed noise floor data and the first smoothed noise floor data to obtain the final noise floor fitting curve.

[0052] In a specific application embodiment, after obtaining the noise floor data, perform smoothing fitting on the noise floor data. Preferably, an exponential weighted fitting smoothing method can be used, and the processing flow is as Figure 8 shown. After obtaining the noise floor data, flip the data left and right, and then perform exponential weighted filtering smoothing respectively. The smoothing coefficient alpha is preferably 0.002. After the two sets of data are smoothed by filtering, the flipped noise floor data needs to be flipped left and right once again, and then summed with the other set of data and divided by 2 to obtain the final fitted and smoothed noise floor data curve. The fitted and smoothed noise floor data graph is as Figure 9 shown.

[0053] It should be noted that there are mainly two methods for the noise floor fitting method: high-order Fourier series fitting of the spectral noise floor and high-order polynomial fitting of the spectral noise floor. High-order Fourier series fitting first uses Fourier series to fit the noise floor of each spectral data to reduce its impact on spectral sensing, and then performs energy detection on the processed spectrum. High-order polynomial fitting is performed among a large number of sample points, that is, discrete points, that is, the noise floor and the corresponding frequency point sequence are separated, and a polynomial that satisfies the sample point distribution is found according to the corresponding relationship. It is very sensitive to the selection of the order of the fitting model. Selecting a suitable order model for specific data can obtain good results, but when the order or data changes, the results are not very ideal, and it can be flexibly selected according to the actual situation. The above fitting methods have strong robustness and have been widely and maturely applied. Additionally, preferably, the method in Chinese Patent Publication No. CN113726348A (A Smoothing Filtering Method and System for Radio Signal Spectrum) can be used for smoothing fitting. This smoothing fitting method does not use the above two high-order fitting algorithms, not only greatly reducing the computational complexity, but also being able to adapt to spectral data in various situations, avoiding the need to continuously adjust the order of the fitting algorithm for different spectral data to obtain a relatively ideal result. It has strong robustness and can adapt to various spectral data. Its specific implementation steps are not elaborated here.

[0054] After obtaining the fitted and smoothed noise floor data graph as Figure 9 shown, preferably, the overall can be increased by 3 dB as the signal detection threshold (the specific increase value can be changed according to actual requirements), as Figure 10 shown. It can be seen that the signal detection threshold adaptively extracted according to the noise floor well fits the change trend of the noise floor in the spectral data.

[0055] As Figure 11The figure shows the signal detection threshold curve obtained by applying the above method of this embodiment to two segments of spectrum data in the actual electromagnetic environment. The abscissa is the frequency and the ordinate is the power. It can be seen that the above method of this embodiment has a good fitting effect, can effectively fit to obtain the noise floor fitting curve, improve the accuracy of communication spectrum signal detection in complex environments, and avoid the influence of uneven noise floor on detection. It can be understood that through the communication spectrum signal detection method based on noise floor adaptive extraction of the present invention, the demarcation point between noise and signal can be effectively and accurately identified, the error that may be brought by artificially setting the threshold is avoided, the accuracy of noise floor extraction is improved, and it can be applied to signal detection in various communication scenarios, with high reliability and good versatility.

[0056] The present invention further provides a communication spectrum signal detection system based on noise floor adaptive extraction, including a microprocessor and a memory connected to each other. The microprocessor is programmed or configured to execute the above communication spectrum signal detection method based on noise floor adaptive extraction.

[0057] The present invention further provides a computer-readable storage medium. A computer program is stored in the computer-readable storage medium. The computer program is used to be programmed or configured by the microprocessor to execute the above communication spectrum signal detection method based on noise floor adaptive extraction. The system and medium of the present invention, corresponding to the above method, also have the advantages as described in the above method.

[0058] The implementation of all or part of the processes in the above-described embodiment methods of the present invention can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium includes: any entity or device capable of carrying computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium, etc. The memory is used to store the computer program and / or module. The processor realizes various functions by running or executing the computer program and / or module stored in the memory, and by calling the data stored in the memory. The memory can include high-speed random access memory, and can also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one magnetic disk storage device, flash device, or other volatile solid-state storage devices, etc.

[0059] The above description is only a preferred embodiment of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions falling within the concept of the present invention belong to the protection scope of the present invention. It should be noted that for those of ordinary skill in the art in the technical field, without departing from the principle of the present invention, several improvements and refinements should also be regarded as within the protection scope of the present invention.

Claims

1. A communication spectrum signal detection method based on adaptive noise floor extraction, characterized in that: include: Acquire the measured communication time domain signal and perform Fourier transform to obtain the communication signal spectrum data; Counting the number of spectrum points of the communication signal spectrum data in each spectrum power value interval, and obtaining power value-spectrum point distribution data of the spectrum points changing with the power value; Determine the power demarcation point value between the spectrum noise and the communication signal spectrum according to the power value-spectrum point number distribution data; Extracting spectrum noise floor data from the communication signal spectrum data according to the power cutoff point value; The extracted spectrum noise floor data is fitted to form a noise floor fitting curve, and the noise floor fitting curve is used as an adaptive detection threshold to detect the final communication spectrum signal from the original communication signal.

2. The communication spectrum signal detection method based on noise floor adaptive extraction according to claim 1 is characterized in that: The determining of the power demarcation point value between the spectrum noise and the communication signal spectrum according to the power value-spectrum point number distribution data comprises: Cumulatively summing the power value-spectrum point number distribution data to obtain cumulative sum distribution data in which the cumulative sum of spectrum points varies with the spectrum power value; Determine the change trend of the cumulative sum distribution data, find out the first power value corresponding to the time when the change from the growth trend begins to weaken, and use the found first power value as the power dividing point value between the spectrum noise and the communication signal spectrum.

3. The communication spectrum signal detection method based on noise floor adaptive extraction according to claim 1 is characterized in that: The step of determining the power demarcation point value between the spectrum noise and the communication signal spectrum according to the power value-spectrum point number distribution state comprises: Finding the maximum value of the spectrum point number statistical data and the second power value corresponding to the maximum value from the power value-spectrum point number distribution data; Traverse and search in the direction greater than the second power value until the number of spectrum points corresponding to the third power value found is less than the maximum statistical value of the specified ratio, and use the third power value found as the power dividing point value between the spectrum noise and the communication signal spectrum.

4. The communication spectrum signal detection method based on noise floor adaptive extraction according to claim 3 is characterized in that: When traversing and searching in a direction greater than the second power value, if the third power value that meets the preset conditions is still not found after the traversal is completed, the last power value in the power value-spectrum point distribution data is used as the power dividing point value between the spectrum noise and the communication signal spectrum.

5. The communication spectrum signal detection method based on noise floor adaptive extraction according to claim 1 is characterized in that: The method further includes performing mean filtering on the communication signal spectrum data after counting the number of spectrum points in each spectrum power value interval and obtaining the power value-spectrum point distribution data of the spectrum points changing with the spectrum power value.

6. The communication spectrum signal detection method based on noise floor adaptive extraction according to any one of claims 1 to 5, characterized in that: The acquiring of the measured communication time domain data and performing Fourier transform to obtain the communication signal spectrum data comprises: According to the number of sampling points of the communication signal time domain data, the measured communication signal is fast Fourier transformed in steps of a preset number of points to obtain short-time Fourier transform time-frequency data, and then the short-time Fourier transform time-frequency data is median filtered at multiple points before and after the time dimension to obtain the filtered communication signal spectrum data.

7. The communication spectrum signal detection method based on noise floor adaptive extraction according to any one of claims 1 to 5, characterized in that: Extracting spectrum noise floor data from the communication signal spectrum data according to the power demarcation point value includes: The spectrum data corresponding to the power values ​​greater than the demarcation point value in the spectrum data of the communication signal are removed, and the spectrum data corresponding to the power values ​​greater than the demarcation point value are filled with the noise spectrum power values ​​of the points adjacent to the frequency edge of the communication signal to obtain the noise floor data.

8. The communication spectrum signal detection method based on noise floor adaptive extraction according to any one of claims 1 to 5, characterized in that: The step of fitting the extracted spectrum noise floor data to form a noise floor fitting curve comprises: Performing exponential weighted filtering and smoothing on the spectrum noise floor data to obtain first smoothed noise floor data; The spectrum noise floor data is flipped left-right, and then subjected to exponential weighted filtering and then flipped left-right again to obtain second smoothed noise floor data; The second smoothed noise floor data and the first smoothed noise floor data are summed and averaged to obtain a final noise floor fitting curve.

9. A communication spectrum signal detection system based on noise floor adaptive extraction, comprising a microprocessor and a memory connected to each other, characterized in that: The microprocessor is programmed or configured to execute the communication spectrum signal detection method based on adaptive noise floor extraction as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored therein, characterized in that: The computer program is used to be programmed or configured by a microprocessor to execute the communication spectrum signal detection method based on adaptive noise floor extraction as described in any one of claims 1 to 8.

Images