CN103856276B - The real-time auto-correction method of frequency spectrum non-flat forms noise floor - Google Patents
The real-time auto-correction method of frequency spectrum non-flat forms noise floor Download PDFInfo
- Publication number
- CN103856276B CN103856276B CN201410104411.5A CN201410104411A CN103856276B CN 103856276 B CN103856276 B CN 103856276B CN 201410104411 A CN201410104411 A CN 201410104411A CN 103856276 B CN103856276 B CN 103856276B
- Authority
- CN
- China
- Prior art keywords
- data
- noise floor
- point
- frequency spectrum
- length
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Image Processing (AREA)
Abstract
The present invention discloses the real-time auto-correction method of a kind of frequency spectrum non-flat forms noise floor, comprising: frequency spectrum data obtaining step, data extraction step to be analyzed, noise floor estimation step, estimated value interpolation procedure and noise floor aligning step.The real-time auto-correction method of frequency spectrum non-flat forms noise floor of the present invention, completes the correction of the uneven noise floor of frequency spectrum in real time, automatically, efficiently, and be applicable to various types of communication system at spectrum monitoring equipment in the process can run.
Description
Technical field
The invention belongs to radio communication and digital signal processing technique field, particularly the real-time auto-correction method of a kind of frequency spectrum non-flat forms noise floor.
Background technology
For the spectrum monitoring of signal of communication, the quantity, bandwidth, modulation system etc. of the useful signal needing real-time analysis identification to be currently received.In order to realize the identification of above-mentioned parameter, first signal must be separated from background noise.At present, simple and practical frequency spectrum detecting method is the threshold detection method of fixed resolution printenv Power estimation, compares by the estimated value of frequency band and thresholding, thus obtains the testing result to interested frequency band (band of Interest, BOI).The precondition that the method is suitable for is that noise floor is smooth.But in actual applications, the noise floor of the signal received exists fluctuating mostly, now above-mentioned threshold detection method will increase false drop rate and the loss of signal, reduces the accuracy rate of subsequent parameter identification, thus causes the entirety of spectrum monitoring performance to worsen.For this reason, need to correct uneven noise floor before carrying out input.
To the calibration of pectrum noise substrate, the change needing to understand noise floor under no signal condition in essence, existing method closes transmitting apparatus, adopt the correction manually completing noise floor, data volume is large, complexity is high, is difficult to the requirement of real-time meeting spectrum monitoring analysis, and manual correction degree of fitting in the short time is low, deviation is large, after once correcting, due to the reason of signal self and the characteristic variations of receiving equipment device, noise floor can change again.Current all types of communication system all needs spectrum monitoring, does not also have the automation noise floor bearing calibration that can be applicable to dissimilar monitoring system at present.
Summary of the invention
The object of the present invention is to provide the real-time auto-correction method of a kind of frequency spectrum non-flat forms noise floor, in real time, automatically, efficiently complete the correction of the uneven noise floor of frequency spectrum in the process can run at spectrum monitoring equipment, and be applicable to various types of communication system.
The technical solution realizing the object of the invention is: the real-time auto-correction method of a kind of frequency spectrum non-flat forms noise floor, comprises the steps:
10) frequency spectrum data obtains: the raw spectroscopy data obtaining signal amplitude spectrum or power spectrum from front end Fourier transform equipment;
20) data to be analyzed are extracted: to the isometric segmentation of raw spectroscopy data, each section is searched for one by one to the maximum of points of its amplitude and power, extracted arrangement, form data to be analyzed;
30) noise floor estimation: utilize morphologic filtering to estimate the variation tendency that data to be analyzed carry out noise floor, obtain noise floor estimation value;
40) estimated value interpolation: carry out forward and backward point-to-point transmission and initial and end interpolation to noise floor estimation value pointwise, obtains the Interpolate estimation value of counting identical with raw spectroscopy data;
50) noise floor corrects: utilize that raw spectroscopy data is corresponding with the pointwise of Interpolate estimation value subtracts each other, and obtains the corrected spectrum that noise floor is smooth.
The present invention compared with prior art, its remarkable advantage:
1, real-time automatic calibration: in spectrum monitoring process, the present invention can realize the automatic calibration of each frame frequency modal data noise floor, carries out long-time and loaded down with trivial details manual calibration without the need to closing transmitting apparatus before each monitoring;
2, versatility is good: the segmentation of the frequency spectrum that first the present invention carries out is extracted, and weakens the impact that the characteristics such as the amplitude leyel of signal, bandwidth are estimated ground noise, can be common to various dissimilar communication system;
3, degree of fitting is high: the frequency spectrum that there is signal is considered as one-dimensional pattern by the present invention, utilizes the morphologic filtering in image procossing, eliminates the signal area that it exists, and simulate noise floor change, pectrum noise substrate variations trend when its result and no signal is consistent;
4, complexity is low: the data volume of single treatment of the present invention is counted lower than signal, and handling process is linear operation, and complexity is o (N), is applicable to hardware implementing.
Below in conjunction with the drawings and specific embodiments, the present invention is described in further detail.
Accompanying drawing explanation
Fig. 1 is the real-time auto-correction method flow chart of non-flat forms noise floor of the present invention.
Fig. 2 is frequency spectrum data stepwise schematic views.
Fig. 3 is structural element schematic diagram.
Fig. 4 is the operation chart utilizing structural element to analyze original signal.
Fig. 5 is the uneven pectrum noise substrate value exemplary plot utilizing the inventive method to estimate.
Fig. 6 is the frequency spectrum exemplary plot after utilizing the inventive method to correct.
Embodiment
As shown in Figure 1, the real-time auto-correction method of frequency spectrum non-flat forms noise floor of the present invention, comprises the steps:
10) frequency spectrum data obtains: the raw spectroscopy data obtaining signal amplitude spectrum or power spectrum from front end Fourier transform equipment.
Obtain raw spectroscopy data from front end Fourier transform equipment, be designated as SpecD, length is N, and wherein SpecD (i) represents amplitude or the performance number of i-th point, is a real number, such as formula (1).
SpecD=[SpecD(1),SpecD(2),…,SpecD(i),…,SpecD(N)] （1）
20) data to be analyzed are extracted: to the isometric segmentation of raw spectroscopy data, each section is searched for one by one to the maximum of points of its amplitude or power, extracted arrangement, form data to be analyzed.
Described data to be analyzed are extracted (20) step and are comprised:
21) frequency spectrum data segmentation: select the window of certain length, this window mobile in raw spectroscopy data, is divided into some non-overlapping copies successively by raw spectroscopy data and length is equal to the frequency spectrum data section of length of window;
Length is selected to be the window of K, this window mobile on frequency spectrum SpecD, the length of some sections of non-overlapping copies is divided into by frequency spectrum SpecD to be all K frequency spectrum data successively, each section is designated as Part (i), it includes K point raw spectroscopy data, operating process as shown in Figure 2, obtains after segmentation terminates:
22) maximum of points search: search for each frequency spectrum data section one by one, find out the maximum of points of amplitude or power in each section; To each segmentation Part (i), in point by point search section, the point of the maximum of amplitude or power, is designated as PMax (i);
23) data composition to be analyzed: the amplitude of each section or power maximum of points are extracted out, is arranged in order, obtains data to be analyzed.
Each section is searched spectrum value maximum of points PMax (i) and extracts composition one frame data to be analyzed out, be designated as PMax:
30) noise floor estimation: utilize morphologic filtering to estimate the variation tendency that data to be analyzed carry out noise floor, obtain noise floor estimation value.
Described noise floor estimation (30) step comprises:
31) structural element is selected: according to the feature of spectrum signal, selects one dimension platypelloid type structural element;
Described structural element selects (31) step to be specially:
Length is the platypelloid type structural element S of L,
S=s
0,s
1,…,s
L-1（4）
Wherein, L is the natural number being less than frequency spectrum data length, s
ibe wherein certain any value, i is from 0 to the integer of (L-1), s
ivalue is any real number.Fig. 3 is a platypelloid type structural element example.Data point to be analyzed in S corresponding to first point is the impact point when pre-treatment.
32) erosion operation result data obtains: utilize structural element data to be analyzed to be done to erosion operation in morphology, obtain the result data of erosion operation;
Described erosion operation result data obtains the operation of (32) step as shown in Figure 4, comprising:
321) impact point in data to be analyzed is selected: if analyze data f to be analyzed first, if data to be analyzed first point is impact point x, otherwise a bit as impact point x under setting current goal point;
322) present analysis data acquisition: from impact point x, the data to be analyzed of acquisition and structural element length L same number;
323) impact point erosion operation result calculates: present analysis data f and structural element S does erosion operation, obtains the erosion operation result of current goal point x;
Described impact point erosion operation result calculates (323) step and is specially:
Wherein, f represents data PMax, x to be analyzed and represents its impact point subscript, is the integer of a non-negative; The one-dimentional structure element of S to be length be L, s
ibe wherein certain any value, i is from 0 to the integer of (L-1).
324) erosion operation result data composition: impact point erosion operation result is arranged in order, obtains the result data ErdVal of erosion operation;
325) erosion operation terminate differentiate: in data f to be analyzed from impact point x position to ED, when total data length is less than structural element length L, terminate computing; Otherwise, go to target setting point (321).
33) noise floor estimation value obtains: utilize structural element erosion operation result data to be done to dilation operation in morphology, obtain noise floor estimation value.
Described noise floor estimation value obtains (33) step and comprises:
331) impact point in erosion operation result data is selected: if analyze erosion operation result data first, if erosion operation result data ErdVal first point is impact point x, otherwise a bit as impact point x under setting current goal point;
332) present analysis data acquisition: from impact point x, the erosion operation result data of acquisition and structural element length L same number;
333) impact point dilation operation result calculates: present analysis data and structural element S do dilation operation, obtain the dilation operation result of current goal point x;
Described impact point dilation operation result calculates (333) step and is specially:
Wherein, g represents the result ErdVal of erosion operation, and x represents the subscript of its impact point, is the integer of a non-negative; The one-dimentional structure element of S to be length be L, s
ibe wherein certain any value, i is from 0 to the integer of (L-1).
334) noise floor estimation value composition: impact point dilation operation result is arranged in order, obtains noise floor estimation value Open_Max;
335) dilation operation terminate differentiate: in erosion operation result data from aiming spot x to ED, when data length is less than structural element length L, terminate computing; Otherwise, go to target setting point (331).
40) estimated value interpolation: carry out forward and backward point-to-point transmission and afterbody interpolation to noise floor estimation value Open_Max pointwise, obtains the Interpolate estimation value of counting identical with raw spectroscopy data;
Described estimated value interpolation (40) step comprises:
41) estimated value point-to-point transmission interpolation result obtains: between former and later two points of estimated value, insert the few value of the frequency spectrum data segment length after compared with segmentation successively, obtain estimated value point-to-point transmission interpolation result;
Described estimated value point-to-point transmission interpolation result InstVMax obtains (41) step and is specially: according to the length of window K of the segmentation of former spectrum signal SpecD, (K-1) point should be inserted in 2, the front and back Open_Max (i) of estimated value and Open_Max (i+1), the value of each point is such as formula (7), and InstV (j) represents the value inserted:
InstV(j)=Open_Max(i)+j*(Open_Max(i+1)-Open_Max(i))/K （7）
Wherein, Open_Max (i) represents the estimated value of the i-th each point in estimated value, InstV (j) representative insert Open_Max (i) and Open_Max (i+1)) value, j=1 ... (K-1), K is the length of after the segmentation of raw spectroscopy data every section.
42) afterbody interpolation result obtains: utilize the value of last point of its afterbody to carry out benefit value, until its length is equal with raw spectroscopy data at the afterbody of estimated value point-to-point transmission interpolation result.
Described afterbody interpolation result obtains (42) step and is specially: the InstVMax compared after point-to-point transmission interpolation counts and the difference of the length N of former data SpecD, with the value of last point, benefit value is carried out to it at InstVMax afterbody, namely the value of adding last point is repeated at afterbody, until InstVMax length is identical with the length N of SpecD, obtain, with the afterbody interpolation result of former spectrum signal equal length, being designated as AllMax.
50) noise floor corrects: utilize that raw spectroscopy data is corresponding with the pointwise of Interpolate estimation value subtracts each other, and obtains the corrected spectrum that noise floor is smooth.
Described noise floor corrects (50) step and is specially: subtract each other with the value of former spectrum value SpecD is corresponding with the estimated value AllMax pointwise after interpolation, namely obtains the smooth frequency spectrum SpecD ' after correcting, shown in (8).
SpecD′(i)=SpecD(i)-AllMax(i)，i=1…N （8）
Fig. 5 gives the uneven pectrum noise substrate value exemplary plot that the inventive method is estimated, in figure, grey description of lines is the uneven spectrum signal of noise floor obtained from leading portion Fourier equipment, and black lines represents the noise floor variation tendency utilizing the present invention automatically to estimate.Fig. 6 gives the frequency spectrum exemplary plot after utilizing the inventive method to correct.Can find out that the method is by carrying out the noise floor of spectrum signal automatically estimating in real time by result, thus achieve the correction to non-flat frequency spectrum noise floor.
Claims (9)
1. the real-time auto-correction method of frequency spectrum non-flat forms noise floor, is characterized in that, comprise the steps:
10) frequency spectrum data obtains: the raw spectroscopy data obtaining signal amplitude spectrum or power spectrum from front end Fourier transform equipment;
20) data to be analyzed are extracted: to the isometric segmentation of raw spectroscopy data, each section is searched for one by one to the maximum of points of its amplitude or power, extracted arrangement, form data to be analyzed;
30) noise floor estimation: utilize morphologic filtering to estimate the variation tendency that data to be analyzed carry out noise floor, obtain noise floor estimation value;
40) estimated value interpolation: carry out forward and backward point-to-point transmission and afterbody interpolation to noise floor estimation value pointwise, obtains the Interpolate estimation value of counting identical with raw spectroscopy data;
50) noise floor corrects: utilize that raw spectroscopy data is corresponding with the pointwise of Interpolate estimation value subtracts each other, and obtains the corrected spectrum that noise floor is smooth;
Described noise floor estimation (30) step comprises:
31) structural element is selected: according to the feature of spectrum signal, selects one dimension platypelloid type structural element;
32) erosion operation result data obtains: utilize structural element data to be analyzed to be done to erosion operation in morphology, obtain the result data of erosion operation;
33) noise floor estimation value obtains: utilize structural element erosion operation result data to be done to dilation operation in morphology, obtain noise floor estimation value.
2. the real-time auto-correction method of frequency spectrum non-flat forms noise floor according to claim 1, is characterized in that, described data to be analyzed are extracted (20) step and comprised:
21) frequency spectrum data segmentation: select the window of certain length, this window mobile in raw spectroscopy data, is divided into some non-overlapping copies successively by raw spectroscopy data and length is equal to the frequency spectrum data section of length of window;
22) maximum of points search: search for each frequency spectrum data section one by one, find out the maximum of points of amplitude or power in each section;
23) data composition to be analyzed: the amplitude of each section or power maximum of points are extracted out, is arranged in order, obtains data to be analyzed.
3. the real-time auto-correction method of frequency spectrum non-flat forms noise floor according to claim 1, is characterized in that, described structural element selects (31) step to be specially:
Length is the platypelloid type structural element S of L,
S＝s
0,s
1,…,s
L-1(1)
Wherein, L is the natural number being less than frequency spectrum data length, s
ibe wherein certain any value, i is from 0 to the integer of (L-1), s
ivalue is any real number.
4. the real-time auto-correction method of frequency spectrum non-flat forms noise floor according to claim 1, is characterized in that, described erosion operation result data obtains (32) step and comprises:
321) impact point in data to be analyzed is selected: if analyze data to be analyzed first, if data to be analyzed first point is impact point, otherwise a bit as impact point under setting current goal point;
322) present analysis data acquisition: from impact point, the data to be analyzed of acquisition and structural element length L same number;
323) impact point erosion operation result calculates: present analysis data and structural element do erosion operation, obtain the erosion operation result of current goal point;
324) erosion operation result data composition: impact point erosion operation result is arranged in order, obtains the result data of erosion operation;
325) erosion operation terminate differentiate: in data to be analyzed from aiming spot to ED, when total data length is less than structural element length L, terminate computing; Otherwise, go to target setting point (321).
5. the real-time auto-correction method of frequency spectrum non-flat forms noise floor according to claim 4, is characterized in that, described impact point erosion operation result calculates (323) step and is specially:
Wherein, f represents data to be analyzed, and x represents its impact point subscript, is the integer of a non-negative; The one-dimentional structure element of S to be length be L, s
ibe wherein certain any value, i is from 0 to the integer of (L-1).
6. the real-time auto-correction method of frequency spectrum non-flat forms noise floor according to claim 1, is characterized in that, described noise floor estimation value obtains (33) step and comprises:
331) impact point in erosion operation result data is selected: if analyze erosion operation result data first, if erosion operation result data first point is impact point, otherwise a bit as impact point under setting current goal point;
332) present analysis data acquisition: from impact point, the erosion operation result data of acquisition and structural element length L same number;
333) impact point dilation operation result calculates: present analysis data and structural element do dilation operation, obtain the dilation operation result of current goal point;
334) noise floor estimation value composition: impact point dilation operation result is arranged in order, obtains noise floor estimation value;
335) dilation operation terminate differentiate: in erosion operation result data from aiming spot to ED, when data length is less than structural element length L, terminate computing; Otherwise, go to target setting point (331).
7. the real-time auto-correction method of frequency spectrum non-flat forms noise floor according to claim 6, is characterized in that, described impact point dilation operation result calculates (333) step and is specially:
Wherein, g represents the result of erosion operation, and x represents the subscript of its impact point, is the integer of a non-negative; The one-dimentional structure element of S to be length be L, s
ibe wherein certain any value, namely i is from 0 to the integer of (L-1).
8. the real-time auto-correction method of frequency spectrum non-flat forms noise floor according to claim 1, is characterized in that, described estimated value interpolation (40) step comprises:
41) estimated value point-to-point transmission interpolation result obtains: between former and later two points of estimated value, insert the few value of the frequency spectrum data segment length after compared with segmentation successively, obtain estimated value point-to-point transmission interpolation result;
42) afterbody interpolation result obtains: utilize the value of last point of its afterbody to carry out benefit value, until its length is equal with raw spectroscopy data at the afterbody of estimated value point-to-point transmission interpolation result.
9. the real-time auto-correction method of frequency spectrum non-flat forms noise floor according to claim 8, is characterized in that, described estimated value point-to-point transmission interpolation result obtains (41) step and is specially:
InstV(j)＝Open_Max(i)+j*(Open_Max(i+1)-Open_Max(i)/K) (4)
Wherein, Open_Max (i) represents the estimated value of the i-th each point in estimated value, the value that InstV (i) representative is inserted, j=1 ... (K-1), K is the length of after the segmentation of raw spectroscopy data every section.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410104411.5A CN103856276B (en) | 2014-03-20 | 2014-03-20 | The real-time auto-correction method of frequency spectrum non-flat forms noise floor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410104411.5A CN103856276B (en) | 2014-03-20 | 2014-03-20 | The real-time auto-correction method of frequency spectrum non-flat forms noise floor |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103856276A CN103856276A (en) | 2014-06-11 |
| CN103856276B true CN103856276B (en) | 2015-09-30 |
Family
ID=50863527
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410104411.5A Active CN103856276B (en) | 2014-03-20 | 2014-03-20 | The real-time auto-correction method of frequency spectrum non-flat forms noise floor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103856276B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108667535B (en) * | 2018-09-03 | 2020-08-11 | 电子科技大学 | Receiver bandwidth measuring method |
| CN110824248B (en) * | 2019-11-28 | 2022-02-25 | 中电科思仪科技股份有限公司 | Electromagnetic spectrum monitoring receiver signal detection template threshold processing method |
| CN112737711B (en) * | 2020-12-24 | 2023-04-18 | 成都戎星科技有限公司 | Broadband carrier detection method based on adaptive noise floor estimation |
| CN115173971B (en) * | 2022-07-08 | 2023-10-03 | 电信科学技术第五研究所有限公司 | Broadband signal real-time detection method based on frequency spectrum data |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3606522B2 (en) * | 2002-03-19 | 2005-01-05 | 日本通信機株式会社 | Frequency conversion apparatus and method |
| CN102624468A (en) * | 2011-12-30 | 2012-08-01 | 成都中安频谱科技有限公司 | Automatic broadband detection method based on dual fast Fourier transformation (FFT) |
| CN103454497A (en) * | 2013-09-10 | 2013-12-18 | 南京理工大学 | Phase difference measuring method based on improved windowing discrete Fourier transform |
-
2014
- 2014-03-20 CN CN201410104411.5A patent/CN103856276B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3606522B2 (en) * | 2002-03-19 | 2005-01-05 | 日本通信機株式会社 | Frequency conversion apparatus and method |
| CN102624468A (en) * | 2011-12-30 | 2012-08-01 | 成都中安频谱科技有限公司 | Automatic broadband detection method based on dual fast Fourier transformation (FFT) |
| CN103454497A (en) * | 2013-09-10 | 2013-12-18 | 南京理工大学 | Phase difference measuring method based on improved windowing discrete Fourier transform |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103856276A (en) | 2014-06-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103856276B (en) | The real-time auto-correction method of frequency spectrum non-flat forms noise floor | |
| CN114025379B (en) | Broadband multi-signal detection method, device and equipment | |
| CN105429719B (en) | Based on power spectrum and multi-scale wavelet transformation analysis high reject signal detection method | |
| CN105785324A (en) | MGCSTFT-based chirp signal parameter estimation method | |
| US9172428B2 (en) | Determining the spectral energy content of a data bus | |
| CN103217676A (en) | Radar target identification method under noise background based on bispectrum de-noising | |
| JP5752481B2 (en) | Passive radar device, guidance device, and radio wave detection method | |
| KR102354060B1 (en) | Method for Path Calibration in Multi-Path Radio-frequency Transceiver and Apparatus Using the Same | |
| CN108900445B (en) | Method and device for estimating signal symbol rate | |
| CN104270331A (en) | BPSK signal blind recognition result effectiveness evaluation method based on HOUGH transformation | |
| EP1865639A1 (en) | A method and an apparatus for eliminating the narrow-band interference in the spread spectrum system | |
| CN114185030A (en) | Radar signal intra-pulse modulation identification method and system | |
| CN108155959B (en) | Unmanned aerial vehicle measurement and control signal detection method under complex network environment | |
| US8175829B2 (en) | Analyzer for signal anomalies | |
| DE102017203543B4 (en) | Method for receiving and monitoring a signal and device for receiving and monitoring signals | |
| CN112929141A (en) | Unmanned aerial vehicle detection and identification method and system based on graph-borne signal matching | |
| CN104408474A (en) | Radiation source recognition method and device | |
| CN115118334B (en) | Method and device for capturing satellite communication frame header, communication equipment and storage medium | |
| WO2023155457A1 (en) | Signal processing method and apparatus, electronic device, and storage medium | |
| US9432226B2 (en) | Channel equalization method and apparatus, and receiver | |
| CN105453505A (en) | Fine timing | |
| CN110943952B (en) | Method and device for detecting amplitude modulation signal | |
| CN111835495B (en) | Method and system for detecting reference signal, readable storage medium and electronic device | |
| CN115811355B (en) | High dynamic carrier capturing method | |
| EP2647249B1 (en) | Method and device for noise floor estimation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |