CN109728822A - A kind of method, apparatus of signal processing, equipment and computer readable storage medium - Google Patents
A kind of method, apparatus of signal processing, equipment and computer readable storage medium Download PDFInfo
- Publication number
- CN109728822A CN109728822A CN201811645974.XA CN201811645974A CN109728822A CN 109728822 A CN109728822 A CN 109728822A CN 201811645974 A CN201811645974 A CN 201811645974A CN 109728822 A CN109728822 A CN 109728822A
- Authority
- CN
- China
- Prior art keywords
- signal
- matrix
- initial signal
- signal processing
- observation
- 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.)
- Pending
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a kind of methods of signal processing, including obtain initial signal, and the initial signal is the digital signal of matrix form；Sparse transformation is carried out to the initial signal, the quantity of all 0 row or all 0 column that make the initial signal increases；Observing matrix is determined according to limited equidistant property, and inputs the observing matrix as input quantity for by the initial signal of sparse transformation, obtains the observation signal.Scheme provided by the invention can a certain row or column in constraint matrix data all 0, it is equivalent to the ranks number for directly reducing the matrix of digital signal, therefore it compares and the prior art, substantially reduce the size of the observation signal finally obtained, the more conducively storage and propagation of signal, processing speed is improved simultaneously, motion artifacts are few, accelerate working efficiency.The present invention additionally provides device, equipment and the computer storage medium of a kind of signal processing with above-mentioned beneficial effect simultaneously.
Description
Technical field
The present invention relates to field of signal processing, more particularly to the method, apparatus, equipment and computer of a kind of signal processing
Readable storage medium storing program for executing.
Background technique
With accelerating development for modern digital information technology, miscellaneous smart machine constantly floods the market, such as
Phone, mobile phone, computer, camera, TV from the beginning etc., the number such as smart home, unmanned plane product, Intelligent bracelet till now
Word product produces increasing influence to our life.It may be said that the present epoch are a digitized epoch,
Acquisition, storage, exchange of all digital signals etc. require to use corresponding digital soft hardware.Modern is for image matter
Amount, video resolution etc. be multimedia to be required higher and higher, therefore corresponding hardware facility such as camera, video camera etc. is believed
Number acquisition equipment configuration requirement it is also higher and higher.Such as, it is desirable to the resolution ratio of video file is higher and higher, it is necessary to image
Built-in more sensors in machine, to acquire higher-quality vision signal.Furthermore black light such as X-ray, gamma ray etc.
The acquisition of the signals such as the acquisition of signal, the acquisition of high-speed video, traditional signal acquisition mode cannot be met the requirements.Therefore
Compressed sensing technology has obtained extensive research and application as a kind of more effectively acquisition, storage, transmission and processing method.
But in the prior art, during compressed sensing to initial signal it is sparse generally use all for L1 model
Number, L1 norm are that dispersibility is sparse, although can reduce the size of signal, cannot change the ranks of the matrix of digital signal
Number, the signal data after causing its sparse is still larger, while the distortion factor is higher.
Summary of the invention
The object of the present invention is to provide a kind of method, apparatus of signal processing, equipment and computer readable storage medium, with
It is insufficient to solve the processed signal sparsity of compressed sensing in the prior art, the excessive problem of signal data.
In order to solve the above technical problems, the present invention provides a kind of method of signal processing, comprising:
Initial signal is obtained, the initial signal is the digital signal of matrix form；
Sparse transformation is carried out to the initial signal, makes all 0 row or all 0 column of the initial signal
Quantity increases；
Observing matrix is determined according to limited equidistant property, and is inputted by the initial signal of sparse transformation as input quantity
The observing matrix obtains the observation signal.
It is optionally, described that sparse transformation is carried out to the initial signal in the method for the signal processing specifically:
Sparse transformation is carried out to the initial signal by L2,1 norm.
Optionally, in the method for the signal processing, the observing matrix is gaussian random matrix, the random square of two-value
Battle array, partial fourier matrix, local any of hadamard matrix or toeplitz matrix.
Optionally, in the method for the signal processing, observing matrix is being determined according to limited equidistant property, and will pass through
The initial signal of sparse transformation inputs the observing matrix as input quantity, after obtaining the observation signal, further includes:
By L2, the observation signal is reconstructed in 1 norm, obtains reconstruction signal.
The present invention also provides a kind of devices of signal processing, comprising:
Module is obtained, for obtaining initial signal, the initial signal is the digital signal of matrix form；
Sparse module, for the initial signal carry out sparse transformation, make the initial signal all 0 row or
The quantity of all 0 column increases；
Module is observed, for determining observing matrix according to limited equidistant property, and will be by the initial signal of sparse transformation
The observing matrix is inputted as input quantity, obtains the observation signal.
Optionally, in the device of the signal processing, the sparse module is specifically used for:
Sparse transformation is carried out to the initial signal by L2,1 norm.
Optionally, in the device of the signal processing, the observation module is specifically used for:
The observing matrix is gaussian random matrix, two-value random matrix, partial fourier matrix, local hadamard matrix
Or any of toeplitz matrix.
Optionally, in the device of the signal processing, the observation module is also used to:
By L2, the observation signal is reconstructed in 1 norm, obtains reconstruction signal.
The present invention also provides a kind of equipment of signal processing, comprising:
Memory, for storing computer program；
Processor realizes the method for signal processing as described in any one of the above embodiments when for executing the computer program
Step.
The present invention also provides be stored with meter on computer readable storage medium described in a kind of storage medium of signal processing
Calculation machine program, when the computer program is executed by processor realize it is above-mentioned it is any as described in signal processing method step
Suddenly.
The method of signal processing provided by the present invention, including initial signal is obtained, the initial signal is matrix form
Digital signal；Sparse transformation is carried out to the initial signal, makes all 0 row or all 0 of the initial signal
The quantity of column increases；Observing matrix is determined according to limited equidistant property, and will pass through the initial signal of sparse transformation as input
Amount inputs the observing matrix, obtains the observation signal.Scheme provided by the invention can a certain row or column in constraint matrix
Data all 0, be equivalent to the ranks number for directly reducing the matrix of digital signal, therefore compare and the prior art, significantly
The size of the observation signal finally obtained, the more conducively storage and propagation of signal are reduced, while improving processing speed, is moved
Artifact is few, and distorted signals is smaller.The present invention additionally provides a kind of device of signal processing with above-mentioned beneficial effect, sets simultaneously
Standby and computer storage medium.
Detailed description of the invention
It, below will be to embodiment or existing for the clearer technical solution for illustrating the embodiment of the present invention or the prior art
Attached drawing needed in technical description is briefly described, it should be apparent that, the accompanying drawings in the following description is only this hair
Bright some embodiments for those of ordinary skill in the art without creative efforts, can be with root
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is a kind of flow diagram of specific embodiment of the method for signal processing provided by the invention；
Fig. 2 is the flow diagram of another specific embodiment of the method for signal processing provided by the invention；
Fig. 3 is the flow diagram of another specific embodiment of the method for signal processing provided by the invention；
Fig. 4 is a kind of flow diagram of specific embodiment of the device of signal processing provided by the invention；
Fig. 5 is the original image that a kind of specific embodiment of the method for signal processing provided by the invention provides；
Fig. 6 is to obtain after a kind of specific embodiment of the method for signal processing provided by the invention handles original image
Image after Fourier transform；
Fig. 7 is to obtain after a kind of specific embodiment of the method for signal processing provided by the invention handles original image
Sampling matrix；
Fig. 8 is the image obtained after a kind of specific embodiment of the method for signal processing provided by the invention reconstructs.
Specific embodiment
Traditional signal acquisition method is the process of analog-to-digital conversion, i.e., will be simulated by the sensor in signal collecting device
Signal (light) is converted to digital signal (sampling of Nyquist theorem), such as the picture signal of N pixel, is then compiled by compression
The signal is transformed to the data indicated with K coefficient by code algorithm, usually K < < N.It obtained before this in conventional methods where N number of
Sampled value finally compresses it into K numerical value further through encryption algorithm, it can be seen that the digital signal tool that this method obtains
There is biggish data volume, this is highly detrimental to store and transmit.Digital signal has many redundant signals, can pass through various volumes
Code method comes into being to its further compression, compressed sensing.Perception compression, that is, directly acquire compressed data.Exist
When acquiring data, directly acquisition has the M measured value most imitated, rather than meets N number of sampled value (M of Nyquist sampling thheorem
<<N).Compressed sensing (Compressed sensing (CS)) is that one kind can restore the new of sparse signal from lack sampling measurement
Type information theory, so far, CS are widely studied and are applied.
When signal is compressible or sparse, so that it may suitable linear projection method be selected to obtain simplifying for signal
Gather (collection process, that is, reduction process), obtained data can be with undistorted or compared with low distortion place formula rebuild original number
Signal (reconstruction process), i.e. Y=Φ X.Y is exactly that compressed signal indicates, Φ indicates the calculation matrix of acquisition, can be one
Random matrix, X represent original digital signal, and collection process is exactly a linear projection process.
The design difficulty of compressed sensing is signal sparse transformation, observing matrix design, the design of restructing algorithm.
1. an important prerequisite of compressed sensing is signal or its transformation in a certain domain is sparse or compressible.
To the emphasis of rarefaction representation research first is that sparse decomposition of the signal under redundant dictionary, i.e., with super complete redundancy
Function library replaces basic function, referred to as redundant dictionary, and the element in dictionary is referred to as atom.The dictionary of selection gets over approximation signal
Structure, the result that can be more got well in subsequent arithmetic.The K item atom that optimum linear combination is selected from redundant dictionary, makees
Nonlinearity for signal approaches or sparse bayesian learning.Sparse table based on signal under redundant dictionary is shown with many research,
It is concentrated mainly on and how to construct one and be suitble to the redundant dictionary of certain class signal and how to design quickly and effectively sparse transformation algorithm.
2. the design of observing matrix
Sampling process will not change according to the variation of signal X in compressed sensing, be one for finding out Θ in given Y
Linear programming problem, but since the number of equation is far less than the number of unknown number, determining solution will not be obtained.However, if
Θ has K rank sparsity, then the problem is expected to find out determining solution.As long as at this point, trying to determine K nonzero coefficient θ i in Θ
Suitable position.Since observation vector Y is the linear combination for the K column vector that these nonzero coefficients θ i corresponds to Υ, so as to
The system of linear equations of a M × K is formed to solve the occurrence of these nonzero terms.
Limited equidistant property (RIP) gives equation, and there may be the sufficient and necessary conditions for determining solution, i.e., to making signal
It being capable of Perfect Reconstruction, it is necessary to guarantee that two different K sparse signals are not observed matrix and are mapped in the same sampling set,
The matrix that the every M column vector selected from observing matrix is constituted must assure that it is nonsingular.How by determining non-zero
The position of coefficient, and then construct M × K system of linear equations with determining solution and be the key of solving the problem.Common observation square
Battle array have random Gaussian calculation matrix, random Bernoulli Jacob's calculation matrix, part Hadamard calculation matrix, sparse random measurement matrix with
And partial orthogonality calculation matrix etc..
3. the reconstruct of signal
Restructing algorithm can all be divided into following three categories:
(1) greedy tracing algorithm: locally optimal solution is selected by iteration, continuous iteration carrys out Step wise approximation original signal.
These algorithms include matching pursuit algorithm, and orthogonal matching pursuit algorithm is segmented orthogonal matching pursuit algorithm (StOMP) and regularization
Orthogonal matching pursuit (ROMP) algorithm.
(2) convex method of relaxation: being converted to convex problem Optimization Solution for non-convex problem, so that approaching for signal is found, such as base
Tracing algorithm, GRADIENT PROJECTION METHODS, interior point method and iteration method etc..
(3) combinational algorithm: requiring the sampling of signal to support quickly to rebuild by grouping test, if Fourier samples, chain
Formula tracking etc..
For sparse decomposition algorithm angle, the match tracing Matching Pursuit algorithm based on greedy iteration thought
Great superiority is shown, but is not globally optimal solution.Donoho et al. proposes base tracking (BP) algorithm, and this method has
The advantages of global optimum, but there is very high time complexity.Occur again later a series of same based on greedy iteration think of
The innovatory algorithm thought, such as orthogonal matching pursuit algorithm (OMP), tree-like match tracing (TMP), two stage cultivation is tracked (StOMP) and is calculated
Method.All these researchs all show that the reconstruction technique based on CS can restore in (time and/or space) rate lower than Nai Kuisi
Rate required by special sampling theory.
In order to enable those skilled in the art to better understand the solution of the present invention, with reference to the accompanying drawings and detailed description
The present invention is described in further detail.Obviously, described embodiments are only a part of the embodiments of the present invention, rather than
Whole embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not making creative work premise
Under every other embodiment obtained, shall fall within the protection scope of the present invention.
Core of the invention is to provide a kind of method of signal processing, and a kind of flow diagram of specific embodiment is such as
Shown in Fig. 1, it is called specific embodiment one, comprising:
Step S101: obtaining initial signal, and above-mentioned initial signal is the digital signal of matrix form.
Step S102: sparse transformation is carried out to above-mentioned initial signal, makes all 0 row or whole of above-mentioned initial signal
Quantity for 0 column increases.
Step S103: determining observing matrix according to limited equidistant property, and using by the initial signal of sparse transformation as
Input quantity inputs above-mentioned observing matrix, obtains above-mentioned observation signal.
Further, a kind of method determining observing matrix according to limited equidistant property is provided, as shown in following equation:
Θ=ΨTX
Y=Φ Θ=Φ ΨTX
Wherein, X is raw digital signal, and Y is that the M that observation obtains is vector, and Ψ is the calculation matrix of acquisition, and Θ is sparse
Transformation matrix, A=Φ ΨT。
Further, find new accidental projection matrix to rebuild required observation number it is few as far as possible.
Further, under conditions of keeping appropriate observation number, so that new accidental projection matrix has more preferably
Property such as sparsity, come simplify the projection in reconstruction process calculate.Shown in the following formula of reconstruction signal model:
min‖ΨTX‖2,1
AX=Y
Further, above-mentioned observing matrix is gaussian random matrix, two-value random matrix, partial fourier matrix, office
Any of portion's hadamard matrix or toeplitz matrix.
The method of signal processing provided by the present invention, including initial signal is obtained, above-mentioned initial signal is matrix form
Digital signal；Sparse transformation is carried out to above-mentioned initial signal, makes all 0 row or all 0 of above-mentioned initial signal
The quantity of column increases；Observing matrix is determined according to limited equidistant property, and will pass through the initial signal of sparse transformation as input
Amount inputs above-mentioned observing matrix, obtains above-mentioned observation signal.Scheme provided by the invention can a certain row or column in constraint matrix
Data all 0, be equivalent to the ranks number for directly reducing the matrix of digital signal, therefore compare and the prior art, significantly
The size of the observation signal finally obtained, the more conducively storage and propagation of signal are reduced, while improving processing speed, is moved
Artifact is few, accelerates working efficiency.
On the basis of specific embodiment one, further the process of above-mentioned sparse transformation is limited, obtains specific reality
Mode two is applied, flow diagram is as shown in Figure 2, comprising:
Step S201: obtaining initial signal, and above-mentioned initial signal is the digital signal of matrix form.
Step S202: by L2,1 norm carries out sparse transformation to above-mentioned initial signal, makes the whole of above-mentioned initial signal
For the increase of the quantity of 0 row or all 0 column.
Step S203: determining observing matrix according to limited equidistant property, and using by the initial signal of sparse transformation as
Input quantity inputs above-mentioned observing matrix, obtains above-mentioned observation signal.
Present embodiment and above-mentioned specific embodiment the difference is that, present embodiment defines pair
Above-mentioned initial signal carries out sparse transformation by L2,1 norm, remaining step is identical as above-mentioned specific embodiment, herein not
It is reinflated to repeat.
Further, a kind of L2 is provided, the method for 1 norm transformation, as shown in following formula:
The wherein matrix that Z is n rank m times, the formula indicate Z Matrix Calculating L2 norm, then seek L1 norm have row sparse
Characteristic.
For present embodiment by L2,1 norm carries out sparse transformation, L2, and it is sparse that 1 norm itself may be implemented full line,
The square matrix of whole row or column all 0 is generated, to reduce the quantity of square matrix row or column, solves the problems, such as sparse dispersibility, and
L2,1 norm constraint problem have good convergence, can obtain only optimal solution.
On the basis of specific embodiment two, subsequent processing further is carried out to above-mentioned observation signal, obtains specific reality
Mode three is applied, flow diagram is as shown in Figure 3, comprising:
Step S301: obtaining initial signal, and above-mentioned initial signal is the digital signal of matrix form.
Step S302: by L2,1 norm carries out sparse transformation to above-mentioned initial signal, makes the whole of above-mentioned initial signal
For the increase of the quantity of 0 row or all 0 column.
Step S303: determining observing matrix according to limited equidistant property, and using by the initial signal of sparse transformation as
Input quantity inputs above-mentioned observing matrix, obtains above-mentioned observation signal.
Step S304: by L2, above-mentioned observation signal is reconstructed in 1 norm, obtains reconstruction signal.
Present embodiment and above-mentioned specific embodiment the difference is that, present embodiment is to above-mentioned sight
It surveys signal to be reconstructed, remaining step is identical as above-mentioned specific embodiment, not reinflated herein to repeat.
It should be noted that should make accidental projection matrix to rebuild required observation number it is few as far as possible；It is protecting
Under conditions of holding appropriate observation number, so that new accidental projection matrix has better properties such as sparsity, to simplify
Projection in reconstruction process calculates.
It further, is that will reconstruct the optimization problem that equation regards a belt restraining as, it is real by various optimization algorithms
The reconstruct of existing signal, most algorithm is the solution based on the convex optimization problem of L1, although based on convex optimized algorithm be effect, it
Required observation is reconstructed than theory.Therefore the present invention proposes to be based on L2, the reconstructing method of 1 norm optimization, mould
Type can be written as following form:
TV (ψ ' x) refers to the total variation penalty term of signal, λ1And λ2It is the regularization of determining data consistency and sparsity
Parameter can solve objective function using checker iteration
Present embodiment is after obtaining above-mentioned observation signal, further by L2,1 norm to above-mentioned observation signal into
Row reconstruct, obtains reconstruction signal, above-mentioned reconstruction signal and the difference of above-mentioned initial signal are the smaller the better, and present embodiment
Since by L2, above-mentioned observation signal is reconstructed in 1 norm, significantly improve the Y-PSNR of above-mentioned reconstructed image, and
Above-mentioned reconstruction signal is smaller relative to original image image distortion, and reconstruction quality is high, in addition, lack sampling, the reconstructed velocity of signal is obviously mentioned
Height reduces test period.
The composition of above-mentioned signal reconstruction can be with are as follows: using a kind of signal reconstruction method of piecemeal observation, by original signal
It is divided into the lesser block of more size, uses identical gaussian random observing matrix to signal with square during signal reconstruction
Mode observed samples, the high problem of computation complexity during signal reconstruction can be reduced.
The device of signal processing provided in an embodiment of the present invention is introduced below, the dress of signal processing described below
Reference can be corresponded to each other with the method for signal process described above by setting.
Fig. 4 is the structural block diagram of the device of signal processing provided in an embodiment of the present invention, referring to the device of Fig. 4 signal processing
May include:
Module 100 is obtained, for obtaining initial signal, above-mentioned initial signal is the digital signal of matrix form；
Sparse module 200 makes all the 0 of above-mentioned initial signal for carrying out sparse transformation to above-mentioned initial signal
It is capable or all 0 column quantity increase；
Module 300 is observed, for determining observing matrix according to limited equidistant property, and will be by the initial letter of sparse transformation
Number above-mentioned observing matrix is inputted as input quantity, obtains above-mentioned observation signal.
The device of signal processing provided by the present invention, including initial signal is obtained by obtaining module 100, it is above-mentioned initial
Signal is the digital signal of matrix form；Sparse transformation is carried out to above-mentioned initial signal by sparse module 200, is made above-mentioned initial
All 0 row of signal or the quantity of all 0 column increase；It is determined by observation module 300 according to limited equidistant property
Observing matrix, and above-mentioned observing matrix is inputted as input quantity using by the initial signal of sparse transformation, obtain above-mentioned observation letter
Number.Scheme provided by the invention can a certain row or column in constraint matrix data all 0, being equivalent to directly reduces number
The ranks number of the matrix of signal, therefore compare and the prior art, the size of the observation signal finally obtained is substantially reduced, it is more sharp
In the storage and propagation of signal, while processing speed is improved, motion artifacts are few, accelerate working efficiency.
Method of the device of the signal processing of the present embodiment for realizing signal processing above-mentioned, therefore the dress of signal processing
The embodiment part of the method for the visible signal processing hereinbefore of specific embodiment in setting, for example, module 100 is obtained, it is dilute
Module 200 is dredged, module 300 is observed, is respectively used to step S101, S102 and S103 in the method for realizing above-mentioned signal processing, institute
With specific embodiment is referred to the description of corresponding various pieces embodiment, and details are not described herein.
Below as an example, the effect for verifying signal processing method provided by the invention, as shown in Fig. 5 to Fig. 8:
Fig. 5 is the original image that a kind of specific embodiment of the method for signal processing provided by the invention provides；
Fig. 6 is to obtain after a kind of specific embodiment of the method for signal processing provided by the invention handles original image
Image after Fourier transform；
Fig. 7 is to obtain after a kind of specific embodiment of the method for signal processing provided by the invention handles original image
Sampling matrix；
Fig. 8 is the image obtained after a kind of specific embodiment of the method for signal processing provided by the invention reconstructs.
Wherein mean square deviation errorx=2.556e-23；
Y-PSNR psnr=3.222e+02.
The present invention also provides a kind of equipment, in above equipment, each component part, which can be shared out the work and helped one another, executes any of the above-described implementation
The method of signal processing described in mode.Remaining content can refer to the prior art, no longer carry out expansion description herein.
Invention additionally provides a kind of computer readable storage medium, it is stored on above-mentioned computer readable storage medium
Computer program, above-mentioned computer program realize signal described in any of the above-described invention embodiment when being executed by processor
The method of processing.Remaining content is referred to the prior art, no longer carries out expansion description herein.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with it is other
The difference of embodiment, same or similar part may refer to each other between each embodiment.For being filled disclosed in embodiment
For setting, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is referring to method part
Explanation.
It should be noted that in the present specification, relational terms such as first and second and the like are used merely to one
A entity or operation with another entity or operate distinguish, without necessarily requiring or implying these entities or operation it
Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant are intended to
Cover non-exclusive inclusion, so that the process, method, article or equipment for including a series of elements not only includes those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or setting
Standby intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in the process, method, article or equipment for including above-mentioned element.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think beyond the scope of this invention.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
The method, apparatus, equipment and computer readable storage medium of signal processing provided by the present invention are carried out above
It is discussed in detail.Used herein a specific example illustrates the principle and implementation of the invention, above embodiments
Explanation be merely used to help understand method and its core concept of the invention.It should be pointed out that for the common of the art
, without departing from the principle of the present invention, can be with several improvements and modifications are made to the present invention for technical staff, these
Improvement and modification are also fallen within the protection scope of the claims of the present invention.
Claims (10)
1. a kind of method of signal processing characterized by comprising
Initial signal is obtained, the initial signal is the digital signal of matrix form；
Sparse transformation is carried out to the initial signal, makes all 0 row of the initial signal or the quantity of all 0 column
Increase；
Observing matrix is determined according to limited equidistant property, and will pass through the initial signal of sparse transformation as described in input quantity input
Observing matrix obtains the observation signal.
2. the method for signal processing as described in claim 1, which is characterized in that described to carry out sparse change to the initial signal
It changes specifically:
Sparse transformation is carried out to the initial signal by L2,1 norm.
3. signal processing method as described in claim 1, which is characterized in that the observing matrix is gaussian random matrix, two
It is worth random matrix, partial fourier matrix, local any of hadamard matrix or toeplitz matrix.
4. signal processing method as described in claim 1, which is characterized in that determining observation square according to limited equidistant property
Battle array, and the observing matrix is inputted as input quantity using by the initial signal of sparse transformation, after obtaining the observation signal,
Further include:
By L2, the observation signal is reconstructed in 1 norm, obtains reconstruction signal.
5. a kind of device of signal processing characterized by comprising
Module is obtained, for obtaining initial signal, the initial signal is the digital signal of matrix form；
Sparse module makes all 0 row or whole of the initial signal for carrying out sparse transformation to the initial signal
Quantity for 0 column increases；
Observe module, for determining observing matrix according to limited equidistant property, and using by the initial signal of sparse transformation as
Input quantity inputs the observing matrix, obtains the observation signal.
6. the device of signal processing as claimed in claim 5, which is characterized in that the sparse module is specifically used for:
Sparse transformation is carried out to the initial signal by L2,1 norm.
7. the device of signal processing as claimed in claim 5, which is characterized in that the observation module is specifically used for:
The observing matrix is gaussian random matrix, two-value random matrix, partial fourier matrix, local hadamard matrix or support
Puli's hereby any of matrix.
8. the device of signal processing as claimed in claim 5, which is characterized in that the observation module is also used to:
By L2, the observation signal is reconstructed in 1 norm, obtains reconstruction signal.
9. a kind of equipment of signal processing characterized by comprising
Memory, for storing computer program；
Processor realizes the side such as the described in any item signal processings of Claims 1-4 when for executing the computer program
The step of method.
10. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium
Program realizes the method such as the described in any item signal processings of Claims 1-4 when the computer program is executed by processor
The step of.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811645974.XA CN109728822A (en) | 2018-12-29 | 2018-12-29 | A kind of method, apparatus of signal processing, equipment and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811645974.XA CN109728822A (en) | 2018-12-29 | 2018-12-29 | A kind of method, apparatus of signal processing, equipment and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109728822A true CN109728822A (en) | 2019-05-07 |
Family
ID=66299383
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811645974.XA Pending CN109728822A (en) | 2018-12-29 | 2018-12-29 | A kind of method, apparatus of signal processing, equipment and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109728822A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111751658A (en) * | 2020-06-24 | 2020-10-09 | 国家电网有限公司大数据中心 | Signal processing method and device |
| CN111953382A (en) * | 2020-08-13 | 2020-11-17 | 广东石油化工学院 | PLC signal reconstruction method and system by utilizing segmented sparsity |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104077791A (en) * | 2014-05-22 | 2014-10-01 | 南京信息工程大学 | Joint reconstruction method for multiple dynamic contrast enhancement nuclear magnetic resonance images |
| CN104242946A (en) * | 2014-07-03 | 2014-12-24 | 河海大学 | Signal reconstruction method of photovoltaic array state monitoring network |
| CN104586394A (en) * | 2014-12-23 | 2015-05-06 | 中国科学院深圳先进技术研究院 | Method and system for removing magnetic resonance diffusion tensor imaging noise |
| CN106137167A (en) * | 2016-07-21 | 2016-11-23 | 浙江师范大学 | A kind of motion artifacts detection method based on photoplethysmographic signal |
-
2018
- 2018-12-29 CN CN201811645974.XA patent/CN109728822A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104077791A (en) * | 2014-05-22 | 2014-10-01 | 南京信息工程大学 | Joint reconstruction method for multiple dynamic contrast enhancement nuclear magnetic resonance images |
| CN104242946A (en) * | 2014-07-03 | 2014-12-24 | 河海大学 | Signal reconstruction method of photovoltaic array state monitoring network |
| CN104586394A (en) * | 2014-12-23 | 2015-05-06 | 中国科学院深圳先进技术研究院 | Method and system for removing magnetic resonance diffusion tensor imaging noise |
| CN106137167A (en) * | 2016-07-21 | 2016-11-23 | 浙江师范大学 | A kind of motion artifacts detection method based on photoplethysmographic signal |
Non-Patent Citations (2)
| Title |
|---|
| JIE CHEN ,XIAOMING HUO: "SPARSE REPRESENTATIONS FOR MULTIPLE MEASUREMENT VECTORS (MMV) IN AN", 《IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, 2005》 * |
| 斯塔克等: "《稀疏图像与信号处理：小波，曲波，形态多元性》", 31 May 2015 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111751658A (en) * | 2020-06-24 | 2020-10-09 | 国家电网有限公司大数据中心 | Signal processing method and device |
| CN111953382A (en) * | 2020-08-13 | 2020-11-17 | 广东石油化工学院 | PLC signal reconstruction method and system by utilizing segmented sparsity |
| CN111953382B (en) * | 2020-08-13 | 2021-06-11 | 广东石油化工学院 | PLC signal reconstruction method and system by utilizing segmented sparsity |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108735279B (en) | Virtual reality upper limb rehabilitation training system for stroke in brain and control method | |
| CN104268934B (en) | Method for reconstructing three-dimensional curve face through point cloud | |
| Liu et al. | Gaussian mixture model based volume visualization | |
| Zheng et al. | Cascaded dilated dense network with two-step data consistency for MRI reconstruction | |
| CN105046672B (en) | A kind of image super-resolution rebuilding method | |
| Bartelmann | Cluster mass estimates from weak lensing | |
| Wohlberg | Convolutional sparse representation of color images | |
| CN109584319A (en) | A kind of compression of images sensing reconstructing algorithm based on non-local low rank and full variation | |
| CN109932816A (en) | Super memory effect range non-intrusion type based on connected domain optimization scatters imaging method | |
| CN111160298A (en) | Robot and pose estimation method and device thereof | |
| CN109887050A (en) | A kind of code aperture spectrum imaging method based on self-adapting dictionary study | |
| CN112017228A (en) | Method for three-dimensional reconstruction of object and related equipment | |
| CN108230280A (en) | Image speckle noise minimizing technology based on tensor model and compressive sensing theory | |
| CN103955904A (en) | Method for reconstructing signal based on dispersed fractional order Fourier transform phase information | |
| CN109728822A (en) | A kind of method, apparatus of signal processing, equipment and computer readable storage medium | |
| Xie et al. | Artifact removal using GAN network for limited-angle CT reconstruction | |
| Yang et al. | Reconstruction of structurally-incomplete matrices with reweighted low-rank and sparsity priors | |
| Mondal et al. | FPGA based accelerated 3D affine transform for real-time image processing applications | |
| CN109559278B (en) | Super resolution image reconstruction method and system based on multiple features study | |
| RU2753591C1 (en) | Method for compression and storage of three-dimensional data (variants) | |
| CN106296583B (en) | Based on image block group sparse coding and the noisy high spectrum image ultra-resolution ratio reconstructing method that in pairs maps | |
| CN117217997A (en) | Remote sensing image super-resolution method based on context perception edge enhancement | |
| CN116612009A (en) | Multi-scale connection generation countermeasure network medical image super-resolution reconstruction method | |
| CN108801457B (en) | Three-dimensional map acquisition and reconstruction method based on coding sampling plate design and secondary energy constraint correction | |
| Cho et al. | Example-based super-resolution using self-patches and approximated constrained least squares filter |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190507 |