CN109360264A - The method for building up and device of image unified model - Google Patents
The method for building up and device of image unified model Download PDFInfo
- Publication number
- CN109360264A CN109360264A CN201811003368.8A CN201811003368A CN109360264A CN 109360264 A CN109360264 A CN 109360264A CN 201811003368 A CN201811003368 A CN 201811003368A CN 109360264 A CN109360264 A CN 109360264A
- Authority
- CN
- China
- Prior art keywords
- value
- gradient
- spectral value
- point
- spectral
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10048—Infrared image
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
- Y02A40/10—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in agriculture
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses the method for building up and device of a kind of image unified model, method comprises determining that spectral value of the high spectrum image on the curve of spectrum of designated flying zone point, and determines the gradient information of the spectral value；According to the spectral value and gradient information, the comprehensive characteristics parameter of high spectrum image is determined, and establish the unified model of high spectrum image based on the comprehensive characteristics parameter, and local feature detection is carried out to high spectrum image using unified model.Compared to existing technologies, the present invention is when carrying out local feature detection to high spectrum image, the comprehensive characteristics parameter including spectral value gradient information corresponding with spectral value is introduced, in the local feature detection for carrying out high spectrum image, facilitates the robustness for promoting whole system.
Description
Technical field
The present invention relates to technical field of image processing more particularly to a kind of method for building up and device of image unified model.
Background technique
High spectrum image is suffered from many fields and is widely applied in recent years, different from gray level image and color image
, high-spectrum seems to be folded by the image of multiple wave bands such as visible light, near-infrared, short-wave infrared, medium-wave infrared, thermal infrared
Add, therefore, high spectrum image can provide more information than gray level image and color image.
High spectrum image can indicate that the data in vector field are by the same coordinate position with trivector field
Different-waveband pixel composition vector.Operation and analysis for multidimensional vector data, traditional multidimensional theorem in Euclid space is
Through complete theory support can not be provided for it, it is difficult to obtain the key message of more high spectrum images, therefore high spectrum image
Robustness in local feature detection process is lower.
Summary of the invention
The present invention provides the method for building up and device of a kind of image unified model, can solve bloom in the prior art
The lower technical problem of robustness of the spectrogram picture in local feature detection process.
First aspect present invention provides a kind of method for building up of image unified model, this method comprises:
It determines spectral value of the high spectrum image on the curve of spectrum of designated flying zone point, and determines the ladder of the spectral value
Spend information；
According to the spectral value and the gradient information, the comprehensive characteristics parameter of the high spectrum image is determined；
The unified model of the high spectrum image is established based on the comprehensive characteristics parameter, and utilizes the unified model
Local feature detection is carried out to the high spectrum image.
Optionally, described according to the spectral value and the gradient information, determine the comprehensive characteristics of the high spectrum image
The step of parameter includes:
According to the spectral value and the gradient information, the spectrum change of gradient vector modulus value of the airspace point is calculated；
The spectral value of the airspace point is modified, the revised amendment spectral value of airspace point is obtained；
Based on the spectrum change of gradient vector modulus value and the amendment spectral value, the comprehensive characteristics parameter is determined.
Optionally, described according to the spectral value and the gradient information, calculate the spectrum change of gradient of the airspace point
The step of vector modulus value includes:
Based on the gradient information of the airspace point, the gradient value of the airspace point is calculated；
Using calculated gradient value, each point in the field of the default size centered on the airspace point is calculated
Average gradient modulus value；
The spectrum change of gradient vector modulus value of the airspace point is determined based on the average gradient modulus value.
Optionally, the spectral value to the airspace point is modified, and obtains the revised amendment light of airspace point
The step of spectrum includes:
Using pre-set spectral value correction amount computational algorithm, the corresponding correction value of the spectral value is calculated；
When the correction value is less than preset correction threshold, the spectral value is modified using the correction value.
Optionally, the step of gradient information of the determination spectral value includes:
Determine the spectral value on the direction of airspace with the change rate on spectral domain direction；
The gradient information of the spectral value is calculated based on the change rate.
What second aspect of the present invention provided a kind of image unified model establishes device, which includes:
First determining module, for determining spectral value of the high spectrum image on the curve of spectrum of designated flying zone point, and really
The gradient information of the fixed spectral value；
Second determining module, for determining the comprehensive of the high spectrum image according to the spectral value and the gradient information
Close characteristic parameter；
Module is established, for establishing the unified model of the high spectrum image based on the comprehensive characteristics parameter, and is utilized
The unified model carries out local feature detection to the high spectrum image.
Optionally, second determining module includes:
Computing module, for according to the spectral value and the gradient information, the spectrum gradient for calculating the airspace point to become
Change vector modulus value；
Correction module is modified for the spectral value to the airspace point, obtains the revised amendment of the airspace point
Spectral value；
Parameter determination module, for determining institute based on the spectrum change of gradient vector modulus value and the amendment spectral value
State comprehensive characteristics parameter.
Optionally, the computing module includes:
First computing module calculates the gradient value of the airspace point for the gradient information based on the airspace point；
Second computing module calculates the default size centered on the airspace point for utilizing calculated gradient value
Field in each point average gradient modulus value；
Third computing module, for determining that the spectrum change of gradient of the airspace point is sweared based on the average gradient modulus value
Measure modulus value.
Optionally, the correction module is specifically used for:
Using pre-set spectral value correction amount computational algorithm, the corresponding correction value of the spectral value is calculated, when described
When correction value is less than preset correction threshold, the spectral value is modified using the correction value.
Optionally, first determining module includes:
Change rate determining module, for determine the spectral value on the direction of airspace with the change rate on spectral domain direction；
The gradient information of the spectral value is calculated based on the change rate for gradient information computing module.
The present invention provides a kind of method for building up of image unified model, this method comprises: determining that high spectrum image is referring to
Determine the spectral value on the curve of spectrum of airspace point, and determines the gradient information of the spectral value；According to above-mentioned spectral value and the ladder
Information is spent, the comprehensive characteristics parameter of high spectrum image is determined, the unified mould of high spectrum image is established based on the comprehensive characteristics parameter
Type, and local feature detection is carried out to high spectrum image using unified model.Compared to existing technologies, the present invention is right
When high spectrum image carries out local feature detection, introduce comprehensive special including spectral value gradient information corresponding with spectral value
Levy parameter, then establish the unified model of high spectrum image, due in high spectrum image only comprising same substance point light
Spectral curve, it is very small in the variation of spectral domain, and include the curve of spectrum of the point of a variety of different materials, spectral domain variation but very
Greatly, therefore by above-mentioned unified model, the substance kind that any point is included in high spectrum image can be effectively detected
Class number facilitates the robustness for promoting whole system that is, in the local feature detection for carrying out high spectrum image.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, below will to embodiment or
Attached drawing needed to be used in the description of the prior art is briefly described, it should be apparent that, the accompanying drawings in the following description is only
Some embodiments of the present invention for those skilled in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is the step flow diagram of the method for building up of image unified model in the embodiment of the present invention；
Fig. 2 is the refinement step flow diagram of step 102 in the embodiment of the present invention；
Fig. 3 is the program module schematic diagram for establishing device of image unified model in the embodiment of the present invention；
Fig. 4 is the elaborator module diagram of the second determining module 302 in the embodiment of the present invention；
Fig. 5 is the structural schematic diagram of the electronic equipment provided in the embodiment of the present invention.
Specific embodiment
In order to make the invention's purpose, features and advantages of the invention more obvious and easy to understand, below in conjunction with this hair
Attached drawing in bright embodiment, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described
Embodiment is only a part of the embodiment of the present invention, and not all embodiments.Based on the embodiments of the present invention, this field skill
Art personnel every other embodiment obtained without making creative work belongs to the model that the present invention protects
It encloses.
In order to better understand the present invention, the embodiment of the present invention provides a kind of Geometrical algebra mould of high spectrum image first
Type, the local feature detection process in the present invention carry out in the high spectrum image of a limited long-wave band, therefore EO-1 hyperion
Image can be expressed as a high spectrum image cube, including spatial information (si) (x, y) and spectrum domain information λ, i.e., for one
Airspace is the high spectrum image F of n having a size of M*N, high spectrum image wave band number, it can be indicated are as follows:
F=f (x, y, λ) (1)
In formula (1), f (x, y, λ) indicates that the function of high spectrum image, (x, y, λ) indicate that 3 dimension coordinates, x and y indicate airspace
Coordinate, 0 < x < M, 0 < y < N, λ indicate spectral domain coordinate, 0 < λ < n.
Therefore the mathematical framework that the present embodiment is indicated and analyzed as high spectrum image using Geometrical algebra is first situated between below
The mathematical notation model of the high spectrum image to continue under Geometrical algebra frame.
If R3It is the three-dimensional theorem in Euclid space in high spectrum image airspace and spectral domain composition, its orthonormal basis is { e1,
e2,e3, then these orthonormal basises are by geometry product at R3On Geometrical algebra space beAndIt can be with
Referred to as the three-dimensional geometry algebraic space of high spectrum image, the present embodiment further part are abbreviated asIts one group of specification base
Are as follows:
E3:={ Ei| i=0,1,2 ..., 23- 1 }={ 1, e1,e2,e3,e1∧e2,e2∧e3,e1∧e3,e1∧e2∧e3} (2)
Wherein, ∧ is Geometrical algebra apposition, e1∧e2,e2∧e3,e1∧e3It is by three orthogonal basis e1,e2,e3It obtains
Three independent double appositions are shown respectively on these three double apposition geometric meaningsThe vector representation of two, space is put down
Face, e1∧e2∧e3It is triple appositions: e1∧e2∧e3=(e1∧e2)e3, geometric interpretation is: double apposition e1∧e2Along arrow
Measure e3Mobile oriented solid obtained.{e1,e2,e3Be considered as3 dimensional vector subspaces base vector x, y,
λ}。
By e1e2e3It is denoted as I, due to eiej=ei∧ej=eij,ei 2=1, then e1e2=Ie3, e2e3=Ie1And e3e1=
Ie2.And meet
(e1e2)2=(e2e3)2=(e3e1)2=-1 (3)
Then high spectrum image can indicate are as follows:
F=f (p) (4)
WhereinP=xe1+ye2+λe3, indicate the pixel of high spectrum image F.
If coordinate is (x in high spectrum image F1,y1,λ1) and (x2,y2,λ2) two pixel p1,p2,And
p1=x1e1+y1e2+λ1e3, p2=x2e1+y2e2+λ2e3, then their geometry product can indicate are as follows:
p1p2=p1·p2+p1∧p2 (5)
It indicates that the geometry product of two vectors is by inner product (p1·p2) and apposition (p1∧p2) the sum of composition.
?In, p1And p2Distance can be indicated with Δ p, it may be assumed that
Δ p=p1-p2=(x1-x2)e1+(y1-y2)e2+(λ1-λ2)e3 (6)
It indicate one from p2It is directed toward p1Vector, it is not only the measurement of two pixel distances, and can be anti-
Reflect the gradient information in high spectrum image.
Referring to Fig. 1, Fig. 1 is the step flow diagram of the method for building up of image unified model in the embodiment of the present invention,
In the embodiment of the present invention, the above method includes:
Step 101 determines spectral value of the high spectrum image on the curve of spectrum of designated flying zone point, and determines the light
The gradient information of spectrum.
Wherein, the curve of spectrum refers to spectrum in the corresponding entire spectral domain in the certain point (x, y) in high spectrum image in airspace
Value is formed by a curve.For the only curve of spectrum of the point including same substance in airspace, their variations in spectral domain
It is very close, but includes the curve of spectrum of the point of different material, they is but very big in the variation of spectral domain.Therefore I
Will be effective by the fluctuation situation of spectral value and the change rate of the curve of spectrum to complete object detection and recognition
A kind of method.It is calculated in airspace by the change rate of the curve of spectrum to difference, can more effectively search out object
The important information hidden in matter.
Specifically, the step of determining the gradient information of the spectral value in above-mentioned steps 101 includes:
Step a, determine the spectral value on the direction of airspace with the change rate on spectral domain direction；
Step b, the gradient information of the spectral value is calculated based on the change rate.
Specifically, under Geometrical algebra space high spectrum image gradient definition:
Assuming thatAnd p=xe1+ye2+λe3, then HSI (hyperspectral image, high spectrum image) model can
It is defined herein to be expressed as F=f (p):
For the gradient of high spectrum image F P point in Geometrical algebra space, it is denoted as gradf (p), i.e.,
Gradf (p)=fx(p)e1+fy(p)e2+fλ(p)e3 (8)
Wherein fx(p) and fyIt (p) is the local derviation in the direction x and y on airspace, fλIt (p) is the local derviation on spectral domain direction.EO-1 hyperion
Gradient (f on image airspacex(p)e1+fy(p)e2) reflect the marginal information of airspace two dimensional image, and the gradient on spectral domain
(fλ(p)e3) then reflect material information specific to different material.It is emerging in order to detect the empty spectral domain with different classes of property
Interesting point (SSIP, spatial-spectral interest point) will be to probe into emphasis with the gradient on spectral domain, to obtain
SSIP with different material attribute.
Step 102, according to the spectral value and the gradient information, determine the comprehensive characteristics ginseng of the high spectrum image
Number.
Wherein, under normal circumstances, in high spectrum image only comprising same substance point the curve of spectrum have it is identical or
The similar changing rule of person, therefore the variable gradient of their spectral value and the curve of spectrum will be very in certain neighborhood
Close, therefore, integrated spectral value and gradient information determine the comprehensive characteristics parameter of high spectrum image, can be used as EO-1 hyperion
A preferably parameter in the extraction of image local feature.
Specifically, Fig. 2 is the refinement step flow diagram of step 102 in the embodiment of the present invention, above-mentioned step referring to Fig. 2
Rapid 102 can be subdivided into following steps:
Step 201, according to the spectral value and the gradient information, calculate the spectrum change of gradient arrow of the airspace point
Measure modulus value.
Wherein, above-mentioned steps 201 include:
Step a, the gradient information based on the airspace point, calculates the gradient value of the airspace point；
Step b, it using calculated gradient value, calculates each in the field of the default size centered on the airspace point
The average gradient modulus value of a point；
Step c, the spectrum change of gradient vector modulus value of the airspace point is determined based on the average gradient modulus value.
Specifically, it is assumed that
Then: p0=xie1+yje2+λke3, p1=xie1+yje2+(λk+1)e3, p2=xie1+yje2+(λk-1)e3, p3=xie1
+yje2+(λk+2)e3, p4=xie1+yje2+(λk-2)e3。Middle p0The spectrum gradient at placeIs defined as:
Wherein,
Wherein,Reflect point p0Its amplitude of variation is reflected in the gradient magnitude at place and direction using its modulus value,
And its direction can be defined as the curve of spectrum in wave band λkMidpoint p0Tangential direction.In general, in the gradient-norm of the point
Value is bigger, then the curve of spectrum is bigger in the variable quantity of the point, and vice versa.Due to same substance same airspace not
There is the identical or similar curve of spectrum of variation with spectral domain, and the variation of the curve of spectrum is spectrum gradient, if in l × l neighborhood
Interior is same substance, then they have the identical or similar curve of spectrum of variation, and their gradient magnitudes in this neighborhood
Also very close to.
If Si(i=0,1) it indicates in λi=λkWith p in+i planeiCentered on l × l the total l of neighborhood2The set of a point,
S2It indicates in λ2=λkWith p in -1 plane2Centered on l × l the total l of neighborhood2The set of a point, wherein λk∈ [2, n-1], obtains
It arrives:
Wherein gpijForUpper S0Interior any point pijSpectrum gradient.S0Interior average gradient modulus value are as follows:
If in point p0L × l-1 neighborhood in be same substance, then its variance
And
Wherein ε is empirical value, is the threshold value of unit gradient variance.For convenience and simplify subsequent calculating, defines one
Change of gradient vectorAssuming thatIt is spectrum gradient in p0Standard deviation, then the definition of spectral value change of gradient vector modulus value
It is as follows:
Step 202 is modified the spectral value of the airspace point, obtains the revised amendment spectrum of airspace point
Value.
Wherein, above-mentioned steps 202 include:
Step a, using pre-set spectral value correction amount computational algorithm, the corresponding correction value of the spectral value is calculated；
Step b, when the correction value be less than preset correction threshold when, using the correction value to the spectral value into
Row amendment.
Specifically, the standard deviation of spectrum gradient is spectrum change of gradient vectorModulus value,Direction be spectrum ladder
DegreeDirection.On the one hand, due to external environment, the shadow of the noises such as hyperspectral image data acquisition, transport, pretreatment
It rings, the possible curve of spectrum of same substance can have differences, and need to be modified spectral value thus, " different with composing with reduction
Influence when object " is to feature extraction and expression；On the other hand, the curve of spectrum is influenced by change rate (i.e. gradient) and spectral value,
Therefore, for the same substance in HSI, other than gradient, the corrected value for calculating the spectral value on the curve of spectrum is also needed,
It is defined as follows:
Wherein f (p0) it is p0The spectral value at place, fδ(p0) it is revised spectral value,It is fixed for the correction amount of spectral value
Justice forCentered on 3 σ neighborhoods standard deviation, i.e.,
Wherein σ is Gauss scale factor,It is with p0Centered on 3 σ neighborhoods mean value, and
Wherein, ε1It is the standard deviation threshold method of unit spectral value correction amount for empirical value.
Step 203 is based on the spectrum change of gradient vector modulus value and the amendment spectral value, determines described comprehensive special
Levy parameter.
Wherein, for same substance, gradient vectorIt must simultaneously meet (12)-(19) with revised spectral value
Formula is located at then for performancePixel possessed by spectrum value information and change of gradient information containing correction amount,
Define a new Geometrical algebra vector f ' (p0), it is named as spectral value-change of gradient vector of high spectrum image pixel
(spectral value-gradient change vector, SVGCV), i.e., signified comprehensive characteristics ginseng in above-mentioned steps
Number, form are as follows:
Wherein fδ(p0) it is p after amendment0The spectral value at place,For change of gradient vector.It is one and both believed containing scalar
Breath and the vector containing Vector Message, not only reflect spectrum value information, but also reflect gradient direction and its size variation
Situation.
Step 103, the unified model that the high spectrum image is established based on the comprehensive characteristics parameter, and described in utilization
Unified model carries out local feature detection to the high spectrum image.
In the embodiment of the present invention, based on above-mentioned definition and describe, high spectrum image F is transformed to one to be joined with comprehensive characteristics
Number SVGCV is the three-dimensional matrice of element, which can be used as the unified model of high spectrum image, be denoted as:
F '=f ' (p) (21)
It is p that wherein f ' (p), which is independent variable, and dependent variable is the function of SVGCV.
Wherein, local feature detection can be carried out to above-mentioned high spectrum image using above-mentioned unified model.
A kind of method for building up of image unified model provided by the embodiment of the present invention, comprising: determine that high spectrum image exists
Spectral value on the curve of spectrum of designated flying zone point, and determine the gradient information of the spectral value；According to above-mentioned spectral value with it is described
Gradient information determines the comprehensive characteristics parameter of high spectrum image, and the unification of high spectrum image is established based on the comprehensive characteristics parameter
Model, and local feature detection is carried out to high spectrum image using unified model.Compared to existing technologies, the present invention exists
When carrying out local feature detection to high spectrum image, the synthesis including spectral value gradient information corresponding with spectral value is introduced
Then characteristic parameter establishes the unified model of high spectrum image, due to only including the point of same substance in high spectrum image
The curve of spectrum, it is very small in the variation of spectral domain, and include the curve of spectrum of the point of a variety of different materials, spectral domain variation but
It is very big, therefore by above-mentioned unified model, it can effectively detect the substance that any point is included in high spectrum image
Species number facilitates the robustness for promoting whole system that is, in the local feature detection for carrying out high spectrum image.
Further, the embodiment of the present invention also provides a kind of device of establishing of image unified model, is referring to Fig. 3, Fig. 3
The program module schematic diagram for establishing device of image unified model in the embodiment of the present invention, in the embodiment of the present invention, above-mentioned apparatus
Include:
First determining module 301, for determining spectral value of the high spectrum image on the curve of spectrum of designated flying zone point,
And determine the gradient information of the spectral value.
Second determining module 302, for determining the high spectrum image according to the spectral value and the gradient information
Comprehensive characteristics parameter.
Module 303 is established, for establishing the unified model of the high spectrum image based on the comprehensive characteristics parameter, and
Local feature detection is carried out to the high spectrum image using the unified model.
Wherein, the first determining module 301 includes:
Change rate determining module, for determine the spectral value on the direction of airspace with the change rate on spectral domain direction；
The gradient information of the spectral value is calculated based on the change rate for gradient information computing module.
It further, is that the elaborator module of the second determining module 302 in the embodiment of the present invention is shown referring to Fig. 4, Fig. 4
It is intended to, above-mentioned second determining module 302 includes:
Computing module 401, for calculating the spectrum ladder of the airspace point according to the spectral value and the gradient information
Spend diverse vector modulus value.
Wherein, computing module 401 includes:
First computing module calculates the gradient value of the airspace point for the gradient information based on the airspace point；
Second computing module calculates the default size centered on the airspace point for utilizing calculated gradient value
Field in each point average gradient modulus value；
Third computing module, for determining that the spectrum change of gradient of the airspace point is sweared based on the average gradient modulus value
Measure modulus value.
Correction module 402 is modified for the spectral value to the airspace point, and it is revised to obtain the airspace point
Correct spectral value.
Wherein, correction module 402 is specifically used for:
Using pre-set spectral value correction amount computational algorithm, the corresponding correction value of the spectral value is calculated, when described
When correction value is less than preset correction threshold, the spectral value is modified using the correction value.
Parameter determination module 403, for being based on the spectrum change of gradient vector modulus value and the amendment spectral value, really
The fixed comprehensive characteristics parameter.
A kind of image unified model provided by the embodiment of the present invention establishes device, may be implemented: determining high-spectrum
As the spectral value on the curve of spectrum of designated flying zone point, and determine the gradient information of the spectral value；According to above-mentioned spectral value with
The gradient information determines the comprehensive characteristics parameter of high spectrum image, establishes high spectrum image based on the comprehensive characteristics parameter
Unified model, and local feature detection is carried out to high spectrum image using unified model.Compared to existing technologies, this hair
It is bright to high spectrum image carry out local feature detection when, introduce including spectral value gradient information corresponding with spectral value
Then comprehensive characteristics parameter establishes the unified model of high spectrum image, due to only including same substance in high spectrum image
The curve of spectrum of point, it is very small in the variation of spectral domain, and include the curve of spectrum of the point of a variety of different materials, in the change of spectral domain
Change is but very big, therefore by above-mentioned unified model, can effectively detect that any point is included in high spectrum image
Substance classes number facilitates the robustness for promoting whole system that is, in the local feature detection for carrying out high spectrum image.
Further, the embodiment of the present invention also provides a kind of electronic equipment, including memory, processor and being stored in is deposited
On reservoir and the computer program that can run on a processor, when processor executes computer program, realize that image of the present invention is united
The method for building up of one model corresponds to each step in each embodiment.
The embodiment of the present invention also provides a kind of readable storage medium storing program for executing, which is computer-readable storage medium
Matter is stored thereon with computer program, when computer program is executed by processor, realizes building for image unified model of the present invention
Cube method corresponds to each step in each embodiment.
It in order to better understand the present invention, is the structure of the electronic equipment provided in the embodiment of the present invention referring to Fig. 5, Fig. 5
Schematic diagram.As shown in figure 5, the electronic equipment 05 of the embodiment specifically includes that processor 50, memory 51 and is stored in
In reservoir 51 and the computer program 52 that can run on processor 50, such as image unified model establishes program.Processor
The step in each embodiment of method for building up of above-mentioned image unified model is realized when 50 execution computer program 52, such as Fig. 1 is extremely
Step shown in 2 any example of attached drawing.Alternatively, processor 50 realizes above-mentioned each Installation practice when executing computer program 52
In each module/unit function, such as the function of each module shown in Fig. 3.
Computer program 52 can be divided into one or more module/units, one or more module/unit quilt
It is stored in memory 51, and is executed by processor 50, to complete the present invention.One or more module/units can be can
The series of computation machine program instruction section of specific function is completed, which is calculating equipment for describing computer program 52
Implementation procedure in 05.For example, computer program 52 can be divided into the first determining module 301, the second determining module 302,
The function of detection module 303 (module in virtual bench).
Calculating equipment 05 may include, but are not limited to processor 50, memory 51.It will be understood by those skilled in the art that
Fig. 5 is only the example for calculating equipment 05, does not constitute the restriction to equipment 05 is calculated, and may include more or more than illustrating
Few component perhaps combines certain components or different components, for example, calculate equipment can also include input-output equipment,
Network access equipment, bus etc..
Alleged processor 50 can be central processing unit (Central Processing Unit, CPU), can also be
Other general processors, digital signal processor (Digital Signal Processor, DSP), specific integrated circuit
(Application Specific Integrated Circuit, ASIC), ready-made programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic device
Part, discrete hardware components etc..General processor can be microprocessor or the processor is also possible to any conventional processing
Device etc..
Memory 51 can be the internal storage unit for calculating equipment 05, such as calculate the hard disk or memory of equipment 05.It deposits
Reservoir 51 is also possible to calculate the External memory equipment of equipment 05, such as calculates the plug-in type hard disk being equipped in equipment 05, intelligence
Storage card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card, flash card (Flash Card)
Deng.Further, memory 51 can also both include calculating the internal storage unit of equipment 05 or including External memory equipment.
Memory 51 is for other programs and data needed for storing computer program and calculating equipment.Memory 51 can also be used
In temporarily storing the data that has exported or will export.
In several embodiments provided herein, it should be understood that disclosed device and method can pass through
Other modes are realized.For example, the apparatus embodiments described above are merely exemplary, for example, the module is drawn
Point, only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple module or components
It can be combined or can be integrated into another system, or some features can be ignored or not executed.Another point, it is shown or
The mutual coupling, direct-coupling or communication connection discussed can be through some interfaces, device or module it is indirect
Coupling or communication connection can be electrical property, mechanical or other forms.
The module as illustrated by the separation member may or may not be physically separated, as module
The component of display may or may not be physical module, it can and it is in one place, or may be distributed over more
On a network module.Some or all of the modules therein can be selected to realize this embodiment scheme according to the actual needs
Purpose.
It, can also be in addition, each functional module in each embodiment of the present invention can integrate in a processing module
It is that modules physically exist alone, can also be integrated in two or more modules in a module.Above-mentioned integrated mould
Block both can take the form of hardware realization, can also be realized in the form of software function module.
If the integrated module is realized in the form of software function module and sells or use as independent product
When, it can store in a computer readable storage medium.Based on this understanding, technical solution of the present invention essence
On all or part of the part that contributes to existing technology or the technical solution can be with the shape of software product in other words
Formula embodies, which is stored in a storage medium, including some instructions are used so that a calculating
Machine equipment (can be personal computer, server or the network equipment etc.) executes each embodiment the method for the present invention
All or part of the steps.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only
Memory), random access memory (RAM, Random Access Memory), magnetic or disk etc. are various can store
The medium of program code.
It should be noted that for the various method embodiments described above, describing for simplicity, therefore, it is stated as a systems
The combination of actions of column, but those skilled in the art should understand that, the present invention is not limited by the sequence of acts described,
Because according to the present invention, certain steps can use other sequences or carry out simultaneously.Secondly, those skilled in the art also answer
This knows that the embodiments described in the specification are all preferred embodiments, and related actions and modules might not all be
Necessary to the present invention.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, there is no the portion being described in detail in some embodiment
Point, it may refer to the associated description of other embodiments.
The above are be situated between to a kind of method for building up of image unified model provided by the present invention and device, equipment and storage
The description of matter, for those skilled in the art, thought according to an embodiment of the present invention in specific embodiment and applies model
Place that there will be changes, to sum up, the contents of this specification are not to be construed as limiting the invention.
Claims (10)
1. a kind of method for building up of image unified model, which is characterized in that the described method includes:
It determines spectral value of the high spectrum image on the curve of spectrum of designated flying zone point, and determines the gradient letter of the spectral value
Breath；
According to the spectral value and the gradient information, the comprehensive characteristics parameter of the high spectrum image is determined；
The unified model of the high spectrum image is established based on the comprehensive characteristics parameter, and using the unified model to described
High spectrum image carries out local feature detection.
2. the method as described in claim 1, which is characterized in that it is described according to the spectral value and the gradient information, it determines
The step of comprehensive characteristics parameter of the high spectrum image includes:
According to the spectral value and the gradient information, the spectrum change of gradient vector modulus value of the airspace point is calculated；
The spectral value of the airspace point is modified, the revised amendment spectral value of airspace point is obtained；
Based on the spectrum change of gradient vector modulus value and the amendment spectral value, the comprehensive characteristics parameter is determined.
3. method according to claim 2, which is characterized in that it is described according to the spectral value and the gradient information, it calculates
The step of spectrum change of gradient vector modulus value of the airspace point includes:
Based on the gradient information of the airspace point, the gradient value of the airspace point is calculated；
Using calculated gradient value, the average ladder of each point in the field of the default size centered on the airspace point is calculated
Spend modulus value；
The spectrum change of gradient vector modulus value of the airspace point is determined based on the average gradient modulus value.
4. method according to claim 2, which is characterized in that the spectral value to the airspace point is modified, and is obtained
The step of airspace point revised amendment spectral value includes:
Using pre-set spectral value correction amount computational algorithm, the corresponding correction value of the spectral value is calculated；
When the correction value is less than preset correction threshold, the spectral value is modified using the correction value.
5. the method as described in Claims 1-4 any one, which is characterized in that the gradient of the determination spectral value is believed
The step of breath includes:
Determine the spectral value on the direction of airspace with the change rate on spectral domain direction；
The gradient information of the spectral value is calculated based on the change rate.
6. a kind of image unified model establishes device, which is characterized in that described device includes:
First determining module for determining spectral value of the high spectrum image on the curve of spectrum of designated flying zone point, and determines institute
State the gradient information of spectral value；
Second determining module, for determining the comprehensive special of the high spectrum image according to the spectral value and the gradient information
Levy parameter；
Module is established, for establishing the unified model of the high spectrum image based on the comprehensive characteristics parameter, and described in utilization
Unified model carries out local feature detection to the high spectrum image.
7. device as claimed in claim 6, which is characterized in that second determining module includes:
Computing module, for calculating the spectrum change of gradient arrow of the airspace point according to the spectral value and the gradient information
Measure modulus value；
Correction module is modified for the spectral value to the airspace point, obtains the revised amendment spectrum of airspace point
Value；
Parameter determination module, for determining described comprehensive based on the spectrum change of gradient vector modulus value and the amendment spectral value
Close characteristic parameter.
8. device as claimed in claim 7, which is characterized in that the computing module includes:
First computing module calculates the gradient value of the airspace point for the gradient information based on the airspace point；
Second computing module calculates the neck of the default size centered on the airspace point for utilizing calculated gradient value
The average gradient modulus value of each point in domain；
Third computing module, for determining the spectrum change of gradient Vector Mode of the airspace point based on the average gradient modulus value
Value.
9. device as claimed in claim 7, which is characterized in that the correction module is specifically used for:
Using pre-set spectral value correction amount computational algorithm, the corresponding correction value of the spectral value is calculated, when the amendment
When value is less than preset correction threshold, the spectral value is modified using the correction value.
10. the device as described in claim 6 to 9 any one, which is characterized in that first determining module includes:
Change rate determining module, for determine the spectral value on the direction of airspace with the change rate on spectral domain direction；
The gradient information of the spectral value is calculated based on the change rate for gradient information computing module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811003368.8A CN109360264B (en) | 2018-08-30 | 2018-08-30 | Method and device for establishing unified image model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811003368.8A CN109360264B (en) | 2018-08-30 | 2018-08-30 | Method and device for establishing unified image model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109360264A true CN109360264A (en) | 2019-02-19 |
| CN109360264B CN109360264B (en) | 2023-05-26 |
Family
ID=65350192
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811003368.8A Active CN109360264B (en) | 2018-08-30 | 2018-08-30 | Method and device for establishing unified image model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109360264B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2187341A1 (en) * | 2008-11-12 | 2010-05-19 | Fundación Robotiker | Method for modelling electromagnetic spectra |
| US8897570B1 (en) * | 2011-03-31 | 2014-11-25 | Raytheon Company | Detection of targets from hyperspectral imagery |
| CN105139412A (en) * | 2015-09-25 | 2015-12-09 | 深圳大学 | Hyperspectral image corner detection method and system |
| CN107633216A (en) * | 2017-09-07 | 2018-01-26 | 深圳大学 | The feature extracting method and device of high-spectrum remote sensing |
| CN107742114A (en) * | 2017-11-09 | 2018-02-27 | 深圳大学 | high spectrum image feature detection method and device |
| CN108009550A (en) * | 2017-11-09 | 2018-05-08 | 深圳大学 | High spectrum image feature detection method and device based on curve of spectrum fitting |
-
2018
- 2018-08-30 CN CN201811003368.8A patent/CN109360264B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2187341A1 (en) * | 2008-11-12 | 2010-05-19 | Fundación Robotiker | Method for modelling electromagnetic spectra |
| US8897570B1 (en) * | 2011-03-31 | 2014-11-25 | Raytheon Company | Detection of targets from hyperspectral imagery |
| CN105139412A (en) * | 2015-09-25 | 2015-12-09 | 深圳大学 | Hyperspectral image corner detection method and system |
| CN107633216A (en) * | 2017-09-07 | 2018-01-26 | 深圳大学 | The feature extracting method and device of high-spectrum remote sensing |
| CN107742114A (en) * | 2017-11-09 | 2018-02-27 | 深圳大学 | high spectrum image feature detection method and device |
| CN108009550A (en) * | 2017-11-09 | 2018-05-08 | 深圳大学 | High spectrum image feature detection method and device based on curve of spectrum fitting |
Non-Patent Citations (2)
| Title |
|---|
| 雷武虎等: "基于谱域-空域结合的高光谱异常探测", 《光电技术应用》 * |
| 高恒振等: "一种基于谱域-空域组合特征支持向量机的高光谱图像分类算法", 《宇航学报》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109360264B (en) | 2023-05-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chen et al. | Matrix completion via an alternating direction method | |
| CN110909611B (en) | Method and device for detecting attention area, readable storage medium and terminal equipment | |
| Yan et al. | Efficient computation of clipped Voronoi diagram for mesh generation | |
| Dufour et al. | Shape, displacement and mechanical properties from isogeometric multiview stereocorrelation | |
| CN109325430B (en) | Real-time behavior identification method and system | |
| Boscaini et al. | Shape‐from‐operator: Recovering shapes from intrinsic operators | |
| Chen et al. | The Fiedler vector of a Laplacian tensor for hypergraph partitioning | |
| CN104809695B (en) | A kind of method and device of data de-noising | |
| CN105608478B (en) | image feature extraction and classification combined method and system | |
| Gundogdu et al. | Garnet++: Improving fast and accurate static 3d cloth draping by curvature loss | |
| Wang et al. | A novel GCN-based point cloud classification model robust to pose variances | |
| Raudies | Optic flow | |
| CN104751112B (en) | A kind of fingerprint template and fingerprint identification method based on fuzzy characteristics point information | |
| Sun et al. | Multi-structure KELM with attention fusion strategy for hyperspectral image classification | |
| CN106127179A (en) | Based on the Classification of hyperspectral remote sensing image method that adaptive layered is multiple dimensioned | |
| Villalpando et al. | FPGA implementation of stereo disparity with high throughput for mobility applications | |
| CN108875903A (en) | Method, apparatus, system and the computer storage medium of image detection | |
| Deng et al. | A generalized low-rank double-tensor nuclear norm completion framework for infrared small target detection | |
| Seversky et al. | Harmonic point cloud orientation | |
| Rouxel-Labbé et al. | Discretized riemannian delaunay triangulations | |
| CN109360264A (en) | The method for building up and device of image unified model | |
| Fan et al. | Explicit shape regression with characteristic number for facial landmark localization | |
| US20160292529A1 (en) | Image collation system, image collation method, and program | |
| Lejemble et al. | Stable and efficient differential estimators on oriented point clouds | |
| US20180330514A1 (en) | Selective 3d registration |
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 |