CN109447930A - Wavelet field light field total focus image generation algorithm - Google Patents
Wavelet field light field total focus image generation algorithm Download PDFInfo
- Publication number
- CN109447930A CN109447930A CN201811259275.1A CN201811259275A CN109447930A CN 109447930 A CN109447930 A CN 109447930A CN 201811259275 A CN201811259275 A CN 201811259275A CN 109447930 A CN109447930 A CN 109447930A
- Authority
- CN
- China
- Prior art keywords
- image
- total focus
- light field
- wavelet
- focus image
- 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 by the use of more than one image, e.g. averaging, subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10052—Images from lightfield camera
-
- 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
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Abstract
Wavelet field light field total focus image generation algorithm of the invention belongs to total focus image co-registration field, the present invention effectively avoids the blocking artifact of traditional airspace light field image blending algorithm, obtain the light field total focus image of better quality, spatial alternation and projection are carried out by the 4D light field data obtained to microlens array light-field camera, obtain the multiple focussing image for total focus image co-registration, wavelet decomposition is carried out to each frame multiple focussing image and extracts height, low frequency subgraph image set, it is proposed Region homogenization Laplace operator, pixel visibility function constructs the height of blending image respectively, low-frequency wavelet coefficients realize image co-registration, its performance is better than traditional region sharpness evaluation function, the experimental verification correctness and validity of mentioned method of the invention, fusion total focus image is calculated using the initial data of Lytro light-field camera, with tradition Image Fusion is compared, and human eye vision effect is more preferable, and objective image index is also improved.
Description
Technical field
The invention belongs to total focus image co-registration fields, generate and calculate more particularly to a kind of wavelet field light field total focus image
Method.
Background technique
Along with the development for the rise and optical field imaging theory for calculating photography new discipline field, light-field camera becomes close
The hot spot of domestic and international numerous areas concern over 10 years.Relative to traditional camera, lenticule light-field camera joined after main lens
Microlens array records the position and direction information of space light simultaneously.Various dimensions field information is recorded as the light-field camera later period
The processing and application of image are provided convenience, and light-field camera number refocusing, the generation of total focus image and depth information are such as utilized
Calculate etc..Light-field camera can calculate the refocusing image of any depth in space after single shot, and it is general to be that light-field camera is able to
All over the most prominent technological highlights of concern.Based on this, all kinds of light field high quality texture images are obtained and high accuracy depth information calculates
It is furtherd investigate.It is limited since shooting of repeatedly focusing not by traditional camera obtains multiple focussing image, is based on light field number weight
The total focus image co-registration of focusing technology becomes an important branch application of light-field camera, for later period texture image and depth
The generation for spending the super-resolution reconstruction, light field video file of image is also of great significance.
Currently, the total focus fusion for traditional images is broadly divided into spatial domain and transform domain, spatial domain be based on pixel or
Block carries out clarity evaluation, extracts high-quality pixel from different images constituting total focus image, calculating the time fastly but depositing
In blocking artifact problem.Picture breakdown is different resolution layer or the subgraph of different frequency bands by transform domain, passes through evaluation reconstruction point
It distinguishes layer or subgraph building refocusing image, can effectively avoid blocking artifact.As a kind of common method of transform domain, wavelet transformation
By picture breakdown to be fused into a series of frequency channels, high and low frequency subgraph is constructed using its pyramidal structure after decomposing, point
It is other high and low frequency subgraph is merged after through inverse wavelet transform obtain total focus image.The quality of Wavelet Transform Fusion image
It is decided by the fusion rule selection of high and low frequency subgraph: for low frequency subgraph picture, fusion is generally realized using mean value computation method；
For high frequency subgraph, evaluation is carried out frequently with Sobel operator, Prewitt operator and Laplace operator etc. and establishes fusion rule
Then.
Most probably there is the opposite situation of symbol in the second dervative in the direction x and the direction y in traditional Laplace operator, and
And the anti-noise ability of existing Laplace operator algorithm is low, lenticule calibrated error can cause refocusing image to generate local make an uproar
Sound.In addition, the existing diffusion-weighted method of average of low frequency signal can reduce the contrast of blending image and lose some in original image
Useful information.
Summary of the invention
The present invention is not high for existing civil light-field camera shooting picture contrast, through the acquisition of digital refocusing technology
Multiple focussing image collection limited resolution and there are problems that the local noise as caused by calibrated error, it is desirable to provide one kind is based on
Wavelet field light field total focus image generation algorithm, this algorithm is by establishing high frequency coefficient Region homogenization Laplace operator, low frequency
The sharpness evaluation function of coefficient pixel visibility function realizes that light field total focus image generates, and effectively realizes light field original number
According to the conversion for arriving total focus image, and image co-registration quality increases compared with traditional algorithm.
In order to solve the above technical problems, the technical solution adopted by the present invention are as follows: wavelet field light field total focus image, which generates, to be calculated
Method is realized according to the following steps:
Step 1): obtaining 4D light field for light field original image after data decode, and selects different αn(n=1,2,3 ...) is utilized
Digital refocusing technology obtains the refocusing image of different spaces depth
Step 2) calculates the small echo high and low frequency subgraph of each frame refocusing image
Region homogenization Laplce BL operator and pixel visibility PV function is respectively adopted to high and low frequency subgraph in step 3)
As image co-registration clarity evaluation index, the fusion of high and low frequency coefficient is realized；
Low-and high-frequency coefficient is obtained fused total focus image by wavelet inverse transformation by step 4).
Further, the BL operator used in step 3) is Region homogenization Laplace operator, and the expression formula of the operator is such as
Under:
Wherein, wherein S × T indicates equalization region size, and S, T can only take odd number；S, t indicates horizontal vertical direction second order
Lead step-length；Indicate weight factor, the closer point of distance center point, weight factor is bigger, to Laplace operator
Value contribution is bigger, conversely, distance center point is remoter, contributes Laplace operator value smaller.
Further, PV function is pixel visibility function in step 3), and expression is as follows:
Wherein, S × T indicates the rectangular neighborhood centered on current pixel point, and S, T can only take odd number；S, t is indicated in square
The scanning step in horizontal vertical direction in shape neighborhood；Indicate the average gray value of S × T area pixel.
Further, low frequency coefficient is identical with high frequency coefficient fusion rule, and by taking high frequency coefficient merges as an example, rule is as follows:
Wherein,Represent each refocusing image of different spaces depth respective high frequency subgraph after wavelet decomposition, n=1,
2,3 ... N, N indicate to participate in the refocusing number of image frames of total focus image co-registration；Represent any 2 panel height frequency subgraph
As the difference of the balanced Laplace operator of corresponding points；Max [], min [] be maximized, minimum Value Operations；HHFor certainly
Define threshold value (HH0.1 is taken, because indicating the two Region homogenization Laplacian values difference when the two difference is less than 0.1 very
It is small to can be ignored), when the minimum value of difference is greater than threshold value, take balanced Laplce's energy in N frame image maximum
High frequency coefficient corresponding to person is multiplied when the two difference is less than threshold value by multiple image high frequency coefficient as fusion coefficients
Last fusion coefficients are determined with weight factor, wherein weight factor
The present invention carries out the fusion of image using the method for wavelet transformation.4D light field is decoded first and is met again using number
Burnt algorithm obtains the multiple focussing image of different depth, by carrying out wavelet decomposition and tower reconstruct building to each multiple focussing image collection
High and low frequency subgraph image set finally proposes that Region homogenization Laplace operator, pixel visibility function construct blending image respectively
High and low frequency wavelet coefficient realizes image co-registration.This algorithm effectively realizes the conversion of light field initial data to total focus image, has
Effect avoids the blocking artifact of traditional spatial domain picture blending algorithm, obtains the light field total focus image of better quality, image co-registration quality
It increases compared with traditional algorithm.
Detailed description of the invention
The present invention will be further described in detail with reference to the accompanying drawing.
Fig. 1 is the flow chart of inventive algorithm.
Fig. 2 is light field biplane parameterized model.
Fig. 3 is light-field camera number refocusing schematic diagram.
Fig. 4 is BL operator schematic diagram.
Fig. 5 is the fusion process demonstration graph of Leaves sample image.
Fig. 6 is Flower sample image difference blending algorithm comparison diagram.
Fig. 7 is Forest sample image difference blending algorithm comparison diagram.
Fig. 8 is Zither sample image difference blending algorithm comparison diagram.
Specific embodiment
It is understandable to enable objects, features and advantages of the present invention to become apparent, with reference to the accompanying drawing to tool of the invention
Body embodiment is described in detail.
As shown in Figure 1, the detailed process of this algorithm are as follows:
Step 1) light field original image obtains 4D light field after data decode, and selects different αn(n=1,2,3 ...), utilizes number
Word refocusing technology obtains the refocusing image of different spaces depth
Step 2) calculates the small echo high and low frequency subgraph of each frame refocusing image
BL operator and PV function is respectively adopted as image co-registration clarity evaluation index in step 3) high and low frequency subgraph,
Realize the fusion of low-and high-frequency coefficient；
Step 4) most obtains fused total focus image through wavelet inverse transformation afterwards.
Detailed process is as follows for step 1): according to the biplane parameterized model of light field, as shown in Fig. 2, any light in space
Line can be used the intersection point of itself and two planes to determine, if the main lens plane of light-field camera is the face (u, v), sensor plane be (x,
Y) face, the 4D light field of light-field camera record are LFFull light camera focal plane can be obtained by classical light radiation formula in (x, y, u, v)
Integral image:
Wherein F indicates the distance between main lens plane and focal plane, indicates 4D light field matrix L with X × Y × U × VF(x,y,
U, v) size.If by F F ', new 4D light field matrix L will be moved to as planeF′(x ', y ', u ', v ') is indicated, this phase
Machine focal plane refocusing image is expressed as:
Enable F '=αnF takes a section in the space 4D in order to facilitate graphical representation to obtain the geometry between coordinate and close
System, as shown in Figure 3.According to similar triangle theory, the coordinate that new light field and original light field can be obtained meets:
X '=u+ (x-u) αn=αn·x+(1-αn)·u (3)
U '=u (4)
Similarly it can be obtained:
Y '=v+ (y-v) αn=αn·y+(1-αn)·v (5)
V '=v (6)
Formula (3)-(6) are represented by matrix form:
Wherein, [x ', y ', u ', v ']TIndicate the transposition of row vector [x ', y ', u ', v '],Indicates coordinate transformation matrix,
Concrete form is as follows:
Formula (7) is also equivalent to following formula:
According to formula (9), formula (2) be can be rewritten as:
Change αnValue, that is, can reach change as plane position purpose, then obtain the weight of different spaces depth
Focus picture.
Detailed process is as follows for step 2): according to wavelet image blending theory, by wavelet transformation by figure to be fused
As decomposing in a series of frequency channels, high and low frequency subgraph is constructed using its pyramidal structure after decomposing, which can describe
Are as follows:
Wherein (x, y) indicates that image coordinate system, (i, j) indicate that wavelet field coordinate system, W [] indicate that small echo tower decomposes behaviour
It accords with, WH[] indicates to extract high frequency coefficient (high frequency subgraph) after small echo tower decomposes, WLAfter [] indicates that small echo tower decomposes
It extracts low frequency coefficient (low frequency subgraph picture).
For the BL operator used in step 3) for Region homogenization Laplace operator, the expression formula of the operator is as follows:
Wherein, wherein S × T indicates equalization region size, and S, T can only take odd number；S, t indicates horizontal vertical direction second order
Lead step-length；Indicate weight factor, the closer point of distance center point, weight factor is bigger, to Laplace operator
Value contribution is bigger, conversely, distance center point is remoter, contributes Laplace operator value smaller.It is equal when Fig. 4 is S=5, T=5
Weigh Laplace operator.
The jump in brightness characteristic of the high frequency subgraph reflection image of wavelet transformation, i.e. borderline properties, Laplace operator energy
The boundary and lines of any trend are sharpened, and keep isotropic characteristics, is commented carrying out clarity to high frequency subgraph
It is widely used when valence.Most probably there are the opposite feelings of symbol in the second dervative in the direction x and the direction y for Laplace operator
Condition, while having fully considered influence of the peripheral point to current location sharpness evaluation function, it is general that the present invention proposes that Region homogenization is drawn
Laplacian operater realizes balancing energy by increasing quantity and the direction that second order is led.
In view of lenticule calibrated error can cause refocusing image to generate local noise, Laplace operator is quick to noise
The shortcomings that sense, first carries out bilateral filtering pretreatment before merging to high frequency subgraph, and the present invention is based on Region homogenization drawing is general
The high frequency coefficient fusion rule of Laplacian operater is as follows:
Wherein,Represent each refocusing image of different spaces depth respective high frequency subgraph after wavelet decomposition, n=1,
2,3N, N indicate to participate in the refocusing number of image frames of total focus image co-registration；D(BLαn(i, j)) represent any 2 panel height frequency
The difference of the balanced Laplace operator of subgraph corresponding points；Max [], min [] be maximized, minimum Value Operations；HH
For customized threshold value, (HH takes 0.1, because indicating the two Region homogenization Laplce's value difference when the two difference is less than 0.1
Different very little can be ignored), when the minimum value of difference is greater than threshold value, take balanced Laplce's energy in N frame image
High frequency coefficient corresponding to the maximum is as fusion coefficients, when the two difference is less than threshold value, by multiple image high frequency system
Number determines last fusion coefficients multiplied by weight factor, wherein weight factor
The low frequency coefficient that small echo tower decomposes in step 2), the key reaction average gray feature of original image.Meter
The most straightforward procedure for calculating low frequency fusion coefficients is weighted mean method, but weighted mean method can reduce the contrast of blending image and lose
Lose some useful informations in original image.In addition, some methods for calculating gradient such as spatial frequency method, point acutance operator are also applied
Into the calculating of low frequency fusion coefficients.In light field image low frequency coefficient fusion of the invention, use for reference special based on human vision
The concept of the image visibility (Image visibility, abbreviation VI) of property, is defined as follows:
Wherein P × Q indicates the size of image I (i, j)；Indicate the average value of image I (i, j)；γ indicates vision constant,
Its value range is that the value of 0.6~0.7, VI is bigger, and representative image visibility is higher.
In the fusion process of low frequency subgraph picture, if directlying adopt the calculating of (15) formula, the VI of entire image can only obtain
Value is not used to region class or the Pixel-level fusion of multiple image.In order to rationally establish effective low frequency coefficient evaluation index, I
Formula (15) is improved, establish image visibility function pixel-based (Pixel visibility, abbreviation PV), specifically
Expression formula is as follows:
Wherein, S × T indicates the rectangular neighborhood centered on current pixel point, and S, T can only take odd number；S, t is indicated in square
The scanning step in horizontal vertical direction in shape neighborhood；Indicate the average gray value of S × T area pixel.It is merged in low frequency coefficient
In the process, using fusion rule identical with high frequency coefficient.
Fused low-and high-frequency coefficient is finally obtained into fused total focus image through wavelet inverse transformation.
Wavelet field light field total focus image generation algorithm of the present invention is described in detail above, below by specific example
To verify the validity of this algorithm.
The present invention is tested using the original image of Lytro light-field camera shooting.Fig. 5 (a) be light field original image, Fig. 5 (b),
(c), three width different spaces depth being calculated when (d) is respectively α=0.52, α=0.78, α=0.98 according to formula (10)
Multiple focussing image, the depth of focus tapers to background from prospect.Fig. 5 (e) is complete to be calculated using the method for the present invention
Focusedimage, red dotted line institute frame region clarity are apparently higher than (b) figure corresponding region, and dotted yellow line institute frame region clarity is bright
Aobvious to be higher than (c) figure corresponding region, white dashed line institute frame region clarity is apparently higher than (d) figure corresponding region, it is seen that the present invention calculates
Method can efficiently use light field original image and obtain total focus image.
In order to visually evaluate the advantage of the proposed algorithm of the present invention, three kinds of classical images based on wavelet transformation are chosen
Fusion method (Sobel algorithm, Prewitt algorithm, Laplace algorithm) is compared with the mentioned algorithm of the present invention, experimental data
Using three groups of light field original images (Flower, Forest, Zither).
Fig. 6, Fig. 7, Fig. 8 are corresponding experimental result: (a), (b) of every width figure are respectively that three width light field original images correspond to α=1, α
The refocusing image obtained when=2, the depth of focus transform to background from prospect；(c), (d), (e), (f) of Fig. 6-8 is Sobel
The total focus image that algorithm, Prewitt algorithm, traditional Laplacian algorithm and inventive algorithm obtain.In terms of visual effect, figure
Clarity of the blending image that 6Sobel algorithm, Prewitt algorithm obtain in dotted line institute's frame rectangular area is obviously not so good as the present invention
Algorithm；Fig. 7 dotted line institute frame region Sobel algorithm, Prewitt algorithm obtain clarity also obviously not as good as inventive algorithm；Fig. 8 is adopted
With the clarity of the plant leaf respective dashed institute frame region of Prewitt algorithm fusion (position of verification dotted line frame) also it is obvious not
Such as inventive algorithm；Illustrate that the mentioned light field total focus image interfusion method of the present invention has some superiority in visual effect.
Furthermore, it is contemplated that human eye vision limits, the present invention further has chosen some indexs that objectively evaluate to picture quality
It is evaluated, verifies the superiority of inventive algorithm.The present invention chooses comentropy (E), average gradient (AG), image clearly respectively
It spends (FD) and edge strength (EI) is used as evaluation index, to the matter for the total focus image that methods a variety of in Fig. 6, Fig. 7, Fig. 8 obtain
Amount is evaluated.
Wherein E is a physical quantity of metric size, and value is bigger, and expression amount of image information is bigger.AG can be sensitive
Reaction image to minor detail contrast ability, value is higher, and the ability for representing it is stronger.FD representative image readability,
Value is higher, and it is better to represent its readability.EI reflects the edge strength of image, and value is higher, and representative image edge is more clear,
Specific evaluation index corresponds to result as shown in table 1, table 2, table 3.
Data in contrast table are superior to other three kinds it is found that inventive algorithm objectively evaluates in index in four kinds of image
Traditional small wave converting method embodies the feasibility and validity of inventive algorithm.
1 Flower sample image difference blending algorithm performance indicator of table compares
Table 1 Comparison ofperformance indexes ofdifferentfusion algorithms
forFlower sample images
| E | FD | AG | EI | |
| Sobel algorithm | 6.8676 | 6.8991 | 6.2470 | 66.5340 |
| Prewitt algorithm | 6.8634 | 6.3270 | 5.8326 | 62.6420 |
| Laplace algorithm | 6.8830 | 7.6837 | 6.8668 | 72.3073 |
| Inventive algorithm | 6.8896 | 7.8498 | 7.0055 | 73.7203 |
2 Cucurbit sample image difference blending algorithm performance indicator of table compares
Table 2 Comparison ofperformance indexes ofdifferent fusion
algorithms for Cucurbit sample images
| E | FD | AG | EI | |
| Sobel algorithm | 5.7544 | 2.9136 | 2.5328 | 26.5157 |
| Prewitt algorithm | 5.7492 | 2.5766 | 2.2905 | 24.2735 |
| Laplace algorithm | 5.8011 | 3.5235 | 3.0018 | 31.0134 |
| Inventive algorithm | 5.8099 | 3.6305 | 3.0875 | 31.9033 |
3 Zither sample image difference blending algorithm performance indicator of table compares
Table 3 Comparison ofperformance indexes ofdifferent fusion
algorithms forZither sample images
| E | FD | AG | EI | |
| Sobel algorithm | 6.2935 | 5.1865 | 4.4675 | 48.3854 |
| Prewitt algorithm | 6.2695 | 4.5182 | 4.0184 | 43.7566 |
| Laplace algorithm | 6.2716 | 5.6773 | 4.8649 | 52.4474 |
| Inventive algorithm | 6.2987 | 6.2987 | 4.9425 | 53.1501 |
The present invention completes the calculating of light field original image to total focus image, using the method evaluated based on wavelet field clarity
The fusion for realizing total focus image avoids blocking artifact caused by traditional spatial domain picture blending algorithm.Decoding is obtained first
4D light field data carries out spatial alternation and projection, the multiple focussing image for total focus image co-registration is obtained, then by each
Multiple focussing image collection carries out wavelet decomposition and tower reconstruct building high and low frequency subgraph image set realizes image co-registration.In small echo high frequency
In the fusion of subgraph, the invention proposes the sharpness evaluation functions based on Region homogenization Laplace operator；It is low in small echo
In the fusion of frequency subgraph, the invention proposes the sharpness evaluation function based on pixel visibility, Lai Tigao total focus image
Fusion mass.Show that the mentioned method of the present invention compared with the traditional algorithm based on wavelet transformation, is finally melted by experiment above
Close image from subjective vision to objective indicator on be all improved.
The embodiment of the present invention is explained in detail above in conjunction with attached drawing, but the present invention is not limited to above-mentioned implementations
Example, within the knowledge of a person skilled in the art, can also make without departing from the purpose of the present invention
Various change out.
Claims (4)
1. wavelet field light field total focus image generation algorithm, which is characterized in that realize according to the following steps:
Step 1): obtaining 4D light field for light field original image after data decode, and selects different αn(n=1,2,3 ...), utilizes number
Refocusing technology obtains the refocusing image of different spaces depth
Step 2) calculates the small echo high and low frequency subgraph of each frame refocusing image
Region homogenization Laplce BL operator and pixel visibility PV function conduct is respectively adopted to high and low frequency subgraph in step 3)
Image co-registration clarity evaluation index realizes the fusion of low-and high-frequency coefficient；
Low-and high-frequency coefficient is obtained fused total focus image by wavelet inverse transformation by step 4).
2. wavelet field light field total focus image generation algorithm according to claim 1, it is characterised in that:
BL operator in step 3) is Region homogenization Laplace operator, and the expression formula of the operator is as follows:
Wherein, wherein S × T indicates equalization region size, and S, T can only take odd number；S, t expression level, vertical direction second order are led
Step-length；Indicate weight factor, the closer point of distance center point, weight factor is bigger, to Laplace operator value
Contribution is bigger, conversely, distance center point is remoter, contributes Laplace operator value smaller.
3. wavelet field light field total focus image generation algorithm according to claim 1, it is characterised in that:
PV function is pixel visibility function in step 3), and expression is as follows:
Wherein, S × T indicates the rectangular neighborhood centered on current pixel point, and S, T can only take odd number；S, t is indicated in rectangle neighbour
The scanning step in horizontal vertical direction in domain；Indicate the average gray value of S × T area pixel.
4. wavelet field light field total focus image generation algorithm according to claim 2 or 3, it is characterised in that: low frequency coefficient
Identical with high frequency coefficient fusion rule, by taking high frequency coefficient merges as an example, rule is as follows:
Wherein,Represent each refocusing image of different spaces depth respective high frequency subgraph after wavelet decomposition, n=1,2,
3 ... N, N indicate to participate in the refocusing number of image frames of total focus image co-registration；Represent any 2 width high frequency subgraph
The difference of the balanced Laplace operator of corresponding points；Max [], min [] be maximized, minimum Value Operations；HHTo make by oneself
Adopted threshold value takes balanced Laplce's energy the maximum institute in N frame image right when the minimum value of difference is greater than threshold value
The high frequency coefficient answered is as fusion coefficients, when the two difference is less than threshold value, by multiple image high frequency coefficient multiplied by weight
The factor determines last fusion coefficients, wherein weight factor
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811259275.1A CN109447930B (en) | 2018-10-26 | 2018-10-26 | Wavelet domain light field full-focusing image generation algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811259275.1A CN109447930B (en) | 2018-10-26 | 2018-10-26 | Wavelet domain light field full-focusing image generation algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109447930A true CN109447930A (en) | 2019-03-08 |
| CN109447930B CN109447930B (en) | 2021-08-20 |
Family
ID=65547793
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811259275.1A Active CN109447930B (en) | 2018-10-26 | 2018-10-26 | Wavelet domain light field full-focusing image generation algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109447930B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110662014A (en) * | 2019-09-25 | 2020-01-07 | 江南大学 | Light field camera four-dimensional data large depth-of-field three-dimensional display method |
| CN111145134A (en) * | 2019-12-24 | 2020-05-12 | 太原科技大学 | Block effect-based microlens light field camera full-focus image generation algorithm |
| CN112132771A (en) * | 2020-11-02 | 2020-12-25 | 西北工业大学 | Multi-focus image fusion method based on light field imaging |
| CN112330757A (en) * | 2019-08-05 | 2021-02-05 | 复旦大学 | Complementary color wavelet measurement for evaluating color image automatic focusing definition |
| CN112801913A (en) * | 2021-02-07 | 2021-05-14 | 佛山中纺联检验技术服务有限公司 | Method for solving field depth limitation of microscope |
| CN113487526A (en) * | 2021-06-04 | 2021-10-08 | 湖北工业大学 | Multi-focus image fusion method for improving focus definition measurement by combining high and low frequency coefficients |
| CN116847209A (en) * | 2023-08-29 | 2023-10-03 | 中国测绘科学研究院 | Log-Gabor and wavelet-based light field full-focusing image generation method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5952957A (en) * | 1998-05-01 | 1999-09-14 | The United States Of America As Represented By The Secretary Of The Navy | Wavelet transform of super-resolutions based on radar and infrared sensor fusion |
| CN101877125A (en) * | 2009-12-25 | 2010-11-03 | 北京航空航天大学 | Wavelet domain statistical signal-based image fusion processing method |
| CN108537756A (en) * | 2018-04-12 | 2018-09-14 | 大连理工大学 | Single image to the fog method based on image co-registration |
| CN108581869A (en) * | 2018-03-16 | 2018-09-28 | 深圳市策维科技有限公司 | A kind of camera module alignment methods |
-
2018
- 2018-10-26 CN CN201811259275.1A patent/CN109447930B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5952957A (en) * | 1998-05-01 | 1999-09-14 | The United States Of America As Represented By The Secretary Of The Navy | Wavelet transform of super-resolutions based on radar and infrared sensor fusion |
| CN101877125A (en) * | 2009-12-25 | 2010-11-03 | 北京航空航天大学 | Wavelet domain statistical signal-based image fusion processing method |
| CN108581869A (en) * | 2018-03-16 | 2018-09-28 | 深圳市策维科技有限公司 | A kind of camera module alignment methods |
| CN108537756A (en) * | 2018-04-12 | 2018-09-14 | 大连理工大学 | Single image to the fog method based on image co-registration |
Non-Patent Citations (2)
| Title |
|---|
| Y. CHAI: "Multifocus image fusion scheme using focused region detection", 《OPTICS COMMUNICATIONS》 * |
| 叶明: "区域清晰度的小波变换图像融合算法研究", 《电子测量与仪器学报》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112330757A (en) * | 2019-08-05 | 2021-02-05 | 复旦大学 | Complementary color wavelet measurement for evaluating color image automatic focusing definition |
| CN110662014A (en) * | 2019-09-25 | 2020-01-07 | 江南大学 | Light field camera four-dimensional data large depth-of-field three-dimensional display method |
| CN111145134A (en) * | 2019-12-24 | 2020-05-12 | 太原科技大学 | Block effect-based microlens light field camera full-focus image generation algorithm |
| CN111145134B (en) * | 2019-12-24 | 2022-04-19 | 太原科技大学 | Block effect-based microlens light field camera full-focus image generation algorithm |
| CN112132771A (en) * | 2020-11-02 | 2020-12-25 | 西北工业大学 | Multi-focus image fusion method based on light field imaging |
| CN112132771B (en) * | 2020-11-02 | 2022-05-27 | 西北工业大学 | Multi-focus image fusion method based on light field imaging |
| CN112801913A (en) * | 2021-02-07 | 2021-05-14 | 佛山中纺联检验技术服务有限公司 | Method for solving field depth limitation of microscope |
| CN113487526A (en) * | 2021-06-04 | 2021-10-08 | 湖北工业大学 | Multi-focus image fusion method for improving focus definition measurement by combining high and low frequency coefficients |
| CN113487526B (en) * | 2021-06-04 | 2023-08-25 | 湖北工业大学 | Multi-focus image fusion method for improving focus definition measurement by combining high-low frequency coefficients |
| CN116847209A (en) * | 2023-08-29 | 2023-10-03 | 中国测绘科学研究院 | Log-Gabor and wavelet-based light field full-focusing image generation method and system |
| CN116847209B (en) * | 2023-08-29 | 2023-11-03 | 中国测绘科学研究院 | Log-Gabor and wavelet-based light field full-focusing image generation method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109447930B (en) | 2021-08-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109447930A (en) | Wavelet field light field total focus image generation algorithm | |
| CN106408524B (en) | Depth image enhancement method based on two-dimensional image assistance | |
| Dudhane et al. | RYF-Net: Deep fusion network for single image haze removal | |
| Gao et al. | Gradient-based interpolation method for division-of-focal-plane polarimeters | |
| CN106339998B (en) | Multi-focus image fusing method based on contrast pyramid transformation | |
| CN107203985B (en) | A kind of more exposure image fusion methods under end-to-end deep learning frame | |
| CN111145134B (en) | Block effect-based microlens light field camera full-focus image generation algorithm | |
| CN105894483B (en) | A kind of multi-focus image fusing method based on multi-scale image analysis and block consistency checking | |
| Alireza Golestaneh et al. | Algorithm for JPEG artifact reduction via local edge regeneration | |
| CN107135388A (en) | A kind of depth extraction method of light field image | |
| JP2013531268A (en) | Measuring distance using coded aperture | |
| CN104200450B (en) | A kind of thermal-induced imagery definition enhancing method | |
| CN110533607A (en) | A kind of image processing method based on deep learning, device and electronic equipment | |
| CN109118544A (en) | Synthetic aperture imaging method based on perspective transform | |
| CN110648302B (en) | Light field full-focus image fusion method based on edge enhancement guide filtering | |
| CN113763300B (en) | Multi-focusing image fusion method combining depth context and convolution conditional random field | |
| CN117274759A (en) | Infrared and visible light image fusion system based on distillation-fusion-semantic joint driving | |
| CN116091314A (en) | Infrared image stitching method based on multi-scale depth homography | |
| An et al. | Image super-resolution reconstruction algorithm based on significant network connection-collaborative migration structure | |
| CN117197627B (en) | Multi-mode image fusion method based on high-order degradation model | |
| CN103873773B (en) | Primary-auxiliary synergy double light path design-based omnidirectional imaging method | |
| CN105303544A (en) | Video splicing method based on minimum boundary distance | |
| CN101726829A (en) | Method for automatically focusing zoom lens | |
| CN109801273B (en) | Light field image quality evaluation method based on polar plane linear similarity | |
| CN116433525A (en) | Underwater image defogging method based on edge detection function variation model |
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 |