CN109509163A - A kind of multi-focus image fusing method and system based on FGF - Google Patents
A kind of multi-focus image fusing method and system based on FGF Download PDFInfo
- Publication number
- CN109509163A CN109509163A CN201811194833.0A CN201811194833A CN109509163A CN 109509163 A CN109509163 A CN 109509163A CN 201811194833 A CN201811194833 A CN 201811194833A CN 109509163 A CN109509163 A CN 109509163A
- Authority
- CN
- China
- Prior art keywords
- source images
- image
- fgf
- fusion
- detail
- 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
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
- G06T2207/20192—Edge enhancement; Edge preservation
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
本发明属于光学图像处理技术领域，公开了一种基于快速导向滤波(Fast Guided Filter，FGF)的多聚焦图像融合方法及系统，用均值滤波对源图像进行平滑，移除源图像中的小结构，并对源图像进行分解得到源图像的基础层和细节层；将拉普拉斯滤波和高斯低通滤波先后对源图像进行滤波处理，得到源图像的显著图；通过对比源图像显著图像素大小得到对应源图像的权值图；将源图像作为引导图像并利用FGF对权值图进行分解优化，分别得到优化的权值图基础层和细节层；根据一定的融合规则，将基础层和细节层对应像素融合；将融合后的基础层和细节层合并，得到融合图像。本发明不但能够有效提高源图像中的聚焦区域检测准确性，且能极大提高融合图像的主客观品质。
The invention belongs to the technical field of optical image processing, and discloses a multi-focus image fusion method and system based on Fast Guided Filter (FGF), which uses mean value filtering to smooth a source image and remove small structures in the source image , and decompose the source image to obtain the base layer and detail layer of the source image; filter the source image successively by Laplacian filtering and Gaussian low-pass filtering to obtain the saliency map of the source image; compare the pixels of the saliency map of the source image According to certain fusion rules, the base layer and The detail layer corresponds to pixel fusion; the fused base layer and the detail layer are combined to obtain a fusion image. The invention can not only effectively improve the detection accuracy of the focus area in the source image, but also greatly improve the subjective and objective quality of the fusion image.
Description
Technical field
The invention belongs to optical image security technical field, a kind of multi-focus image fusing method more particularly to one are designed
Multi-focus image fusing method and system of the kind based on FGF.
Background technique
Since the target in focus can only be clearly imaged in picture plane by optical sensor imaging system, and focus
Target imaging outside point is fuzzy.Therefore, the limited problem of focusing range is easy to cause optical imaging system can not be to field
Scape target whole blur-free imaging.If understood completely entire scene objects, need to analyze a considerable amount of similar diagrams
Picture was not only wasted time but also was frittered away energy, and will also result in the waste on memory space.It is same that a width is obtained by image interfusion method
All objects all clearly images in one scene, make its more comprehensively, true reflection scene information is for accurate point of image
Analysis and understanding are of great significance, and multi-focus image fusion is one of the effective technology means for realizing this target.
Multi-focus image fusion is exactly to be obtained under the identical image-forming condition to process registration about more in a certain scene
Width focusedimage detects by activity measurement and extracts the clear area of every width focusedimage, so using certain blending algorithm
These region merging techniques are generated by all objects all clearly images in the width scene according to certain fusion rule afterwards.It is more
Focusedimage integration technology can be characterized extraction with the characterization scene objects information of complete display, and target identification and tracking etc. are established
Good basis is determined, so that the utilization rate and system that effectively improve image information are to the reliable of object table detection identification
Property, space-time unique is extended, uncertainty is reduced.
The key of Multi-focus image fusion is to make accurate characterization to focal zone characteristic, is accurately positioned and extracts
Region in focusing range or pixel out, this is also that asking of solving very well is not yet received in multi-focus image fusion technology so far
One of topic.Currently, image co-registration research continue for for more than 30 years.With the continuous development of computer and imaging technique, state
Inside and outside researcher determines for focal zone present in multi-focus image fusion technology and extracts problem, proposes several hundred kinds
The blending algorithm haveing excellent performance.These Multi-focus image fusions are broadly divided into two classes: spatial domain multi-focus image fusion is calculated
Method and transform domain Multi-focus image fusion.Wherein, spatial domain Image Fusion is according to the gray scale of pixel in source images
It is worth size, is come out the pixel of focal zone or extracted region using different focal zone evaluating characteristics, according to melts
Normally obtain blending image.The advantages of algorithm is that method is simple, is easily performed, and computation complexity is low, blending image packet
Raw information containing source images.The disadvantage is that being also easy to produce " blocking artifact " vulnerable to noise jamming.Transform domain image blending algorithm pair
Source images are converted, and are handled according to fusion rule transformation coefficient, and by treated, transformation coefficient progress inverse transformation is obtained
To blending image.Its shortcoming is mainly manifested in decomposable process complexity, time-consuming, and high frequency coefficient space hold is big, fusion process
In easily cause information to lose.If changing a transformation coefficient of blending image, the airspace gray value of whole image all will
It changes, as a result during enhancing some image-region attributes, introduces unnecessary artificial interference.It is more common
Pixel-level Multi-focus image fusion have it is following several:
(1) it is based on the multi-focus image fusing method of laplacian pyramid (Laplacian Pyramid, LAP).Its
Main process is to carry out Laplacian pyramid to source images, suitable fusion rule is then used, by high and low frequency
Coefficient is merged, and the progress inverse transformation of fused pyramid coefficient is obtained blending image.This method has good time-frequency
Local characteristics achieve good results, but each inter-layer data that decomposes has redundancy, can not determine the data phase on each decomposition layer
Guan Xing.It is poor to extract detailed information ability, decomposable process medium-high frequency information is lost seriously, and fused image quality is directly affected.
(2) it is based on the multi-focus image fusing method of wavelet transformation (Discrete Wavelet Transform, DWT).
Its main process is to carry out wavelet decomposition to source images, then uses suitable fusion rule, high and low frequency coefficient is carried out
Fused wavelet coefficient progress wavelet inverse transformation is obtained blending image by fusion.This method has good time-frequency part special
Property, it achieves good results, but 2-d wavelet base is made of by way of tensor product one-dimensional wavelet basis, for image
In the expression of singular point be optimal, but the line and face unusual for image can not carry out rarefaction representation.In addition DWT belongs to
It is converted in down-sampling, lacks translation invariance, the loss of information is easily caused in fusion process, blending image is caused to be distorted.
(3) contourlet transform based on non-lower sampling (Non-sub-sampled Contourlet Transform,
NSCT multi-focus image fusing method).Its main process is to carry out NSCT decomposition to source images, then uses and suitably melts
Normally, high and low frequency coefficient is merged, the progress NSCT inverse transformation of fused wavelet coefficient is obtained into fusion figure
Picture.This method can obtain good syncretizing effect, but the speed of service is slower, and decomposition coefficient needs to occupy a large amount of memory space.
(4) it is based on the multi-focus image fusion of principal component analysis (Principal Component Analysis, PCA)
Method.Its main process is source images to be preferentially converted into column vector according to row major or column, and calculate covariance, according to
Covariance matrix seeks feature vector, determines the corresponding feature vector of first principal component and determines therefrom that each source images fusion
Weight is weighted fusion according to weight.When this method has certain common characteristics between source images, it can obtain preferably
Syncretizing effect；And the feature difference between source images it is larger when, then be easy to introduce false information in blending image,
Fusion results are caused to be distorted.This method calculates simply, and speed is fast, but where the gray value of single pixel point can not indicate
The focus characteristics of image-region cause blending image soft edge, the low problem of contrast occur.
(5) it is based on the multi-focus image fusing method of spatial frequency (Spatial Frequency, SF).Its main process
It is that source images are subjected to block segmentation, then calculates each piece of SF, the SF of source images corresponding blocks is compared, by the big correspondence image of SF value
Merged block obtains blending image.This method is simply easy to implement, but piecemeal size is difficult to adaptive determination, and piecemeal is too big, Yi Jiang
Pixel outside focus is all included, and reduces fusion mass, is declined blending image contrast, is also easy to produce blocking artifact, piecemeal is too
It is small limited to region readability characterization ability, easily there is the wrong choice of block, so that consistency is poor between adjacent sub-blocks, is handing over
Occur obvious detail differences at boundary, generates " blocking artifact ".In addition, the focus characteristics of image subblock are difficult to accurate description, it is how sharp
With the focus characteristics of the image subblock local feature accurate description sub-block, will directly affect the accuracy for focusing sub-block selection and
The quality of blending image.
(6) multiple focussing image of convolution rarefaction representation (convolutional sparse representation, CSR)
Fusion method.Its main process is to carry out CSR decomposition to source images, the basal layer and levels of detail of source images is obtained, then to base
Plinth layer and levels of detail are merged, and finally merge the basal layer of fusion and levels of detail to obtain blending image.This method is not direct
Dependent on the focus characteristics of source images, but source images are determined by source images basal layer and the significant characteristics of levels of detail
Focal zone, to noise have robustness.
(7) multi-focus of (cartoon-texture decomposition, CTD) is decomposed based on cartoon-texture image
Image interfusion method.Its main process is that multi-focus source images are carried out with cartoon-texture image respectively to decompose, and obtains multi-focus
The cartoon ingredient and texture ingredient of source images, and the cartoon ingredient and texture ingredient of multi-focus source images are merged respectively,
Merge fused cartoon ingredient and texture ingredient obtains blending image.Its fusion rule be cartoon ingredient based on image and
The focus characteristics design of texture ingredient, the focus characteristics of source images are not directly dependent on, to have to noise and scratch breakage
There is robustness.
(8) multi-focus image fusing method based on Steerable filter (Guided Filter Fusion, GFF).It is main
Process be using guiding image filter by picture breakdown be the basal layer comprising large scale Strength Changes and comprising small scale it is thin
The levels of detail of section, then using the conspicuousness and Space Consistency of basal layer and levels of detail building blending weight figure, and as
Basis merges the basal layer of source images and levels of detail respectively, finally the basal layer of fusion and levels of detail is merged to obtain final
Blending image, this method can obtain good syncretizing effect, but lack robustness to noise.
Above-mentioned eight kinds of methods are more common multi-focus image fusing methods, but in these methods, wavelet transformation
(DWT) geometrical characteristic possessed by image data itself cannot be made full use of, cannot optimal or most " sparse " expression image,
Blending image is easily caused offset and information Loss occur；Contourlet transform (NSCT) method based on non-lower sampling due to
Decomposable process is complicated, and the speed of service is slower, and in addition decomposition coefficient needs to occupy a large amount of memory space.Principal component analysis (PCA)
Method is easily reduced blending image contrast, influences fused image quality.Convolution rarefaction representation (CSR), cartoon texture image point
Solution (CTD), Steerable filter (GFF) are all the new methods proposed in recent years, good syncretizing effect are all achieved, wherein being oriented to
Filtering (GFF) is that edge holding and translation invariant operation are carried out based on local nonlinearity model, and computational efficiency is high；It can use
While iteration frame restores large scale edge, the small details of adjacent edges is eliminated；Preceding four kinds of common fusion methods all exist
Different disadvantages, be difficult to reconcile between speed and fusion mass, limit the application and popularization of these methods, the 8th kind of method
It is the more excellent blending algorithm of current fusion performance, but Steerable filter is not filtered source images directly, is easy
Lost part source image information, at the same average weight integration technology with affect fusion performance to a certain extent.
In conclusion problem of the existing technology is:
In the prior art, (1) traditional Space domain mainly uses region partitioning method to carry out, region division size
It is excessive to will lead to exterior domain in focus and be located at the same area, cause fused image quality to decline；Region division is undersized, son
Provincial characteristics cannot sufficiently reflect the provincial characteristics, be easy to cause the judgement inaccuracy of focal zone pixel and generate and falsely drop, make
Consistency is poor between obtaining adjacent area, obvious detail differences occurs in intersection, generates " blocking artifact ".(2) traditional based on more rulers
It spends in the multi-focus image fusion method decomposed, is always handled whole picture multi-focus source images as single entirety, detailed information
It extracts imperfect, the detailed information such as source images Edge texture cannot be preferably indicated in blending image, affect blending image pair
The integrality of source images potential information description, and then influence fused image quality.
Summary of the invention
In view of the problems of the existing technology, the present invention provides one kind can effectively eliminate " blocking artifact ", expansion optical
The imaging system depth of field and can the subjective and objective quality of significant increase blending image multi-focus image fusing method based on FGF and be
System.It overcomes focal zone present in multi-focus image fusion and determines inaccuracy, it cannot effective extraction source image border texture
Information, blending image minutia characterize imperfect, part loss in detail, " blocking artifact ", the problems such as contrast decline.
(1) source images are carried out smoothly with mean filter, is decomposed to obtain the basal layer of source images and thin to source images
Ganglionic layer；(2) Laplce's filtering and Gassian low-pass filter are successively filtered source images, obtain the aobvious of source images
Write figure；(3) the weight figure of corresponding source images is obtained by comparing source images notable figure pixel size；(4) using source images as drawing
It leads image and DECOMPOSED OPTIMIZATION is carried out to weight figure using FGF, respectively obtain the weight figure basal layer and levels of detail of optimization；(5) root
Basal layer and levels of detail respective pixel are melted using the weight figure basal layer and levels of detail of optimization according to certain fusion rule
It closes；(6) fused basal layer and levels of detail are merged, obtains blending image.
The invention is realized in this way basal layer and levels of detail will be decomposed into source images with mean filter first；Then
Conspicuousness detection is carried out to source images using Laplce's filtering and Gassian low-pass filter, obtains the notable figure of source images；Then
The weight figure of corresponding source images is obtained by comparing source images notable figure pixel size；And simultaneously using source images as navigational figure
DECOMPOSED OPTIMIZATION is carried out to weight figure using FGF, respectively obtains the weight figure basal layer and levels of detail of optimization；It is then based on decision
Matrix respectively merges basal layer and levels of detail respective pixel according to certain fusion rule；Finally by fused basal layer
Merge with levels of detail, obtains blending image.
Further, the multi-focus image fusing method based on FGF, to the multiple focussing image I after registration1And I2It carries out
Fusion, I1And I2It is gray level image, and I1, I2∈□M×N,It is the space that size is M × N, M and N are positive integer,
It specifically includes:
(1) using mean value wave device AF respectively to multiple focussing image I1And I2Smooth operation is carried out, source images I is removed1And I2
In small structure, obtain source images basal layer (B1, B2), source images levels of detail (D1, D2).Wherein: (B1, B2)=AF (I1, I2),
(D1, D2)=(I1, I2)-(B1, B2)。
(2) source images are filtered with LF, obtain the high-pass filtering image H of source images1And H2, with GLF to H1
And H2Low-pass filtering treatment obtains source images notable figure S1And S2.Wherein: (H1, H2)=LF (I1, I2), (S1, S2)=GLF (H1,
H2)。
(3) according to source images I1And I2Pixel S in corresponding notable figure1(i, j) and S2(i, j) size constructs source images pair
The weight matrix P answered1And P2.Wherein:
S1(i, j) is source images I1Notable figure pixel (i, j)；
S2(i, j) is source images I2Notable figure pixel (i, j)；
P1(i, j) is source images I1Weight matrix element (i, j)；
P2(i, j) is source images I2Weight matrix element (i, j)；
I=1,2,3 ..., M；J=1,2,3 ..., N；
S (i, j) is the element of the i-th row of matrix notable figure S, jth column；
(4) by source images I1And I2As navigational figure, with FGF to weight matrix P1And P2DECOMPOSED OPTIMIZATION is carried out, is obtained
Weight matrix W1 B, W2 B, W1 DAnd W2 D.Wherein: (W1 B, W1 D)=FGF (P1, I1), (W2 B, W2 D)=FGF (P2, I2)。
(5) it is based on source images basal layer (B1, B2) and levels of detail (D1, D2), according to the weight matrix W of optimization1 B, W2 B, W1 D
And W2 DConstruct blending image basal layer FB,With levels of detail FD,, obtain fused basal layer FB
With levels of detail FD.Wherein, FB=W1 BB1+W2 BB2, FD=W1 DD1+W2 DD2。
(6) blending image F is constructed,Obtain fused gray level image, in which: F=FB+FD。
Further, corrosion expansive working processing is carried out to the eigenmatrix constructed in step (4), and using treated
Eigenmatrix constructs blending image.
The multi-focus image fusion system based on FGF that another object of the present invention is to provide a kind of.
Another object of the present invention is to provide a kind of intelligence using the above-mentioned multi-focus image fusing method based on FGF
Intelligent city multi-focus image fusion system.
Another object of the present invention is to provide a kind of doctors using the above-mentioned multi-focus image fusing method based on FGF
Treat imaging multi-focus image fusion system.
Another object of the present invention is to provide a kind of peaces using the above-mentioned multi-focus image fusing method based on FGF
Full monitoring multi-focus image fusion system.
Advantages of the present invention and good effect are as follows:
(1) the invention firstly uses mean filters, and basal layer and levels of detail will be decomposed into source images, then general using drawing
Lars high-pass filtering and Gassian low-pass filter carry out conspicuousness detection to source images, obtain the notable figure of source images；Then pass through
Comparison source images notable figure pixel size obtains the weight figure of corresponding source images；And source images as navigational figure and are utilized
FGF carries out DECOMPOSED OPTIMIZATION to weight figure, respectively obtains the weight figure basal layer and levels of detail of optimization, utilizes the basis of weight figure
Layer and levels of detail respectively merge source images basal layer and levels of detail, then melt fused basal layer and levels of detail
Conjunction obtains the blending image of source images.Secondary fusion is carried out to source images, is improved to the judgement of source images focal zone characteristic
Accuracy rate is conducive to the extraction of clear area target, can have preferably from detailed information such as source images transfer Edge textures
Effect improves the subjective and objective quality of blending image.
(2) in the present invention, image co-registration frame is flexible, easy to implement, can be used for other kinds of image co-registration task.
(3) when this blending algorithm carries out smooth operation to source images with mean filter, can effectively inhibit in source images
Influence of the noise to fused image quality.
Image interfusion method frame of the present invention is flexible, determines accuracy rate with higher to source images focal zone characteristic,
Focal zone target detail can be accurately extracted, it is clear to indicate image detail feature, while " blocking artifact " is effectively eliminated,
Effectively improve the subjective and objective quality of blending image.
Detailed description of the invention
Fig. 1 is the multi-focus image fusing method flow chart based on FGF that case study on implementation of the present invention provides.
Fig. 2 is source images to be fused ' Disk ' effect picture that case study on implementation 1 of the present invention provides.
Fig. 3 is that case study on implementation offer of the present invention is Laplce (LAP), wavelet transformation (DWT), based on non-lower sampling
Contourlet transform (NSCT), principal component analysis (PCA) method, spatial frequency (SF), convolution rarefaction representation (CSR), cartoon line
Manage picture breakdown (CTD), Steerable filter (GFF) and of the invention (Proposed) totally nine kinds of image interfusion methods to multi-focus
The syncretizing effect figure of image ' Disk ' Fig. 3 (a) and (b).
Fig. 4 is image to be fused ' Book ' effect picture that case study on implementation 2 of the present invention provides 2；
Fig. 5 be Laplce (LAP), wavelet transformation (DWT), the contourlet transform (NSCT) based on non-lower sampling, it is main at
Analysis (PCA) method, spatial frequency (SF), convolution rarefaction representation (CSR), cartoon texture image decompose (CTD), guiding filter
Wave (GFF) and nine kinds of fusion methods of (Proposed) image of the invention melt multiple focussing image ' Book ' Fig. 4 (a) with (b)
Close effect image.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with case study on implementation, to this
Invention is further elaborated.It should be appreciated that specific implementation case described herein is only used to explain the present invention,
It is not intended to limit the present invention.
In the prior art, blending algorithm determines inaccuracy to source images focal zone in multi-focus image fusion field, carefully
It is imperfect to save information extraction, the detailed information such as source images Edge texture, syncretizing effect cannot be preferably indicated in blending image
Difference.
Application principle of the invention is described in detail with reference to the accompanying drawing.
As shown in Figure 1, the multi-focus image fusing method based on FGF that case study on implementation of the present invention provides, comprising:
S101: basal layer and levels of detail will be decomposed into source images first with mean filter.
S102: and then conspicuousness detection is carried out to source images using Laplce's high-pass filtering and Gassian low-pass filter, it obtains
The weight figure of corresponding source images is obtained to the notable figure of source images, and by comparing corresponding source image saliency map pixel size.
S103: DECOMPOSED OPTIMIZATION is carried out to weight figure using source images as navigational figure and using FGF, respectively obtains optimization
Weight figure basal layer and levels of detail, and using weight figure basal layer and levels of detail respectively by source images basal layer and details
Layer fusion.
S104: finally fused basal layer and levels of detail are merged, obtain blending image.
Below with reference to detailed process, the invention will be further described.
The multi-focus image fusing method based on FGF that case study on implementation of the present invention provides, detailed process include:
Using mean value wave device AF respectively to multiple focussing image I1And I2Smooth operation is carried out, source images I is removed1And I2In
Small structure obtains source images basal layer (B1, B2), source images levels of detail (D1, D2).Wherein: (B1, B2)=AF (I1, I2), (D1,
D2)=(I1, I2)-(B1, B2)；
Source images are filtered with LF, obtain the high-pass filtering image H of source images1And H2, with GLF to H1And H2
Low-pass filtering treatment obtains source images notable figure S1And S2.Wherein: (H1, H2)=LF (I1, I2), (S1, S2)=GLF (H1, H2)。
And according to source images I1And I2Pixel S in corresponding notable figure1(i, j) and S2(i, j) size, the corresponding weight square of building source images
Battle array P1And P2.Wherein:
S1(i, j) is source images I1Notable figure pixel (i, j)；
S2(i, j) is source images I2Notable figure pixel (i, j)；
P1(i, j) is source images I1Weight matrix element (i, j)；
P2(i, j) is source images I2Weight matrix element (i, j)；
I=1,2,3 ..., M；J=1,2,3 ..., N；
S (i, j) is the element of the i-th row of matrix notable figure S, jth column；
Respectively by source images I1And I2As navigational figure, with FGF to weight matrix P1And P2DECOMPOSED OPTIMIZATION is carried out, is obtained
Weight matrix W1 B, W2 B, W1 DAnd W2 D.Wherein: (W1 B, W1 D)=FGF (P1, I1), (W2 B, W2 D)=FGF (P2, I2)。
Based on source images basal layer (B1, B2) and levels of detail (D1, D2), according to the weight matrix W of optimization1 B, W2 B, W1 DWith
W2 DConstruct blending image basal layer FB,With levels of detail FD,, obtain fused basal layer FBWith
Levels of detail FD.Wherein, FB=W1 BB1+W2 BB2, FD=W1 DD1+W2 DD2。
Blending image F is constructed,Obtain fused gray level image, in which: F=FB+FD。
Below with reference to specific implementation case, the invention will be further described.
Fig. 2 is source images to be fused ' Disk ' effect picture that case study on implementation 1 of the present invention provides.
Case study on implementation 1
The solution of the present invention is followed, which carries out fusion treatment to two width source images shown in Fig. 2 (a) and (b),
Processing result is as shown in the Propose in Fig. 3.Simultaneously using Laplce (LAP), wavelet transformation (DWT), based on adopting under non-
The contourlet transform (NSCT) of sample, principal component analysis (PCA) method, spatial frequency (SF), convolution rarefaction representation (CSR), cartoon
Texture image decompose (CTD), eight kinds of image interfusion methods of Steerable filter (GFF) to two width source images shown in Fig. 2 (a) and (b) into
Row fusion treatment carries out quality evaluation to the blending images of different fusion methods, and processing calculates to obtain result shown in table 1.
Table 1 multiple focussing image ' Disk ' fused image quality evaluates
Case study on implementation 2:
The solution of the present invention is followed, which carries out fusion treatment to two width source images shown in Fig. 4 (a) and (b),
Processing result is as shown in the Proposed in Fig. 5.
Simultaneously Laplce (LAP), wavelet transformation (DWT), the contourlet transform (NSCT) based on non-lower sampling, it is main at
Analysis (PCA) method, spatial frequency (SF), convolution rarefaction representation (CSR), cartoon texture image decompose (CTD), guiding filter
Eight kinds of image interfusion methods of wave (GFF) carry out fusion treatment to two width source images (a) shown in Fig. 4 and (b), merge to Fig. 5 difference
The blending image of method carries out quality evaluation, and processing calculates to obtain result shown in table 2.
Table 2 multiple focussing image ' Book ' fused image quality evaluates
In Tables 1 and 2: Method represents method；Fusion method includes eight kinds: Laplce (LAP), small echo
Convert (DWT), the contourlet transform (NSCT) based on non-lower sampling, principal component analysis (PCA) method, spatial frequency (SF), volume
Product rarefaction representation (CSR), cartoon texture image decompose (CTD), Steerable filter (GFF)；When Running Time represents operation
Between, unit is the second.MI represents mutual information, is that fused image quality based on mutual information objectively evaluates index.QAB/FIt represents from source
The marginal information total amount shifted in image.
From Fig. 3, Fig. 5 can be seen that other method frequency domain methods include Laplce (LAP), wavelet transformation (DWT),
The problem of contourlet transform (NSCT) based on non-lower sampling, blending image all deposit artifact again, fuzzy and poor contrast；
Its blending image contrast of principal component analysis (PCA) method is worst in the method for airspace, the fusion figure of spatial frequency (SF) method
As there is " blocking artifact " phenomenon, and convolution rarefaction representation (CSR), cartoon texture image decompose (CTD), Steerable filter (GFF) melts
It is relatively preferable to close quality, but there is also a small amount of obscure portions.Method of the invention is to multiple focussing image Fig. 3 ' Disk ' and poly
The blending image subjective vision effect of burnt image graph 5 ' Book ' is substantially better than the syncretizing effect of other fusion methods.
It can be seen that from blending image, extractability of the method for the present invention to source images focus area object edge and texture
Other methods are substantially better than, can be good at for the target information of focus area in source images being transferred in blending image, are protected
Deposit the detailed information such as edge and the texture in source images.The target detail information of focal zone can be effectively captured, image is improved
Fusion mass.The method of the present invention has good subjective attribute.
As can be seen from Table 1 and Table 2, the picture quality of the method for the present invention blending image objectively evaluates index MI than other
The blending image of method corresponds to index and is averagely higher by 1.5, and the picture quality of blending image objectively evaluates index QAB/FThan its other party
The blending image of method corresponds to index and is higher by 0.04.Illustrate that this method obtains blending image and has good objective figures.
The foregoing is merely preferable case study on implementation of the invention, are not intended to limit the invention, all of the invention
Made any modifications, equivalent replacements, and improvements etc. within spirit and principle, should be included in protection scope of the present invention it
It is interior.
Claims (7)
1. a kind of multi-focus image fusing method based on FGF, which is characterized in that the multi-focus image fusion based on FGF
Method and system the following steps are included:
(1) source images are decomposed with mean filter (Average Filtering, AF), respectively obtains the basis of source images
Layer and levels of detail；
(2) Laplce is filtered into (Laplacian Filter, LF) and Gassian low-pass filter (Gaussian Low-pass
Filter, GLF) successively source images are filtered, obtain the notable figure of source images；
(3) the weight figure that notable figure pixel size constructs corresponding source images is corresponded to according to source images, schemed source images as guidance
Picture simultaneously carries out DECOMPOSED OPTIMIZATION to weight figure using FGF, respectively obtains the weight figure basal layer and levels of detail of optimization；
(4) basal layer and levels of detail based on weight figure, according to certain fusion rule respectively by source images basal layer and details
Layer respective pixel fusion；
(5) fused basal layer and levels of detail are merged, obtains blending image.
2. as described in claim 1 based on the multi-focus image fusing method of FGF, which is characterized in that described based on the more of FGF
Focusedimage fusion method, to the multiple focussing image I after registration1And I2It is merged, I1And I2It is gray level image, and I1, I2
∈□M×N,It is the space that size is M × N, M and N are positive integer, it specifically includes:
(1) using mean value wave device AF respectively to multiple focussing image I1And I2Smooth operation is carried out, source images I is removed1And I2In it is small
Structure obtains source images basal layer (B1, B2), source images levels of detail (D1, D2).Wherein: (B1, B2)=AF (I1, I2), (D1, D2)
=(I1, I2)-(B1, B2)。
(2) source images are filtered with LF, obtain the high-pass filtering image H of source images1And H2, with GLF to H1And H2It is low
Pass filter handles to obtain source images notable figure S1And S2.Wherein: (H1, H2)=LF (I1, I2), (S1, S2)=GLF (H1, H2)。
(3) according to source images I1And I2Pixel S in corresponding notable figure1(i, j) and S2(i, j) size, the corresponding power of building source images
Value matrix P1And P2.Wherein:
S1(i, j) is source images I1Notable figure pixel (i, j)；
S2(i, j) is source images I2Notable figure pixel (i, j)；
P1(i, j) is source images I1Weight matrix element (i, j)；
P2(i, j) is source images I2Weight matrix element (i, j)；
I=1,2,3 ..., M；J=1,2,3 ..., N；
S (i, j) is the element of the i-th row of matrix notable figure S, jth column；
(4) by source images I1And I2As navigational figure, with FGF to weight matrix P1And P2DECOMPOSED OPTIMIZATION is carried out, weight square is obtained
Battle array W1 B, W2 B, W1 DAnd W2 D.Wherein: (W1 B, W1 D)=FGF (P1, I1), (W2 B, W2 D)=FGF (P2, I2)。
(5) it is based on source images basal layer (B1, B2) and levels of detail (D1, D2), according to the weight matrix W of optimization1 B, W2 B, W1 DAnd W2 D
Construct blending image basal layer FB,With levels of detail FD,Obtain fused basal layer FBWith it is thin
Ganglionic layer FD.Wherein, FB=W1 BB1+W2 BB2, FD=W1 DD1+W2 DD2。
(6) blending image F is constructed,Obtain fused gray level image, in which: F=FB+FD。
3. as claimed in claim 2 based on the multi-focus image fusing method of FGF, which is characterized in that building in step (4)
Weight matrix optimize resolution process, and using the basal layer and levels of detail of treated weight matrix fusion source images,
And then construct blending image.
4. a kind of multi-focus image fusion based on FGF of the multi-focus image fusing method based on FGF as described in claim 1
System.
5. a kind of melted using the smart city multiple focussing image of the multi-focus image fusing method described in claim 1 based on FGF
Collaboration system.
6. a kind of melted using the imaging of medical multiple focussing image of the multi-focus image fusing method described in claim 1 based on FGF
Collaboration system.
7. a kind of melted using the security monitoring multiple focussing image of the multi-focus image fusing method described in claim 1 based on FGF
Collaboration system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811194833.0A CN109509163B (en) | 2018-09-28 | 2018-09-28 | A method and system for multi-focus image fusion based on FGF |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811194833.0A CN109509163B (en) | 2018-09-28 | 2018-09-28 | A method and system for multi-focus image fusion based on FGF |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109509163A true CN109509163A (en) | 2019-03-22 |
| CN109509163B CN109509163B (en) | 2022-11-11 |
Family
ID=65746461
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811194833.0A Active CN109509163B (en) | 2018-09-28 | 2018-09-28 | A method and system for multi-focus image fusion based on FGF |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109509163B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110211081A (en) * | 2019-05-24 | 2019-09-06 | 南昌航空大学 | A kind of multi-modality medical image fusion method based on image attributes and guiding filtering |
| CN110648302A (en) * | 2019-10-08 | 2020-01-03 | 太原科技大学 | A light field all-focus image fusion method based on edge enhancement guided filtering |
| CN111223069A (en) * | 2020-01-14 | 2020-06-02 | 天津工业大学 | An image fusion method and system |
| CN111815549A (en) * | 2020-07-09 | 2020-10-23 | 湖南大学 | A night vision image colorization method based on guided filter image fusion |
| CN112884690A (en) * | 2021-02-26 | 2021-06-01 | 中国科学院西安光学精密机械研究所 | Infrared and visible light image fusion method based on three-scale decomposition |
| CN117391985A (en) * | 2023-12-11 | 2024-01-12 | 安徽数分智能科技有限公司 | Multi-source data information fusion processing method and system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011051134A1 (en) * | 2009-10-30 | 2011-05-05 | Siemens Aktiengesellschaft | A body fluid analyzing system and an imaging processing device and method for analyzing body fluids |
| CN103455991A (en) * | 2013-08-22 | 2013-12-18 | 西北大学 | Multi-focus image fusion method |
| CN104036479A (en) * | 2013-11-11 | 2014-09-10 | 西北大学 | Multi-focus image fusion method based on non-negative matrix factorization |
| CN107909560A (en) * | 2017-09-22 | 2018-04-13 | 洛阳师范学院 | A kind of multi-focus image fusing method and system based on SiR |
| CN108230282A (en) * | 2017-11-24 | 2018-06-29 | 洛阳师范学院 | A kind of multi-focus image fusing method and system based on AGF |
-
2018
- 2018-09-28 CN CN201811194833.0A patent/CN109509163B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011051134A1 (en) * | 2009-10-30 | 2011-05-05 | Siemens Aktiengesellschaft | A body fluid analyzing system and an imaging processing device and method for analyzing body fluids |
| CN103455991A (en) * | 2013-08-22 | 2013-12-18 | 西北大学 | Multi-focus image fusion method |
| CN104036479A (en) * | 2013-11-11 | 2014-09-10 | 西北大学 | Multi-focus image fusion method based on non-negative matrix factorization |
| CN107909560A (en) * | 2017-09-22 | 2018-04-13 | 洛阳师范学院 | A kind of multi-focus image fusing method and system based on SiR |
| CN108230282A (en) * | 2017-11-24 | 2018-06-29 | 洛阳师范学院 | A kind of multi-focus image fusing method and system based on AGF |
Non-Patent Citations (2)
| Title |
|---|
| 刘先红等: "结合引导滤波和卷积稀疏表示的红外与可见光图像融合", 《光学精密工程》 * |
| 刘帅奇等: "结合向导滤波与复轮廓波变换的多聚焦图像融合算法", 《信号处理》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110211081A (en) * | 2019-05-24 | 2019-09-06 | 南昌航空大学 | A kind of multi-modality medical image fusion method based on image attributes and guiding filtering |
| CN110648302A (en) * | 2019-10-08 | 2020-01-03 | 太原科技大学 | A light field all-focus image fusion method based on edge enhancement guided filtering |
| CN110648302B (en) * | 2019-10-08 | 2022-04-12 | 太原科技大学 | Light field full-focus image fusion method based on edge enhancement guide filtering |
| CN111223069A (en) * | 2020-01-14 | 2020-06-02 | 天津工业大学 | An image fusion method and system |
| CN111223069B (en) * | 2020-01-14 | 2023-06-02 | 天津工业大学 | A method and system for image fusion |
| CN111815549A (en) * | 2020-07-09 | 2020-10-23 | 湖南大学 | A night vision image colorization method based on guided filter image fusion |
| CN112884690A (en) * | 2021-02-26 | 2021-06-01 | 中国科学院西安光学精密机械研究所 | Infrared and visible light image fusion method based on three-scale decomposition |
| CN112884690B (en) * | 2021-02-26 | 2023-01-06 | 中国科学院西安光学精密机械研究所 | A Fusion Method of Infrared and Visible Light Images Based on Three Scale Decomposition |
| CN117391985A (en) * | 2023-12-11 | 2024-01-12 | 安徽数分智能科技有限公司 | Multi-source data information fusion processing method and system |
| CN117391985B (en) * | 2023-12-11 | 2024-02-20 | 安徽数分智能科技有限公司 | Multi-source data information fusion processing method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109509163B (en) | 2022-11-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109509163A (en) | A kind of multi-focus image fusing method and system based on FGF | |
| Du et al. | Image segmentation-based multi-focus image fusion through multi-scale convolutional neural network | |
| CN106339998B (en) | Multi-focus image fusion method based on contrast pyramid transformation | |
| CN108399611B (en) | Multi-focus image fusion method based on gradient regularization | |
| CN110334779B (en) | Multi-focus image fusion method based on PSPNet detail extraction | |
| CN109509164B (en) | Multi-sensor image fusion method and system based on GDGF | |
| CN107909560A (en) | A kind of multi-focus image fusing method and system based on SiR | |
| CN106886977A (en) | A kind of many figure autoregistrations and anastomosing and splicing method | |
| CN108230282A (en) | A kind of multi-focus image fusing method and system based on AGF | |
| CN103295221B (en) | The waterborne target method for testing motion of simulation compound eye visual mechanism and polarization imaging | |
| Du et al. | Multi-focus image fusion using deep support value convolutional neural network | |
| CN111462027B (en) | Multi-focus image fusion method based on multi-scale gradient and matting | |
| Xiao et al. | Image Fusion | |
| Zhuang et al. | A novel 3D position measurement and structure prediction method for RFID tag group based on deep belief network | |
| Zhang et al. | Medical Image Fusion Based on Low‐Level Features | |
| CN118397074B (en) | Fish target length detection method based on binocular vision | |
| Liu et al. | Multi-level wavelet network based on CNN-Transformer hybrid attention for single image deraining | |
| CN109242797A (en) | Image de-noising method, system and the medium merged based on homogeneous and heterogeneous areas | |
| Yan et al. | Multiscale fusion and aggregation pcnn for 3d shape recovery | |
| Ayub et al. | CNN and Gaussian Pyramid-Based Approach For Enhance Multi-Focus Image Fusion | |
| Li et al. | An improved image registration and fusion algorithm | |
| Zhao et al. | Multi-scale fusion algorithm of intensity and polarization-difference images based on edge information enhancement | |
| Shiralkar et al. | A self-organization based optical flow estimator with GPU implementation | |
| Yang et al. | A new multi‐focus image fusion algorithm based on BEMD and improved local energy | |
| CN113409214A (en) | Rapid repairing method for damaged fingerprint image |
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 |