CN107358635A - A kind of Color-scale Morphology image processing method based on fuzzy comparability - Google Patents
A kind of Color-scale Morphology image processing method based on fuzzy comparability Download PDFInfo
- Publication number
- CN107358635A CN107358635A CN201710591789.6A CN201710591789A CN107358635A CN 107358635 A CN107358635 A CN 107358635A CN 201710591789 A CN201710591789 A CN 201710591789A CN 107358635 A CN107358635 A CN 107358635A
- Authority
- CN
- China
- Prior art keywords
- colour
- color
- phasor
- vector
- colour phasor
- 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
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of Color-scale Morphology image processing method based on fuzzy comparability, is related to technical field of image processing.Pending coloured image is corresponded to RGB color space by this method first, it is determined that represent that the fuzzy comparability of degree of similarity between two colour phasors estimates (i.e. FSM) function, then using a pixel as index, obtain construction unit and its colour phasor collection one by one in RGB color space, the supremum vector infimum vector of the colour phasor collection is determined using FSM as criterion, the basic operation of Color-scale Morphology is built and is applied to coloured image.Morphology thought is applied to Color Image Processing by the present invention, have the characteristics that stability is good, practical in the processing of actual chromatic image, it is not only able to smooth target color, and can be very good to handle the minutia of homogeneous region pixel inconsistency, it is finally reached the target of Color Image Analysis and processing.
Description
Technical field
The present invention relates to technical field of image processing, more particularly to a kind of Color-scale Morphology image based on fuzzy comparability
Processing method.
Background technology
Morphologic basic conception appeared in for 19th century earliest, and morphology is applied into image processing field, exactly utilized
Structural element (circle, square, line segment etc.) extracts and collected image information, special with the geometry of recognisable image target to analyze
Seek peace structure.As a kind of Nonlinear image processing and analysis method, mathematical morphology is successfully applied to binary and grayscale image,
And form complete morphology theory.Wherein, binary morphology regards bianry image as set, is transported with simplest set
(such as intersecting and merging, benefit and translation) is calculated to detect original image.Because the computing of use is based on gathering, therefore two-value shape
State learning aid have clear principle, calculate it is simple, be easy to extend, and the features such as be suitable for parallel computation, and be widely used in denoising, side
The image processing fields such as boundary's detection, skeletonizing, region segmentation.Gray scale morphology is developed by binary morphology, is only needed
Intersecting and merging in binary morphology are changed into the minimum and maximum value of pixel grey scale in construction unit respectively.
In recent years, coloured image is extensively with human being's production, the every field of life.Compared with binary and grayscale image
Compared with addition to monochrome information, coloured image also includes the color information that the mankind can perceive, therefore Color Image Processing is by more
Carry out more concerns.At present, mathematical morphology is generalized to coloured image, and is achieved in the filtering, segmentation, spy of coloured image
Extraction, rim detection, image enhancement and restoration etc. are levied, is an important research direction in Color Image Processing field.
Generally, it is to see RGB color image gray scale morphology to be extended into the most plain mode of Color-scale Morphology
Formed into by the width monochrome image of red, green, blue three, this three width monochrome image is handled respectively using gray scale morphology, finally by form
Learn result and be reduced to RGB color image.But on the one hand colour that such result will change in original image, so as to lose
Lose or distort original image information；On the other hand due to not accounting for the correlation between each component of coloured image, and red, green, blue is made
Three width images become the independent image being not in contact with.Another method for gray scale morphology being extended to Color-scale Morphology is exactly base
Colour in multivariate data ranking criteria, such as edge sequence, condition sequence, region sequence and degeneration sequence, ordering structure unit, is used in combination
The principle of minimum and maximum value defines grown form operation in similar gray scale morphology, but for colourful coloured image
Suitable, general colour phasor sort method is difficult to find that, so these current Color-scale Morphology methods have very big office
It is sex-limited.
The content of the invention
The technical problem to be solved in the present invention is to be directed to above-mentioned the deficiencies in the prior art, there is provided one kind is based on fuzzy comparability
Color-scale Morphology image processing method, in RGB color space, based on the fuzzy similar of the degree of similarity portrayed between vector
Property estimates concept, defines a kind of new type colorful morphological operation, is not only able to smooth target color, and can be very good to handle
The minutia of homogeneous region pixel inconsistency, it is finally reached the target of Color Image Analysis and processing.
In order to solve the above technical problems, the technical solution used in the present invention is：
A kind of Color-scale Morphology image processing method based on fuzzy comparability, comprises the following steps：
Step 1：Read pending coloured image；
Step 2：Pending coloured image is corresponded into RGB color space, i.e., by the pixel of coloured image with the colour
The vector representation in space, and each colour phasor is formed by red (R), green (G), blue (B) three kinds of chrominance components；
Step 3：It is determined that represent that the fuzzy comparability of degree of similarity between any two colour phasor of color space is estimated (i.e.
FSM) function, and the parameter in function is determined；When the value that fuzzy comparability is estimated is bigger, two colour phasors are similar
Property is higher, conversely, two colour phasor similitudes are lower；
Step 4：A pixel is obtained in color space, and current structure unit is constructed as shifting using centered on the pixel
Dynamic window, and the colour phasor collection formed in current structure unit；The size of current structure unit according to be actually needed choose；
Step 5：Determine that the supremum vector infimum of colour phasor collection that current pixel formed is sweared using FSM as criterion
Amount；
Step 6：Utilize the basic of the supremum and infimum of current pixel and its colour phasor collection structure Color-scale Morphology
Operation, including expansive working, etching operation, opening operation and closed operation；
Step 7：The basic operation for the Color-scale Morphology that step 6 is proposed is applied to coloured image, with supremum vector or
Infimum vector replaces current color vector, i.e., current pixel is replaced with supremum or infimum pixel, as current structure list
The morphology output of member；
Step 8：Judge whether pending coloured image all pixels have been handled, such as untreated complete, then return to step 4,
Residual pixel is handled, untill all pixels have been handled.
Preferably, in the step 3, with the distance of colour phasor and angle jointly constructs and two colour phasors are portrayed
Between the fuzzy comparability of similitude estimate FSM, as shown in formula (1), between zero and one, two colour phasors are same to its functional value
During vector, its value is 1, and when two colour phasors are entirely different, its value is 0；
Wherein, k1=[0, ∞) and k2=[0,1] is fuzzy control parameter, is artificially set two to obscure according to actual conditions
Control parameter k1And k2Value, parameter k1And k2Various combination directly determine the similarity degrees of two colour phasors, with k1、
k2Increase, the similarity measure between colour phasor reduces, and similarity degree reduces；Conversely, similarity degree raises；Parameter k1Compare k2
Influence to similarity measure is bigger, and similarity measure is for k1Change it is more sensitive；d(vi, vi) represent to appoint in coloured image
Anticipate two colour phasor viAnd vjBetween vector distance；θ(vi, vj) represent any two colour phasor v in coloured imageiAnd vjBetween
Vector angle；
When two colour phasors in distance and angle closer to when, i.e., its distance and angle are smaller, then similarity measure
Value it is bigger, represent two colour phasor similitudes it is higher；Conversely, when distance between two colour phasors and bigger angle, then
The value of similarity measure is smaller, and two colour phasor similitudes are lower.
Preferably, the step 5 comprises the following steps that：
Step 5.1：According between the colour phasor concentration any two colour phasor in formula (1) calculating current structure unit
FSM functional values；
Step 5.2：The FSM functional values calculated based on step 5.1 determine colored arrow least similar in current structure unit
Amount therefrom randomly selects a pair of least similar colour phasors pair, according to the least similar colour phasor centering two to set
The size of individual colour phasor modulus value, it is determined that maximum colour phasor and minimum color vector, the big person of modulus value is maximum colour phasor, mould
It is minimum color vector to be worth small person；
Step 5.3：Using maximum colour phasor and minimum color vector as core, colour that current structure unit is formed
Vector set is divided into two subsets, i.e., maximum colour phasor subset and minimum color subset of vectors；For current structure unit institute structure
Into colour phasor concentrate any colour phasor, if the similarity measure of any colour phasor and maximum colour phasor is big
In the similarity measure of any colour phasor and minimum color vector, then any colour phasor belongs to maximum colour phasor
Collection, otherwise, any colour phasor belongs to minimum color subset of vectors；
Step 5.4：The highest similitude calculated respectively in maximum colour phasor subset and minimum color subset of vectors is colored
Vector；
Step 5.5：The highest similitude colour phasor obtained in minimum color subset of vectors is defined as infimum vector,
The highest similitude colour phasor obtained in maximum colour phasor subset is defined as supremum vector.
Preferably, when obtaining pixel in step 4, pressed from left to right, from top to bottom in pending coloured image
Order obtains pixel one by one.
It is using beneficial effect caused by above-mentioned technical proposal：One kind provided by the invention obscures with reference to color space
The Color-scale Morphology image processing method of similitude, this method define one kind in RGB color space and feel to be consistent with the mankind,
Fuzzy comparability for measuring colour phasor similarity relationships is estimated, and proposes a kind of form applied to coloured image accordingly
Method.Morphology thought is really applied to Color Image Processing by the present invention, not only avoid the problem of colour phasor sequence,
And the problem of being lost in Color Image Processing or distorting original image information is efficiently solved, in the processing of actual chromatic image
Have the characteristics that stability is good, practical, be not only able to smooth target color, and can be very good to handle homogeneous region picture
The minutia of plain inconsistency, be finally reached the target of Color Image Analysis and processing, be the filtering of coloured image, segmentation,
Feature extraction, rim detection, image enhancement and restoration etc. provide new approaches.
Brief description of the drawings
Fig. 1 is the Color-scale Morphology image processing method flow chart provided in an embodiment of the present invention based on fuzzy comparability；
Fig. 2 is the corresponding colour phasor table into rgb space of color image pixel in step 2 provided in an embodiment of the present invention
Show；
Fig. 3 is the schematic diagram of determination FSM functions provided in an embodiment of the present invention；
Fig. 4 is influence schematic diagram of the fuzzy control parameter of FSM functions provided in an embodiment of the present invention to similitude；
Fig. 5 is the particular flow sheet of step 5 provided in an embodiment of the present invention；
Fig. 6 is high-resolution remote sensing image provided in an embodiment of the present invention；
Fig. 7 is Color-scale Morphology operating result provided in an embodiment of the present invention, wherein, (a) is etching operation result, (b)
For expansive working result, (c) is opening operation operating result, and (d) is closed operation operating result；
Fig. 8 is the evaluation provided in an embodiment of the present invention to Color-scale Morphology operating characteristics, wherein (a) is and original color shadow
The average value contrast schematic diagram of gray level image similarity measure as corresponding to, (b) are to show with the contrast of the degree of similarity of former image
It is intended to.
Embodiment
With reference to the accompanying drawings and examples, the embodiment of the present invention is described in further detail.Implement below
Example is used to illustrate the present invention, but is not limited to the scope of the present invention.
A kind of Color-scale Morphology image processing method based on fuzzy comparability, as shown in figure 1, comprising the following steps：
Step 1：Read pending coloured image；
In the present embodiment, it is X={ x to define pending coloured imagei, i=1 ..., N }, wherein, xiIt is ith pixel,
I is pixel index, and N is total pixel number, and X is 150 × 150 pixels, total pixel number N=22500.
Step 2：Pending coloured image is corresponded into RGB color space, i.e., by the pixel of coloured image with the colour
The vector representation in space, and each colour phasor is formed by red (R), green (G), blue (B) three kinds of chrominance components.
For pending coloured image X, RGB color space is corresponded to, its N number of pixel forms colour phasor set
For VX={ vi, i=1 ..., N }, wherein, vi=(viR, viG, viB) be colour element i colour phasor, viR、viG、viBTable respectively
Show colour phasor viRed, green, blue component, as shown in Figure 2.
Step 3：It is determined that the fuzzy comparability for portraying degree of similarity between two colour phasors estimates (FSM) function, and it is right
Function parameter therein is determined, when the value that fuzzy comparability is estimated is bigger, then it is assumed that they have higher similar journey
The colour of degree, i.e. two corresponding to it pixel is more similar, and this is also just consistent with the visually-perceptible of the mankind.FSM functions
As shown in Figure 3.
In RGB color space, the colour phasor v of any two pixel i and j in coloured image are representediAnd vjIt is similar
Property is expressed as by distance between vector：
d(vi, vj)=((viR-vjR)2+(viG-vjG)2+(viB-vjB)2)1/2
As the distance d (v of two colour phasorsi, vj) get over hour, represent that two vector similarity degree are high, i.e. two pixels
Point i and j colors are visually closer, and visually difference is larger on the contrary then expression pixel i and j color.Similarly, it is also possible to
The angle (i.e. vector relative direction) of colour phasor carries out the description of similitude, is shown below.
As angle, θ (v between the vector that two pixels i and j are formedi, vj) smaller, then it represents that two pixel similitudes are got over
Height, when angle is bigger between the vector that two pixels i and j are formed, then it represents that two pixel similitudes are lower.But both the above
Function is defined because certain deficiency and limitation be present only considering the factor of one side.
In the present embodiment, with the distance and angle jointly constructs of colour phasor and similitude between two colour phasors is portrayed
Fuzzy comparability estimate FSM, it is defined as follows：
Wherein, k1=[0, ∞) and k2=[0,1] is fuzzy control parameter, and two are artificially flexibly set according to actual conditions
Fuzzy control parameter k1And k2Value, parameter k1And k2Artificial regulation fuzzy comparability can be made to estimate FSM not only can be from objective
On portray similitude between two vectors, it is and consistent with the visually-perceptible that the mankind obscure, i.e., for two same coloured silks
Vector in colour space, cognition of the different people to its similarity degree have otherness.Assuming that the distance and angle of two colour phasors are respectively d
=1 and θ=π/4, similarity measure is with parameter k between two vectors1And k2Change as shown in figure 4, can be clearly from figure
Go out, for two same colour phasors, as parameter k1And k2When taking different value, estimated by the fuzzy comparability obtained by formula (1)
It is different, parameter k1And k2Various combination directly determine the similarity degrees of two vectors, with k1、k2Increase, vector
Between similarity measure reduce, similarity degree reduces, conversely, similarity degree raises, and parameter k1Compare k2To similarity measure
Influence is bigger, and similarity measure is for k1Change it is more sensitive.As seen from the above analysis, this phase for meeting colour phasor
Like property because of people, environment and different feature, can meet and meet during Color Image Processing because the visually-perceptible of people is brought
Uncertainty and ambiguity.
FSM is similar to fuzzy membership function, and its functional value is between zero and one.As two colour phasor viAnd vjIn distance
With in angle closer to when, i.e., its distance and angle are smaller, and the value of similarity measure is bigger, represent both with higher similar
Property；Conversely, when distance between two colour phasors and bigger angle, then the value of its similarity measure is smaller, represents similarity degree
It is relatively low.When two colour phasors overlap (being same vector), distance and angle are 0, then similarity measure is 1, is characterized
Its similarity degree highest, these are all consistent with actual visually-perceptible, while also overcome and only defined and brought by distance or angle
Limitation.
Step 4：In color space VXOne pixel i of middle acquirement, and construct current structure unit centered on pixel i and make
For moving window, the size of current structure unit is chosen according to being actually needed, and the colour phasor formed in current structure unit
Collection Wherein include Current central vector viQ colour phasor inside.In the present embodiment, obtain
During pixel, in pending coloured image VXIn by order from left to right, from top to bottom obtain pixel one by one.
Step 5：Determine that the supremum vector infimum of colour phasor collection that current pixel formed is sweared using FSM as criterion
Amount, as shown in figure 5, comprising the following steps that.
Step 5.1：According between the colour phasor concentration any two colour phasor in formula (1) calculating current structure unit
FSM functional values.
Step 5.2：The FSM functional values calculated based on step 5.1 determine colored arrow least similar in current structure unit
Amount is to set Vids, a pair of least similar colour phasors pair are therefrom randomly selected, according to the least similar colour phasor centering
The size of two colour phasor modulus value, it is determined that maximum colour phasor and minimum color vector, the big person of modulus value is maximum colour phasor,
The small person of modulus value is minimum color vector, and it is defined as follows shown in formula：
Wherein, p ViMiddle pixel index, q are colour phasor numbers in the set.According to above-mentioned definition, set VidsIt is Vi×
ViThe minimum vector of all colour phasor centering similarity measures is to set.In VidsIn, a vector is randomly selected to as least
Similar colour phasor pair, is designated as (vids1, vids2)∈Vids.By the size v of its mouldids1And vids2It is referred to as maximum colored arrow
Measure (the big person of modulus value) vimaxWith minimum color vector (the small person of modulus value) vimin。
Step 5.3：Using maximum colour phasor and minimum color vector as core, colour that current structure unit is formed
Vector set is divided into two subsets, is divided into maximum colour phasor subset and minimum color subset of vectors, i.e., respectively with minimum and maximum
Colour phasor vimaxAnd viminFor core, by ViIn all colour phasors be divided into two classes, be designated as CLimaxAnd CLimin.For any color
Vector in colour space vip∈ViIf vipWith vimaxSimilarity measure be more than vipWith viminSimilarity measure, then vipBelong to CLimax,
Otherwise, vipBelong to CLimin, i.e.,：
CLimin={ vip, μ (vip, vimin)≥μ(vip, vimax), vip∈Vi} (3)
CLimax={ vip, μ (νip, vimax)≥μ(vip, vimin), vip∈Vi} (4)
Step 5.4：The two subsets CL is calculated respectivelyimaxAnd CLiminIn highest similitude colour phasor.
In CLiminAnd CLimaxIn, define highest similitude colour phasor vcliminAnd vclimaxRespectively：
Step 5.5：By subset CLiminThe highest similitude vector definition of middle acquisition is infimum vector, by subset CLimax
The highest similitude vector definition of middle acquisition is supremum vector.
According to the definition of formula (5) and (6) highest similitude colour phasor, so that it may structural texture unit ViIn infimum behaviour
Make ∧ and supremum operation ∨ is respectively：
∧Vi=∧ { vi1, vi2..., viq}=vclimin (7)
∨Vi=∨ { vi1, vi2..., viq}=vclimax (8)
Infimum operates ∧ it can be seen from formula (7) and (8) and supremum operates ∨ output colour phasor respectively coloured silk
Vector in colour space class CLiminAnd CLjmaxMiddle highest similar color vector vcliminAnd vclimax, they respectively with CLiminAnd CLimaxMiddle institute
There is the similarity measure sum of other colour phasors maximum.In addition, the colored arrow of the supremum and infimum operation output of definition
Amount be acquisition coloured image in colour phasor in construction unit, i.e. the operation can't produce new colour phasor, very
Original image information is effectively maintained in big degree, this is also the basis for building Color-scale Morphology method.
Step 6：Utilize the basic of the supremum and infimum of current pixel and its colour phasor collection structure Color-scale Morphology
Operation, including expansive working, etching operation, opening operation and closed operation.
The colour phasor ensemble space V formed for pending coloured image X, coloured imageX, the present embodiment defined
The morphological operation of coloured image include expansive working δX, etching operation εX, closed operation χX, opening operation oX, it is respectively defined as：
δX(VX)={ ∨ Vi, i=1,2 ..., N } and (9)
εX(VX)={ ∧ Vi, i=1,2 ..., N } and (10)
χX(VX)=εX(δX(VX)) (11)
oX(VX)=δX(εX(VX)) (12)
By above Color-scale Morphology operate it can be seen from for the vectorial set V of coloured imageXExpansive working be exactly tying
Structure unit ViIn find out the supremum of the colour phasor subset；Similarly, etching operation is exactly in construction unit ViIn find out the colour
The infimum of subset of vectors；Open and close operator is expansion, the compound operation of corrosion, simply calculates the order on upper and lower really boundary every time not
With.So, the Morphological scale-space to coloured image not only can be effectively completed, and largely remains artwork
The pixel color vector essential characteristic of picture.
Step 7：The basic operation for the Color-scale Morphology that step 6 is proposed is applied to coloured image, with supremum vector or
Infimum vector replaces current color vector, i.e., current pixel is replaced with supremum or infimum pixel, as current structure list
The morphology output of member.
Step 8：Judge whether pending coloured image all pixels have been handled, such as untreated complete, then return to step 4,
One other pixel is obtained by order from left to right, from top to bottom, same treatment is carried out to residual pixel, at all pixels
Untill having managed, the processing of coloured image is realized, for the filtering of follow-up chromatic image, segmentation, feature extraction, rim detection, figure
Image intensifying and recovery etc. lay the foundation.
In the present embodiment, operated and tested as Color-scale Morphology as 150 × 150 pixel high-resolution remote sensing images using yardstick
Image, to observe and analyze its disposal ability and effect to coloured image, remote sensing image is as shown in Figure 6.Fuzzy comparability is surveyed
Spend parameter k in FSM1=0.001 and k2=0.8, by theoretical and experiment simulation, when parameter takes this value for the image processing
Effect is preferable., can be according to the specific value of two parameters of real image processing requirement real time modifying in specific implementation.Construction unit
For the square of the pixel of yardstick 3 × 3, for its Color vector morphological operating result as shown in fig. 7, wherein (a) is corrosion, (b) is expansion,
(c) it is opening operation, (d) is closed operation.From Fig. 7 (a) as can be seen that etching operation expands shade in original image (i.e. background) area
Big, other targets are compressed by, and small gap originally is padded, while the color of each target and edge become more to put down
It is sliding；From Fig. 7 (b) as can be seen that the result of expansive working is just with etching operation on the contrary, the shade (background) of original image partly quilt
It has compressed, other targets are extended, i.e., light color is expanded, and dark color is compressed, and " burr " in original image is eliminated, same mesh
Target color and edge are also smoothened.It is open and close operator, similar to burn into dilation operation, as shown in Fig. 7 (c) and (d), pass through
Top-operation result can be seen that the Color-scale Morphology operation that the present embodiment is proposed and meet morphological operation general principle, and
Can preferably solve the problems, such as that homogeneous region pixel is inconsistent, this will make the color images based on Color-scale Morphology, filtering,
Objective extraction etc. is possibly realized.
The Color-scale Morphology that the present embodiment is proposed is operated with the Color-scale Morphology by gray scale morphology expansion, based on region
The morphology of sequence is contrasted, and still selects RS Color Image shown in Fig. 6 as test image.What gray scale morphology was expanded
Chromatic image is exactly regarded as and is made up of the width monochrome image of red, green, blue three by Color-scale Morphology, is handled respectively using gray scale morphology
This three width monochrome image, finally reduce chromatic image with Morphological scale-space result.Morphology based on region ordering is exactly to work as
The gradient of vector is ranked up and minimum, the maximum vector of gradient is defeated as morphology in forefoot area (region is 3 × 3 pixels)
Go out.The parameter k that fuzzy comparability is estimated in the Color-scale Morphology operation that the present embodiment proposes1=0.001 and k2=0.8, structure list
Member is 3 × 3 pixels.For grayscale morphologic theory, carry out burn into expansion, open and close etc. with three kinds of different shape methods and grasp
Chromatic image after work is converted into its corresponding gray level image, seeks its gray level image similarity measure corresponding to original color image
Average value, as shown in Fig. 8 (a), it should be apparent that the operation such as burn into expansion, open and close can change former image it is right
Half-tone information, the Color-scale Morphology of the present embodiment and the Color-scale Morphology expanded by gray scale morphology are answered, completes same morphology
Gray level image after processing has higher similitude and uniformity, and the Color-scale Morphology that the present embodiment proposes with former gray level image
Change to chromatic image is gentler, that is, the details for being more beneficial for image greyscale information is kept, and based on region ordering
Morphology is less desirable in gray value change.In addition, contrasted from chromatic image similitude angle, computation of morphology behaviour
The fuzzy comparability for making result and former image respective pixel estimates average value, i.e., corrosion-artwork, expansion-artwork, open-artwork,
- artwork is closed, its degree of similarity with former image is evaluated, as shown in Fig. 8 (b), it can be seen that burn into expansion, open and close etc. are grasped
Make change former image, but similitude highest after the Color-scale Morphology operation of the present embodiment proposition, that is to say, that it is same reaching
In the case of sample chromatic image processing target, the morphological operation of the present embodiment proposition is minimum to former image detail characteristic change,
Thus can greatly reduce because of operation and caused by chromatic image distortion phenomenon.By quantifying to contrast, it can be seen that proposition
, not only can be with the two-value and grayscale morphologic theory of classics when Color-scale Morphology based on fuzzy comparability is handled image
It is consistent, and the minutia of former image can be kept well, there is preferable treatment effect, this is also this Color vector morphological
Promotion and application are laid a good foundation.
Finally it should be noted that：The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
The present invention is described in detail with reference to the foregoing embodiments, it will be understood by those within the art that：It still may be used
To be modified to the technical scheme described in previous embodiment, either which part or all technical characteristic are equal
Replace；And these modifications or replacement, the essence of appropriate technical solution is departed from the model that the claims in the present invention are limited
Enclose.
Claims (4)
- A kind of 1. Color-scale Morphology image processing method based on fuzzy comparability, it is characterised in that：Comprise the following steps：Step 1：Read pending coloured image；Step 2：Pending coloured image is corresponded into RGB color space, i.e., by the pixel of coloured image with the color space Vector representation, and each colour phasor is formed by red (R), green (G), blue (B) three kinds of chrominance components；Step 3：It is determined that represent that the fuzzy comparability of degree of similarity between any two colour phasor of color space is estimated (i.e. FSM) Function, and the parameter in function is determined；When the value that fuzzy comparability is estimated is bigger, two colour phasor similitudes are got over Height, conversely, two colour phasor similitudes are lower；Step 4：A pixel is obtained in color space, and constructs current structure unit using centered on the pixel and is used as Moving Window Mouthful, and the colour phasor collection formed in current structure unit；The size of current structure unit according to be actually needed choose；Step 5：The supremum vector infimum vector of colour phasor collection that current pixel formed is determined using FSM as criterion；Step 6：The basic operation of Color-scale Morphology is built using the supremum and infimum of current pixel and its colour phasor collection, Including expansive working, etching operation, opening operation and closed operation；Step 7：The basic operation for the Color-scale Morphology that step 6 is proposed is applied to coloured image, with supremum vector or lower true Boundary's vector replaces current color vector, i.e., current pixel is replaced with supremum or infimum pixel, as current structure unit Morphology exports；Step 8：Judge whether pending coloured image all pixels have been handled, such as untreated complete, then return to step 4, to surplus After image element is handled, untill all pixels have been handled.
- 2. the Color-scale Morphology image processing method according to claim 1 based on fuzzy comparability, it is characterised in that： In the step 3, with the distance of colour phasor and angle jointly constructs and the fuzzy phase of similitude between two colour phasors is portrayed Estimate FSM like property, as shown in formula (1), between zero and one, when two colour phasors are same vector, its value is 1 to its functional value, When two colour phasors are entirely different, its value is 0；<mrow> <mi>&mu;</mi> <mrow> <mo>(</mo> <msub> <mi>v</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>v</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <msup> <mi>e</mi> <mrow> <mo>-</mo> <msub> <mi>k</mi> <mn>1</mn> </msub> <mi>d</mi> <mrow> <mo>(</mo> <msub> <mi>v</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>v</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> </mrow> </msup> <mi>c</mi> <mi>o</mi> <mi>s</mi> <mrow> <mo>(</mo> <msub> <mi>k</mi> <mn>2</mn> </msub> <mi>&theta;</mi> <mo>(</mo> <mrow> <msub> <mi>v</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>v</mi> <mi>j</mi> </msub> </mrow> <mo>)</mo> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow>Wherein, k1=[0, ∞) and k2=[0,1] is fuzzy control parameter, and two fuzzy controls are artificially set according to actual conditions Parameter k1And k2Value；d(vi, vj) represent any two colour phasor v in coloured imageiAnd vjBetween vector distance；θ(vi, vj) represent any two colour phasor v in coloured imageiAnd vjBetween vector angle；When two colour phasors in distance and angle closer to when, i.e., its distance and angle are smaller, then the value of similarity measure It is bigger, represent that two colour phasor similitudes are higher；Conversely, when distance between two colour phasors and bigger angle, then it is similar The value that property is estimated is smaller, and two colour phasor similitudes are lower.
- 3. the Color-scale Morphology image processing method according to claim 2 based on fuzzy comparability, it is characterised in that：Institute State comprising the following steps that for step 5：Step 5.1：According between the colour phasor concentration any two colour phasor in formula (1) calculating current structure unit FSM functional values；Step 5.2：The FSM functional values calculated based on step 5.1 determine colour phasor pair least similar in current structure unit Set, therefrom randomly selects a pair of least similar colour phasors pair, according to least similar two coloured silks of colour phasor centering The size of vector in colour space modulus value, it is determined that maximum colour phasor and minimum color vector, the big person of modulus value is maximum colour phasor, and modulus value is small Person is minimum color vector；Step 5.3：Using maximum colour phasor and minimum color vector as core, colour phasor that current structure unit is formed Collection is divided into two subsets, i.e., maximum colour phasor subset and minimum color subset of vectors；Formed for current structure unit Any colour phasor that colour phasor is concentrated, if the similarity measure of any colour phasor and maximum colour phasor is more than and is somebody's turn to do The similarity measure of any colour phasor and minimum color vector, then any colour phasor belong to maximum colour phasor subset, Otherwise, any colour phasor belongs to minimum color subset of vectors；Step 5.4：The highest similitude colour phasor in maximum colour phasor subset and minimum color subset of vectors is calculated respectively；Step 5.5：The highest similitude colour phasor obtained in minimum color subset of vectors is defined as infimum vector, will most The highest similitude colour phasor obtained in big colour phasor subset is defined as supremum vector.
- 4. the Color-scale Morphology image processing method according to claim 1 based on fuzzy comparability, it is characterised in that： When pixel is obtained in the step 4, picture is obtained one by one by order from left to right, from top to bottom in pending coloured image Element.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710591789.6A CN107358635B (en) | 2017-07-19 | 2017-07-19 | Color morphological image processing method based on fuzzy similarity |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710591789.6A CN107358635B (en) | 2017-07-19 | 2017-07-19 | Color morphological image processing method based on fuzzy similarity |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107358635A true CN107358635A (en) | 2017-11-17 |
| CN107358635B CN107358635B (en) | 2020-11-03 |
Family
ID=60284624
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710591789.6A Active CN107358635B (en) | 2017-07-19 | 2017-07-19 | Color morphological image processing method based on fuzzy similarity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107358635B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108428235A (en) * | 2018-02-07 | 2018-08-21 | 西北大学 | Line feature dividing method in Pictorial Map |
| CN109658358A (en) * | 2018-12-25 | 2019-04-19 | 辽宁工程技术大学 | A kind of quick bayer color reconstruction method based on more Steerable filters |
| CN110767310A (en) * | 2019-11-06 | 2020-02-07 | 西安交通大学医学院第二附属医院 | Lung cancer screening system based on big data |
Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5542025A (en) * | 1991-01-16 | 1996-07-30 | Hewlett-Packard Company | Precision Z-interpolation method and apparatus |
| US20020174222A1 (en) * | 2000-10-27 | 2002-11-21 | Cox Earl D. | Behavior experts in e-service management |
| US20050226506A1 (en) * | 2004-04-09 | 2005-10-13 | Shmuel Aharon | GPU multi-label image segmentation |
| US20080103744A1 (en) * | 2006-10-25 | 2008-05-01 | Ai Medical Semiconductor | Heart Simulator |
| CN101210966A (en) * | 2006-12-30 | 2008-07-02 | 深圳迈瑞生物医疗电子股份有限公司 | Priority coding after-treatment method and device colorful blood stream imaging |
| CN101246551A (en) * | 2008-03-07 | 2008-08-20 | 北京航空航天大学 | Fast license plate locating method |
| CN101246586A (en) * | 2008-02-22 | 2008-08-20 | 华南师范大学 | Vector map watermarking method based on curve segmentation |
| CN101303767A (en) * | 2007-11-15 | 2008-11-12 | 复旦大学 | Method for registration of digital cucoloris image based on self-adaption sort of block image contents |
| CN101574267A (en) * | 2008-05-08 | 2009-11-11 | 重庆医科大学超声影像学研究所 | Method for counting ultrasonic microvesicles |
| CN101609452A (en) * | 2009-07-10 | 2009-12-23 | 南方医科大学 | The fuzzy SVM feedback that is used for the medical image target identification is estimated method |
| CN101609550A (en) * | 2009-07-10 | 2009-12-23 | 南方医科大学 | Method for registering images based on stream shape subspace |
| CN101847264A (en) * | 2010-05-28 | 2010-09-29 | 北京大学 | Image interested object automatic retrieving method and system based on complementary significant degree image |
| CN102024029A (en) * | 2010-11-30 | 2011-04-20 | 辽宁师范大学 | Local visual attention-based color image retrieving method |
| CN102196119A (en) * | 2010-02-01 | 2011-09-21 | 夏普株式会社 | Image processing method, image processing apparatus and image forming apparatus |
| CN102436656A (en) * | 2011-09-05 | 2012-05-02 | 同济大学 | Animal-diversity monitoring method based on computer vision |
| CN102601950A (en) * | 2012-01-16 | 2012-07-25 | 上海智觉光电科技有限公司 | Method and system for monitoring and detecting target based on fuzzy algorithm |
| CN103136729A (en) * | 2013-01-15 | 2013-06-05 | 西北工业大学 | Fuzzy vector morphological filtering method based on hypercomplex description |
| CN103699532A (en) * | 2012-09-27 | 2014-04-02 | 中国电信股份有限公司 | Image color retrieval method and system |
| CN104899280A (en) * | 2015-05-29 | 2015-09-09 | 山西大学 | Fuzzy-correlated asynchronous image retrieval method based on color histogram and NSCT (Non-Subsampled Contourlet Transform) |
-
2017
- 2017-07-19 CN CN201710591789.6A patent/CN107358635B/en active Active
Patent Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5542025A (en) * | 1991-01-16 | 1996-07-30 | Hewlett-Packard Company | Precision Z-interpolation method and apparatus |
| US20020174222A1 (en) * | 2000-10-27 | 2002-11-21 | Cox Earl D. | Behavior experts in e-service management |
| US20050226506A1 (en) * | 2004-04-09 | 2005-10-13 | Shmuel Aharon | GPU multi-label image segmentation |
| US20080103744A1 (en) * | 2006-10-25 | 2008-05-01 | Ai Medical Semiconductor | Heart Simulator |
| CN101210966A (en) * | 2006-12-30 | 2008-07-02 | 深圳迈瑞生物医疗电子股份有限公司 | Priority coding after-treatment method and device colorful blood stream imaging |
| CN101303767A (en) * | 2007-11-15 | 2008-11-12 | 复旦大学 | Method for registration of digital cucoloris image based on self-adaption sort of block image contents |
| CN101246586A (en) * | 2008-02-22 | 2008-08-20 | 华南师范大学 | Vector map watermarking method based on curve segmentation |
| CN101246551A (en) * | 2008-03-07 | 2008-08-20 | 北京航空航天大学 | Fast license plate locating method |
| CN101574267A (en) * | 2008-05-08 | 2009-11-11 | 重庆医科大学超声影像学研究所 | Method for counting ultrasonic microvesicles |
| CN101609452A (en) * | 2009-07-10 | 2009-12-23 | 南方医科大学 | The fuzzy SVM feedback that is used for the medical image target identification is estimated method |
| CN101609550A (en) * | 2009-07-10 | 2009-12-23 | 南方医科大学 | Method for registering images based on stream shape subspace |
| CN102196119A (en) * | 2010-02-01 | 2011-09-21 | 夏普株式会社 | Image processing method, image processing apparatus and image forming apparatus |
| CN101847264A (en) * | 2010-05-28 | 2010-09-29 | 北京大学 | Image interested object automatic retrieving method and system based on complementary significant degree image |
| CN102024029A (en) * | 2010-11-30 | 2011-04-20 | 辽宁师范大学 | Local visual attention-based color image retrieving method |
| CN102436656A (en) * | 2011-09-05 | 2012-05-02 | 同济大学 | Animal-diversity monitoring method based on computer vision |
| CN102601950A (en) * | 2012-01-16 | 2012-07-25 | 上海智觉光电科技有限公司 | Method and system for monitoring and detecting target based on fuzzy algorithm |
| CN103699532A (en) * | 2012-09-27 | 2014-04-02 | 中国电信股份有限公司 | Image color retrieval method and system |
| CN103136729A (en) * | 2013-01-15 | 2013-06-05 | 西北工业大学 | Fuzzy vector morphological filtering method based on hypercomplex description |
| CN104899280A (en) * | 2015-05-29 | 2015-09-09 | 山西大学 | Fuzzy-correlated asynchronous image retrieval method based on color histogram and NSCT (Non-Subsampled Contourlet Transform) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108428235A (en) * | 2018-02-07 | 2018-08-21 | 西北大学 | Line feature dividing method in Pictorial Map |
| CN108428235B (en) * | 2018-02-07 | 2021-08-27 | 西北大学 | Line element dividing method in color topographic map |
| CN109658358A (en) * | 2018-12-25 | 2019-04-19 | 辽宁工程技术大学 | A kind of quick bayer color reconstruction method based on more Steerable filters |
| CN110767310A (en) * | 2019-11-06 | 2020-02-07 | 西安交通大学医学院第二附属医院 | Lung cancer screening system based on big data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107358635B (en) | 2020-11-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106570486B (en) | Filtered target tracking is closed based on the nuclear phase of Fusion Features and Bayes's classification | |
| CN107092870B (en) | A kind of high resolution image Semantic features extraction method | |
| CN107944396A (en) | A kind of disconnecting link state identification method based on improvement deep learning | |
| CN109613006A (en) | A kind of fabric defect detection method based on end-to-end neural network | |
| CN103186904B (en) | Picture contour extraction method and device | |
| CN107578035A (en) | Human body contour outline extracting method based on super-pixel polychrome color space | |
| CN110321923A (en) | Object detection method, system and the medium of different scale receptive field Feature-level fusion | |
| CN110349131A (en) | A kind of color textile fabric retrochromism detection method | |
| CN107330892A (en) | A kind of sunflower disease recognition method based on random forest method | |
| CN107230188A (en) | A kind of method of video motion shadow removing | |
| CN105894513B (en) | Take the remote sensing image variation detection method and system of imaged object change in time and space into account | |
| CN106056155A (en) | Super-pixel segmentation method based on boundary information fusion | |
| CN108319973A (en) | Citrusfruit detection method on a kind of tree | |
| CN107358635A (en) | A kind of Color-scale Morphology image processing method based on fuzzy comparability | |
| CN105894490A (en) | Fuzzy integration multiple classifier integration-based uterine neck cell image identification method and device | |
| CN109325495A (en) | A kind of crop image segmentation system and method based on deep neural network modeling | |
| CN105894503A (en) | Method for restoring Kinect plant color and depth detection images | |
| CN110706294A (en) | Method for detecting color difference degree of colored textile fabric | |
| CN104036468B (en) | Single-frame image super-resolution reconstruction method based on the insertion of pre-amplification non-negative neighborhood | |
| CN109509171A (en) | A kind of Fabric Defects Inspection detection method based on GMM and image pyramid | |
| CN105303200B (en) | Face identification method for handheld device | |
| CN107123130A (en) | Kernel correlation filtering target tracking method based on superpixel and hybrid hash | |
| CN106650663B (en) | The determination method of building true and false variation and pseudo- containing the method change minimizing technology | |
| CN109658391A (en) | A kind of radius of circle measurement method being fitted based on contour mergence and convex closure | |
| CN109360179A (en) | A kind of image interfusion method, device and readable storage medium storing program for executing |
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 |