CN108270495B - Background noise extraction method and system - Google Patents
Background noise extraction method and system Download PDFInfo
- Publication number
- CN108270495B CN108270495B CN201810083729.8A CN201810083729A CN108270495B CN 108270495 B CN108270495 B CN 108270495B CN 201810083729 A CN201810083729 A CN 201810083729A CN 108270495 B CN108270495 B CN 108270495B
- Authority
- CN
- China
- Prior art keywords
- point
- sample
- signal
- points
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000000605 extraction Methods 0.000 title claims description 29
- 238000000034 method Methods 0.000 claims abstract description 56
- 238000005070 sampling Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims description 2
- 238000004422 calculation algorithm Methods 0.000 abstract description 10
- 239000000284 extract Substances 0.000 abstract description 4
- 238000000926 separation method Methods 0.000 abstract description 3
- 238000001228 spectrum Methods 0.000 description 23
- 238000012544 monitoring process Methods 0.000 description 19
- 238000012545 processing Methods 0.000 description 10
- 238000004088 simulation Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 6
- 230000000630 rising effect Effects 0.000 description 6
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 4
- 238000005259 measurement Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000000342 Monte Carlo simulation Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000003064 k means clustering Methods 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 239000002023 wood Substances 0.000 description 2
- 241000543540 Guillardia theta Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/309—Measuring or estimating channel quality parameters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/309—Measuring or estimating channel quality parameters
- H04B17/345—Interference values
Abstract
The invention discloses a method and a system for extracting background noise, wherein the method comprises the following steps: the method comprises the steps of obtaining a level sample data set, determining a plurality of sample subsets, determining a fitting curve of each sample subset according to a least square method, determining a slope characteristic value of each point to be measured according to the fitting curve, replacing a level numerical value of a signal point with a level numerical value of an adjacent noise point according to each slope characteristic value signal point, wherein the adjacent noise sample point is a noise point closest to the signal point. The method adopts a least square linear curve fitting algorithm to process the level data corresponding to each frequency point in the frequency band, extracts the slope characteristic value of a sample, realizes signal-noise separation by comparing the characteristic value with a set signal judgment threshold, and replaces the level value corresponding to the signal frequency point value with the level value corresponding to the adjacent noise frequency point, thereby accurately obtaining the background noise of the frequency band.
Description
Technical Field
The invention relates to the field of radio monitoring, in particular to a method and a system for extracting background noise.
Background
With the rapid development of radio technology and signal processing technology, the application of radio services has penetrated into various fields, radio spectrum resources are increasingly strained, and various radio interferences occur at times. In order to ensure that radio spectrum resources are reasonably and efficiently used and radio interference is timely found and eliminated, a radio management department continuously strengthens monitoring on the radio spectrum resources and pays close attention to the dynamic use of the radio spectrum, and one important work is measuring the occupancy rate of the radio spectrum. By measuring the occupancy rate of the frequency spectrum, a frequency spectrum manager can determine whether a channel or a frequency band is effectively utilized, timely master the actual use condition of the frequency spectrum and provide a basis for assigning the frequency to a new user; the spectrum management department can know the spectrum use trend in time and evaluate the frequency planning. According to technical specification (trial) of ultra-short wave band occupancy test, in the work of measuring the spectrum occupancy, an occupancy statistical threshold is specified to be 5dB above the average noise power level or voltage value of a local receiver. In practice, the superposition of various signals in the air causes the noise levels of different frequency bands to be not unique and to change with time, so that a method for dynamically determining the noise levels must be considered. The background noise in the frequency band is accurately extracted, and the occupancy rate measurement threshold is set according to the background noise, so that the accuracy of the occupancy rate statistical result can be improved. The current commonly used method for manually acquiring background noise has large error, and because the background noise in a frequency band has uneven fluctuation, the background noise must be manually segmented for interpretation, so that the workload is greatly increased, and the error is large. Radio monitoring workers and experts at home and abroad begin to research on the measurement of radio background noise, and provide a reasonable background noise measurement method which is actively applied to monitoring practices, such as background noise extraction based on textural features, a neighborhood value discrimination extraction algorithm, a K-means clustering extraction algorithm and the like.
The adjacent value distinguishing and extracting algorithm is greatly influenced by signal bandwidth and equipment difference, and when the signal is a broadband signal, the extracting effect is poor. The background noise extraction method based on the texture features does not provide a determination method of a threshold value, the size of a smooth window is also an engineering empirical value, and when the bandwidth of a signal in a frequency band is larger than that of the smooth window and the envelope of the signal is relatively flat, the signal is misjudged as noise. The background noise extracted by the K-means clustering extraction algorithm is still a fixed value, and the occupancy degree threshold set according to the fixed value is still a fixed threshold, so that the problem of background noise variation caused by continuous change of the actual electromagnetic environment cannot be solved.
Disclosure of Invention
The invention aims to provide a method and a system for extracting background noise, which can accurately extract the background noise.
In order to achieve the purpose, the invention provides the following scheme:
a method of extracting background noise, the method comprising:
acquiring a level sample data set, wherein the level sample data set comprises a plurality of sample points, and the sample points are divided into signal points and noise points;
determining a plurality of sample subsets, wherein each sample subset comprises N continuous sample points in the level sample data sets, and a first sample point in each sample subset is a point to be measured; two adjacent sample subsets comprise the same sample point;
determining a fitted curve for each of the sample subsets according to a least squares method;
determining a slope characteristic value of each point to be measured according to the fitting curve;
determining a plurality of starting points and a plurality of ending points according to the slope characteristic values, wherein the slope characteristic value of the starting points is greater than or equal to T, the slope characteristic value of a point to be measured which is positioned before the starting points and is closest to the starting points is less than T, the slope characteristic value of the ending points is greater than-T, the slope characteristic value of the point to be measured which is positioned before the ending points and is closest to the ending points is less than or equal to-T, the previous time refers to the time sequence of the points to be measured, and T is a set threshold;
determining all sample points between a first signal point and a second signal point in the level sample data set as third signal points, wherein the first signal point is a starting point, the second signal point is an ending point, the first signal point is located before the second signal point, and the first signal point is the closest starting point to the second signal point;
and replacing the level value of the third signal point with the level value of an adjacent noise point, wherein the adjacent noise sample point is the noise point closest to the first signal point or the second signal point.
Optionally, before the determining the plurality of sample subsets, the method further includes:
acquiring the narrowest signal bandwidth and the resolution bandwidth in the frequency domain service signal;
determining the maximum value N of the fitting point number according to the narrowest signal bandwidth and the resolution bandwidthmax；
Determining the number of fitting points N, wherein 3<N<Nmax。
Optionally, the maximum value N of the number of fitting points is determined according to the narrowest signal bandwidth and the resolution bandwidthmaxIn particularComprises the following steps:
according to formula Nmax＝BWmin/(2 × RBW) maximum number of fitting points N is determinedmax；
Wherein, BWminRBW is the resolution bandwidth for the narrowest signal bandwidth.
Optionally, after replacing the level value of the third signal point with the level value of the adjacent noise point, the method further includes:
calculating the level difference value of two adjacent sample points;
judging whether the absolute value of the level difference is larger than 5 to obtain a first judgment result;
and when the first judgment result shows that the absolute value of the level difference value of two adjacent sample points is greater than 5, replacing the value of the sample point with a larger value with the value of the sample point with a smaller value.
Optionally, after the sample set of the level to be measured is obtained, the method further includes:
and replacing the level value of the first sample point of the level sample data set with the level value of the noise point closest to the first sample point.
Optionally, the determining the slope characteristic value of each point to be measured according to the fitted curve specifically includes:
according to the formula:
determining a slope characteristic value of the point to be measured at the fitting curve;
wherein, KpThe slope characteristic value of the point to be measured of the p sample subset, (x)n，yn) For the nth sample point, x, in the subset of samplesn＝n，ynIs the level value of the nth sample point.
Optionally, N is more than or equal to 5 and less than or equal to 9.
Optionally, N is 6.
Optionally, T is 0.3.
An extraction system of background noise, the extraction system comprising:
the system comprises a level sample data set acquisition module, a level sampling data set acquisition module and a level detection module, wherein the level sample data set to be detected comprises a plurality of sample points, and the sample points are divided into signal points and noise points;
the sample subset determining module is used for determining a plurality of sample subsets, each sample subset comprises N continuous sample points in the level sample data sets, and a first sample point in each sample subset is a point to be measured; two adjacent sample subsets comprise the same sample point;
a fitted curve determining module for determining a fitted curve for each of the sample subsets according to a least squares method;
the slope characteristic value determining module is used for determining the slope characteristic value of each point to be measured according to the fitting curve;
a starting point and ending point determining module, configured to determine multiple starting points and multiple ending points according to the slope characteristic values, where a slope characteristic value of a starting point is greater than or equal to T, a slope characteristic value of a point to be measured, which is located before the starting point and closest to the starting point, is less than T, a slope characteristic value of an ending point is greater than-T, a slope characteristic value of a point to be measured, which is located before the ending point and closest to the ending point, is less than or equal to-T, where the previous is a time sequence of the point to be measured, and T is a set threshold;
a signal point determining module, configured to determine that all sample points in the level sample data set between a first signal point and a second signal point are third signal points, where the first signal point is a start point, the second signal point is an end point, the first signal point is located before the second signal point, and the first signal point is a start point closest to the second signal point;
and the replacing module is used for replacing the level value of the third signal point with the level value of an adjacent noise point, wherein the adjacent noise sample point is the noise point closest to the first signal point or the second signal point.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention provides a method and a system for extracting background noise, which adopt a least square linear curve fitting algorithm to process level data corresponding to each frequency point in a frequency band, extract a slope characteristic value of a sample, realize signal-noise separation by comparing the characteristic value with a set signal judgment threshold, and replace a level value corresponding to a signal frequency point value with a level value corresponding to an adjacent noise frequency point, thereby accurately obtaining the background noise of the frequency band.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a flow chart of a method for extracting background noise according to the present invention;
FIG. 2 is a frequency spectrum diagram of a local FM broadcast band;
FIG. 3 is a schematic diagram of a frequency domain signal;
FIG. 4 is a least squares curve fit plot of N sampling points of the frequency domain signal;
FIG. 5 is a simulation diagram obtained by performing 10 ten thousand simulations on the distribution of the fitted curve slope characteristic values by using a Monte Carlo method;
FIG. 6 is a background noise curve obtained by processing frequency spectrum data of 3-30 MHz short wave frequency band data collected from a place in Wulu wood City in Xinjiang by using the method of the present invention;
FIG. 7 is a background noise curve obtained by processing acquired frequency spectrum data of 88-108 MHz frequency modulation broadcast frequency band data by using the method of the invention;
FIG. 8 is a background noise curve obtained by processing the collected frequency spectrum data of 108-1000 MHz ultrashort wave frequency band data by using the method of the present invention;
fig. 9 is a connection diagram of the structure of the background noise extraction system according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a method and a system for extracting background noise based on least square linear curve fitting.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
Fig. 1 is a flowchart of a background noise extraction method according to the present invention. As shown in fig. 1, the extraction method includes:
step 101: acquiring a level sample data set;
step 102: determining a plurality of sample subsets, wherein each sample subset comprises N continuous sample points in the level sample data sets, and a first sample point in each sample subset is a point to be measured; two adjacent sample subsets comprise the same sample point. During specific operation, the point to be measured of the first sample subset is the first sample point in the level sample data set, the point to be measured of the second sample subset is the second sample point in the level sample data set, and so on.
Before step 102, the method further comprises:
step A1: acquiring the narrowest signal bandwidth and the resolution bandwidth in the frequency domain service;
the service Radio services include: radio communication service, fixed service, satellite fixed service, aviation fixed service, inter-satellite service, space operation service, mobile service, satellite mobile service, land mobile service, satellite land mobile service, water mobile service, satellite water mobile service, port operation service, ship mobile service, aviation mobile (R) service, aviation mobile (OR) service, satellite aviation mobile (R) service, satellite aviation mobile (OR) service, broadcast service, satellite broadcast service, radiodetermination service, satellite radiodetermination service, radio navigation service, satellite radionavigation service, water radionavigation service, satellite water radionavigation service, aviation radionavigation service, satellite radionavigation service, radio positioning service, satellite, Satellite radiolocalization services, weather assistance services, satellite earth sounding services, satellite weather services, standard frequency and time signal services, satellite standard frequency and time signal services, space research services, amateur services, satellite amateur services, radio astronomy services, security services, ad hoc services, etc.
Step A2: determining the maximum value N of the fitting point number according to the narrowest signal bandwidth and the resolution bandwidthmax(ii) a The method specifically comprises the following steps: determining the maximum value N of the number of fitting points according to equation (1)max；
Nmax＝BWmin/(2×RBW) (1)
Wherein, BWminRBW is the resolution bandwidth for the narrowest signal bandwidth.
Step A3: determining the number of fitting points N, wherein 3<N<Nmax. Preferably, 5. ltoreq. N.ltoreq.9. More preferably, N is 6.
Step 103: and determining a fitted curve of each sample subset according to a least square method.
Step 104: and determining the slope characteristic value of each point to be measured according to the fitting curve. The method specifically comprises the following steps:
determining a slope characteristic value of the point to be measured at the fitting curve according to a formula (2);
wherein, KpThe slope characteristic value of the point to be measured of the p sample subset, (x)n，yn) For the nth sample point, x, in the subset of samplesn＝n，ynIs the level value of the nth sample point.
Step 105: determining a plurality of starting points and a plurality of ending points according to the slope characteristic values, wherein the slope characteristic value of the starting points is greater than or equal to T, the slope characteristic value of a point to be measured, which is closest to the starting points, before the starting points is less than T, the slope characteristic value of the ending points is greater than-T, the slope characteristic value of the point to be measured, which is closest to the ending points, before the ending points is less than or equal to-T, and the front is the time sequence of the point to be measured; t is a set threshold value, namely a signal judgment threshold value.
Step 106: determining all sample points between a first signal point and a second signal point in the level sample data set as third signal points, wherein the first signal point is a starting point, the second signal point is an ending point, the first signal point is located before the second signal point, and the first signal point is the closest starting point to the second signal point;
step 107: and replacing the level value of the third signal point with the level value of an adjacent noise point, wherein the adjacent noise sample point is the noise point closest to the first signal point or the second signal point.
Optionally, after step 101, the method may further include: and replacing the level value of the first sample point of the level sample data set with the level value of the noise point closest to the first sample point.
Optionally, after step 107, the method further includes:
step B1: calculating the level difference value of two adjacent sample points;
step B2: judging whether the absolute value of the level difference is larger than 5 to obtain a first judgment result;
step B3: and when the first judgment result shows that the absolute value of the level difference value of two adjacent sample points is greater than 5, replacing the value of the sample point with a larger value with the value of the sample point with a smaller value.
Step B1-step B3 are equivalent to filtering, and missing signal points are searched and replaced by noise points. According to technical specification (trial) of ultrashort wave band occupancy test, in the work of measuring the spectrum occupancy, an occupancy statistical threshold is specified to be 5dB above the average noise power level or voltage value of a local receiver, so that one point is absolutely a signal point when the difference between adjacent points is greater than 5.
In engineering, the main source of noise in a monitoring system is thermal noise, which is typically white gaussian noise, and follows a normal distribution. Noise is randomly varying with time, noise has a piecewise continuity, noise of a channel adjacent to the radio signal is considered continuous, the signal is superimposed on background noise, and the spectrum of the signal is abrupt. Fig. 2 is a frequency spectrum diagram of a certain fm broadcast frequency band.
By analyzing the raw data collected by the monitoring system, although the noise level values of the adjacent channels are different and the level changes nondirectionally, the change amplitude is small. By counting the level values of the channels at the same time and then drawing a curve, it can be seen that the curve of the noise part is relatively flat and the curve of the signal part suddenly becomes steep in the whole frequency band, so that the signal and the noise can be easily distinguished by human eyes, but the computer cannot identify the signal and the noise according to the phenomenon. Therefore, the characteristics of the curve can be extracted by a certain method, and the computer can have the identification capability by judging the specific characteristic value.
Through analysis, the signal in the frequency domain has a distinct characteristic in geometry, that is, there are more distinct rising edges and falling edges on both sides of the signal, as shown in fig. 3. The level value between the data sample points in the rising edge and the falling edge of the signal has large change amplitude, and the change of the amplitude shows obvious rising or falling trend. Assuming that in a vertical coordinate system, N continuous data sample points are taken in the rising edge or the falling edge of a signal, curve fitting is carried out by a least square normal curve fitting algorithm, and the fact that the inclination angle of a slope formed by fitting the N data sample points is larger than that of a slope formed by fitting the N data sample points can be seenLarge, e.g. theta in FIG. 41(ii) a After N continuous data sample points are taken in the noise section and subjected to least square linear curve fitting, the inclination angle of a fitted oblique line is smaller or tends to be zero, such as theta in figure 42Therefore, the data level sample points collected in the frequency band where the background noise is to be extracted can be divided into a plurality of sample subsets with the length of N, and then the slope characteristic value K of each sample subset is extracted by least square linear curve fittingi. Setting a signal decision threshold T, when KiWhen the value is more than or equal to T, the subsample is considered as a signal data point, and when K is greater than or equal to Ti<At T, the subsample is considered to be a noise data point. In addition, as can be seen from fig. 3 and 4, a complete signal should have both rising and falling edges, and when a sample data containing a signal is linearly fitted, if a subsample { S is calculated for the first timei,Si+1,…,SN+i-1The slope characteristic value K ofiIf T is greater than T, the sample is considered as signal data and belongs to the beginning of the rising edge of the signal, and the mark i-m1Continuing to perform linear fitting when the slope characteristic value K isj>-T(j>1) And Kj-1When the value is less than or equal to T, the group of data is considered as signal data and belongs to the end part of the falling edge of the signal, and the mark j is m2Sample point S (m)1) To S (m)2) Considered as a signal point, its value is replaced by the value of the neighboring noise sample point. After the method is used for processing the original level sample data of a certain frequency band, the separation of signal frequency points and noise frequency points can be realized, and then an ideal background noise sample is extracted.
1. Calculation of slope characteristic of curve
Let be a set of N points, where the abscissa
Is determined. The least square fitting curve y ═ f (x) ═ Kx + B satisfies the root mean square error E2(f) The smallest curve. The coefficients of the least squares curve are K and B are the solutions of the following linear equation (3), which is called the normal equation:
the slope of the fitted curve can be calculated by linear equation (3):
2. determination of the number of fitting points N
If the noise of each frequency point is in Gaussian distribution and the noise of the continuous N frequency points is independently and identically distributed, the noise of the N frequency points obtained according to the law of large numbers also follows normal distribution; it can be proved that the slope characteristic value of the least square curve fitting of the noise of the N frequency points which obeys the normal distribution approaches to zero. When the number of the selected samples N is larger, the samples approach to a normal distribution. However, the number of points selected is at least less than half of the number of signal points to provide the desired conditions for application of the method. Suppose the narrowest signal bandwidth of all the services in the frequency band is BWminThe resolution bandwidth is RBW, and the selected point number is less than Nmax＝BWminV. (2. multidot. RBW). So that the number of points N suggested to be selected is finally in the interval (3, N)max) And (4) the following steps. Before the occupancy rate statistics, the resolution bandwidth of the receiver is preferably set reasonably according to the occupancy bandwidth of each service signal in the frequency band to be counted so as to ensure Nmax>3。
The fitting point number N has a large influence on the result of background noise extraction, when curve fitting is carried out to extract the slope characteristic value, if the number of points is too large, the slope characteristic value of the extracted signal point is not obvious, if the number of points is too small, the slope characteristic value of the adjacent noise point with large fluctuation is higher than threshold characteristic response judgment, and through practical application, the fitting point number generally has a good effect of 5-9.
3. Selection of signal decision threshold
The signal judgment threshold T is reasonably set, and if the T is set to be too large, a signal sample point with low signal-to-noise ratio is judged to be noise; if T is set to be too small, the calculated amount is increased, and noise sample data with large fluctuation in the frequency band is judged as a signal. The invention provides an estimation method of a signal judgment threshold T according to the actual monitoring condition. It is assumed that the main source of noise in the monitoring system is thermal noise, which is typically white gaussian noise, and follows a normal distribution. By observing the spectral data, it was found that the level difference between adjacent noise points was substantially within 1 dB. The signal decision threshold T can be estimated by simulation. According to the invention, data for fitting are assumed to obey normal distribution, when the number of fitting points N is 6, 10 ten thousand times of simulation is carried out on the distribution condition of the curve slope characteristic value after fitting by adopting a Monte Carlo method, and the simulation result is shown in figure 5. As can be seen from fig. 5, the slope characteristic values after 6-point curve fitting approximately follow normal distribution, and the probability that the slope characteristic values are all less than 0.3 is 80%. Therefore, when performing background noise extraction, it may be considered to set T to 0.3. Through practical tests, when the fitting point number N is 6, the judgment threshold T is 0.3, and the extracted background noise meets the working requirement of occupancy rate statistics.
In addition, because performance indexes of different receiving devices are different, the signal judgment threshold T estimated according to simulation is not necessarily the optimal value, and when the method is actually applied to extracting background noise, the size of the T can be properly adjusted on the basis of the signal judgment threshold T estimated by simulation, and an experiment is carried out in advance to determine the optimal judgment threshold value suitable for a specific receiving device.
4. Background noise extraction method verification
In order to verify the feasibility and the extraction effect of the method, the invention utilizes an ESMD monitoring receiver of the company Roder and Schwarz and combines a monitoring system based on ESMD integration to construct a set of monitoring platform, actual monitoring is carried out in different frequency bands, and monitoring data of each frequency band is collected and stored for calling and processing by the method provided by the invention. In order to efficiently process monitoring data, the invention writes a background noise curve extraction program in an MATLAB development environment, and realizes automatic processing and visualization of frequency spectrum data. In this experiment, the number of fitting points N used in the background noise automatic extraction algorithm is 6, and the signal determination threshold T is 0.3.
FIG. 6 is a background noise curve obtained by processing frequency spectrum data of short wave frequency band data of 3-30 MHz collected from a place in Wulu wood City in Xinjiang according to the method of the present invention.
As shown in fig. 6, the bottom coarser curve is the background noise curve extracted by the method. When the frequency band data is collected, the ESMD monitoring receiver sets a panoramic scanning mode (pscan), a scanning step (step) is set to be 0.5kHz, and the unit of the signal level is dBuV. It can be seen from the figure that the fluctuation of the background noise in the frequency band is large, and the difficulty of acquiring the background noise by adopting a manual interpretation mode is large. It can be seen that the extracted background noise curve is substantially identical to the actual background noise with little error.
FIG. 7 is a background noise curve obtained by processing acquired frequency spectrum data of 88-108 MHz frequency modulation broadcast frequency band data by using the method of the invention.
As shown in fig. 7, the bottom thicker curve is the background noise curve extracted by the algorithm. When the frequency band data is collected, the ESMD monitoring receiver sets a panoramic scanning mode (pscan), a scanning step (step) is set to be 3.125kHz, and the unit of the signal level is dBuV.
FIG. 8 is a background noise curve obtained by processing the collected frequency spectrum data of 108-1000 MHz ultrashort wave band data by using the method of the present invention, and the thicker curve at the bottom is the background noise curve extracted by the algorithm. When the frequency band data is collected, the ESMD monitoring receiver sets a panoramic scanning mode (pscan), a scanning step (step) is set to be 12.5kHz, and the unit of the signal level is dBuV.
The method for extracting the radio background noise curve can accurately extract the background noise on the scanning frequency band of the whole receiver, has the effect less influenced by factors such as hardware equipment, frequency band background noise distribution, signal bandwidth and the like, and is practical and convenient. The method is applied to daily radio monitoring work of the unit, the effect is ideal through practice verification, the problem that the noise estimation value is lower than the true value in the ITU-RP.372 method is solved, the mode that the ITU-RSM.1753 method depends on manual extraction of the noise sample value is improved, and the automation degree of frequency spectrum occupancy degree statistical threshold setting and the accuracy of the measurement result in radio monitoring are improved.
The invention also provides a system for extracting the background noise, and fig. 9 is a structural connection diagram of the system for extracting the background noise. As shown in fig. 9, the extraction system includes:
a level sample data set obtaining module 901, configured to obtain a level sample data set to be detected, where the level sample data set includes multiple sample points, and the sample points are divided into signal points and noise points;
a sample subset determining module 902, configured to determine a plurality of sample subsets, where each sample subset includes N consecutive sample points in the level sample data set, and a first sample point in the sample subset is a point to be measured; two adjacent sample subsets comprise the same sample point;
a fitted curve determining module 903, configured to determine a fitted curve of each sample subset according to a least square method;
a slope characteristic value determining module 904, configured to determine a slope characteristic value of each point to be measured according to the fitted curve;
a start point and end point determining module 905, configured to determine multiple start points and multiple end points according to each slope feature value, where a slope feature value of a start point is greater than or equal to T, a slope feature value of a point to be measured before the start point and closest to the start point is less than T, a slope feature value of an end point is greater than-T, a slope feature value of a point to be measured before the end point and closest to the end point is less than or equal to-T, where the preceding is a time sequence of the point to be measured, and T is a set threshold;
a signal point determining module 906, configured to determine that all sample points in the level sample data set between a first signal point and a second signal point are third signal points, where the first signal point is a start point, the second signal point is an end point, the first signal point is located before the second signal point, and the first signal point is a start point closest to the second signal point;
a replacing module 907, configured to replace the level value of the third signal point with a level value of an adjacent noise point, where the adjacent noise sample point is a noise point closest to the first signal point or the second signal point.
For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principle and the implementation mode of the invention are explained by applying a specific example, and the description of the embodiment is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (10)
1. A method for extracting background noise, the method comprising:
acquiring a level sample data set, wherein the level sample data set comprises a plurality of sample points, and the sample points are divided into signal points and noise points;
determining a plurality of sample subsets, wherein each sample subset comprises N continuous sample points in the level sample data sets, and a first sample point in each sample subset is a point to be measured; two adjacent sample subsets comprise different sample points;
determining a fitted curve for each of the sample subsets according to a least squares method;
determining a slope characteristic value of each point to be measured according to the fitting curve;
determining a plurality of starting points and a plurality of ending points according to the slope characteristic values, wherein the slope characteristic value of the starting points is greater than or equal to T, the slope characteristic value of a point to be measured which is positioned before the starting points and is closest to the starting points is less than T, the slope characteristic value of the ending points is greater than-T, the slope characteristic value of the point to be measured which is positioned before the ending points and is closest to the ending points is less than or equal to-T, the previous time refers to the time sequence of the points to be measured, and T is a set threshold;
determining all sample points between a first signal point and a second signal point in the level sample data set as third signal points, wherein the first signal point is a starting point, the second signal point is an ending point, the first signal point is located before the second signal point, and the first signal point is the closest starting point to the second signal point;
and replacing the level value of the third signal point with the level value of an adjacent noise sample point, wherein the adjacent noise sample point is a noise point closest to the first signal point or the second signal point.
2. The extraction method according to claim 1, further comprising, before the determining the plurality of sample subsets:
acquiring the narrowest signal bandwidth and the resolution bandwidth in the frequency domain service signal;
determining the maximum value N of the fitting point number according to the narrowest signal bandwidth and the resolution bandwidthmax；
Determining the number of fitting points N, wherein 3<N<Nmax。
3. The extraction method according to claim 2, wherein the maximum value N of the number of fitting points is determined according to the narrowest signal bandwidth and the resolution bandwidthmaxThe method specifically comprises the following steps:
according to formula Nmax＝BWmin/(2 × RBW) maximum number of fitting points N is determinedmax；
Wherein, BWminRBW is the resolution bandwidth for the narrowest signal bandwidth.
4. The extraction method according to claim 1, further comprising, after the replacing the level value of the third signal point with the level value of the adjacent noise point:
calculating the level difference value of two adjacent sample points;
judging whether the absolute value of the level difference is larger than 5 to obtain a first judgment result;
and when the first judgment result shows that the absolute value of the level difference value of two adjacent sample points is greater than 5, replacing the value of the sample point with a larger value with the value of the sample point with a smaller value.
5. The extraction method according to claim 1, further comprising, after acquiring the sample set of levels to be measured, the following steps:
and replacing the level value of the first sample point of the level sample data set with the level value of the noise point closest to the first sample point.
6. The extraction method according to claim 1, wherein the determining the slope characteristic value of each point to be measured according to the fitted curve specifically comprises:
according to the formula:
determining a slope characteristic value of the point to be measured at the fitting curve;
wherein, KpThe slope characteristic value of the point to be measured of the p sample subset, (x)n，yn) For the nth sample point, x, in the subset of samplesn＝n，ynIs the level value of the nth sample point.
7. The extraction method according to claim 2, wherein N is 5. ltoreq. N.ltoreq.9.
8. The extraction method according to claim 7, wherein N is 6.
9. The extraction method according to claim 1, wherein T is 0.3.
10. A background noise extraction system, the extraction system comprising:
the system comprises a level sample data set acquisition module, a level sampling data set acquisition module and a level detection module, wherein the level sample data set to be detected comprises a plurality of sample points, and the sample points are divided into signal points and noise points;
the sample subset determining module is used for determining a plurality of sample subsets, each sample subset comprises N continuous sample points in the level sample data sets, and a first sample point in each sample subset is a point to be measured; two adjacent sample subsets comprise different sample points;
a fitted curve determining module for determining a fitted curve for each of the sample subsets according to a least squares method;
the slope characteristic value determining module is used for determining the slope characteristic value of each point to be measured according to the fitting curve;
a starting point and ending point determining module, configured to determine multiple starting points and multiple ending points according to the slope characteristic values, where a slope characteristic value of a starting point is greater than or equal to T, a slope characteristic value of a point to be measured, which is located before the starting point and closest to the starting point, is less than T, a slope characteristic value of an ending point is greater than-T, a slope characteristic value of a point to be measured, which is located before the ending point and closest to the ending point, is less than or equal to-T, where the previous is a time sequence of the point to be measured, and T is a set threshold;
a signal point determining module, configured to determine that all sample points in the level sample data set between a first signal point and a second signal point are third signal points, where the first signal point is a start point, the second signal point is an end point, the first signal point is located before the second signal point, and the first signal point is a start point closest to the second signal point;
and the replacing module is used for replacing the level value of the third signal point with the level value of an adjacent noise sample point, wherein the adjacent noise sample point is a noise point closest to the first signal point or the second signal point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810083729.8A CN108270495B (en) | 2018-01-29 | 2018-01-29 | Background noise extraction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810083729.8A CN108270495B (en) | 2018-01-29 | 2018-01-29 | Background noise extraction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108270495A CN108270495A (en) | 2018-07-10 |
| CN108270495B true CN108270495B (en) | 2020-12-11 |
Family
ID=62776931
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810083729.8A Expired - Fee Related CN108270495B (en) | 2018-01-29 | 2018-01-29 | Background noise extraction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108270495B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108627807B (en) * | 2018-08-08 | 2022-04-01 | 中国航空工业集团公司雷华电子技术研究所 | Anti-interference method for airborne radar |
| CN110471015A (en) * | 2019-09-05 | 2019-11-19 | 国网北京市电力公司 | Determination method and device, storage medium and the processor of sensor detection threshold |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1685336A (en) * | 2002-07-26 | 2005-10-19 | 摩托罗拉公司 | Method for fast dynamic estimation of background noise |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8421802B2 (en) * | 2008-03-07 | 2013-04-16 | Olympus Ndt | Peak visualization enhancement display system for use with a compressed waveform display on a non-destructive inspection instrument |
| CN103646349B (en) * | 2013-09-27 | 2017-02-08 | 华北电力大学 | Power load curve segmented identification method |
| CN104639252B (en) * | 2013-11-13 | 2017-08-25 | 海思光电子有限公司 | Optical modulator bias point locking means and system |
| CN107184203B (en) * | 2017-07-03 | 2019-07-26 | 重庆大学 | Electrocardiosignal Feature point recognition method based on adaptive set empirical mode decomposition |
-
2018
- 2018-01-29 CN CN201810083729.8A patent/CN108270495B/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1685336A (en) * | 2002-07-26 | 2005-10-19 | 摩托罗拉公司 | Method for fast dynamic estimation of background noise |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108270495A (en) | 2018-07-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107911183B (en) | Unmanned aerial vehicle identification method and system based on real-time signal characteristic analysis | |
| WO2017024692A1 (en) | Method of simulating analog circuit diagnostic fault using single measurement node | |
| CN101262288A (en) | Systems and methods for determining sensing thresholds of a multi-resolution spectrum sensing (MRSS) technique for cognitive radio (CR) systems | |
| CN105785324B (en) | Linear frequency-modulated parameter estimating method based on MGCSTFT | |
| CN107820255B (en) | Improved covariance absolute value cooperative spectrum sensing method | |
| CN107733541A (en) | Method, apparatus, equipment and the computer-readable recording medium of frequency spectrum perception | |
| CN110099019A (en) | LoRa Modulation Signal Detection Method based on deep learning | |
| CN108270495B (en) | Background noise extraction method and system | |
| CN113325277A (en) | Partial discharge processing method | |
| CN105429719B (en) | Based on power spectrum and multi-scale wavelet transformation analysis high reject signal detection method | |
| CN110045322B (en) | Short wave automatic direction finding method based on high-resolution direction finding graph intelligent identification | |
| CN109801236A (en) | A kind of photon point cloud denoising method based on mixed Gauss model | |
| CN106170139B (en) | A kind of frequency spectrum detecting method and system | |
| CN106100762B (en) | A kind of weak signal of communication detection method of cyclo-stationary spectrum analysis | |
| CN109474355A (en) | Adaptive noise THRESHOLD ESTIMATION and method for extracting signal based on spectrum monitoring data | |
| CN112668612A (en) | Partial discharge signal clustering analysis method based on grids | |
| CN108900267B (en) | Single-side right-tail goodness-of-fit inspection spectrum sensing method and device based on characteristic values | |
| CN104977602A (en) | Control method and apparatus for earthquake data acquisition construction | |
| CN112929141A (en) | Unmanned aerial vehicle detection and identification method and system based on graph-borne signal matching | |
| CN108718223B (en) | Blind spectrum sensing method for non-cooperative signals | |
| CN114614922A (en) | Frequency domain-based method for cleaning frequency spectrum data in civil aviation very high frequency communication | |
| CN106503649A (en) | A kind of short-wave signal detection recognition method based on computer vision | |
| CN105491583B (en) | A kind of interference detection method, base station, network element device and Interference Detection system | |
| CN204044085U (en) | A kind of foreign matter detection system based on light spectrum image-forming | |
| CN113824518A (en) | Non-cooperative signal detection method based on frequency spectrum |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20201211 |