CN106960414A - A kind of method that various visual angles LDR image generates high-resolution HDR image - Google Patents
A kind of method that various visual angles LDR image generates high-resolution HDR image Download PDFInfo
- Publication number
- CN106960414A CN106960414A CN201611139687.2A CN201611139687A CN106960414A CN 106960414 A CN106960414 A CN 106960414A CN 201611139687 A CN201611139687 A CN 201611139687A CN 106960414 A CN106960414 A CN 106960414A
- Authority
- CN
- China
- Prior art keywords
- image
- images
- pixel
- hdr
- point
- 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
- 238000000034 method Methods 0.000 title claims abstract description 27
- 230000000007 visual effect Effects 0.000 title claims abstract description 22
- 230000004927 fusion Effects 0.000 claims abstract description 28
- 230000000694 effects Effects 0.000 claims abstract description 12
- 239000013598 vector Substances 0.000 claims description 20
- 239000002131 composite material Substances 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000011160 research Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011065 in-situ storage Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 239000000155 melt Substances 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/40—Scaling the whole image or part thereof
- G06T3/4038—Scaling the whole image or part thereof for image mosaicing, i.e. plane images composed of plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration by the use of histogram techniques
-
- G06T5/70—
-
- 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
Abstract
The invention discloses a kind of method that various visual angles LDR image generates high-resolution HDR image, step (1), the image for obtaining the different visual angles of two width on the same line, and require that visual field content two images adjacent with left and right of an image have intersection;Step (2), the identification and matching for carrying out adjacent two images characteristic point, the common portion in two images is extracted;Step (3), the intersection for having length difference exposure to adjacent two images estimate factor generation weight map using contrast, saturation degree and appropriate light exposure as three, carry out multi-resolution pyramid fusion, image of the generation with HDR effects;The two images of step (4), acquisition with similar HDR effects;Step (5), by the two images with similar HDR effects merged with intersection generation HDR image, generate high-resolution HDR image.The present invention is combined fusion with splicing, improves the real-time and frame per second and visual effect in video system of several splicings.
Description
Technical field
The present invention relates to image processing field, more particularly to a kind of method for generating high-resolution HDR image.
Background technology
The generating algorithm of HDR image and the joining method of panoramic picture are all the research for just having correlation a long time ago, its
In the comparative maturity that has developed of some more classical algorithms.And the HDR panoramic pictures that combine both or video
Generation method research is just to rise recent years, is primarily due to its trend for meeting social development, there is many applications
Value and prospect.The generation method of HDR panoramic pictures is related to multiple links, melts wherein mainly having including image acquisition, image
Conjunction, the several major classes of image mosaic, each section are contemplated that different situations are realized by different methods, though in addition,
So each partly seem relatively independent, but the integration method for each several part is that global design is also researchers institute emphasis
A part of content of concern.
High-resolution HDR and the generation of panorama HDR image have become focus of attention, and numerous scholars and mechanism deploy with regard to this
Research.Such as, HDR panoramas are applied to take photo by plane in the air by Fumio Okura et al., describe a set of complete from HDR is obtained in the air
The holonomic system of scape image, by fixing a stereoscopic camera unmanned plane is each up and down, then each camera passes through automatic exposure
Control obtains LDR image sequence, then obtains HDR image by specific blending algorithm, finally obtains two cameras up and down
HDR image is merged, and obtains last HDR panoramic pictures.
Vladan Popovic et al. devise an image processing system based on FPGA, it is possible to use designed by it
Annular image collecting device realizes the fusion of HDR panoramic frames in real time.It is long per two neighboring first time for exposure of shooting, one
Time for exposure is short, carries out image co-registration using the intersection of its adjacent image, ultimately generates the HDR image frame of a panorama,
Systems approach designed by it is higher to the dependence of hardware, is not easy to directly handle image.
Kvale Stensland et al. are realized to football pitch in-situ match by five cameras of different angles
Covering comprehensively, it is long by automatic exposure by being spliced to form the full-view image to whole football pitch, and in each angle
Two different frames of short exposure, generation HDR are merged to it, then the HDR image of each angle is carried out being spliced to form last HDR panoramas
Picture frame.
For basic, current method is mostly first to obtain time for exposure different a few frame low-dynamic ranges in different visual angles
(LDR) image, it is merged respectively the HDR image under generation different visual angles, then again to the HDR image under these different visual angles
Spliced (as shown in figure), time cost of this method on image is obtained is long, and will fusion and splicing point
Step carries out relatively complicated.
The content of the invention
Based on prior art, the present invention proposes a kind of method that various visual angles LDR image generates high-resolution HDR image,
By using the intersection of obtained different visual angles LDR image, fusion is carried out with splicing to combine, proposed a kind of based on Nogata
Scheme the high-resolution HDR image generation method of matching.
The present invention proposes a kind of method that various visual angles LDR image generates high-resolution HDR image, and this method includes following
Step:
Step (1), the image for obtaining the different visual angles of two width on the same line, and two neighboring camera is set to difference
Time for exposure, have the long and short two kind time for exposure, in adjacent cameras alternately, and require the visual field content of an image with it is left
Right adjacent two image has intersection;
Step (2), identification and matching using the adjacent two images characteristic point of surf operators progress, pass through the spy to matching
Levy and a little do the occurrence that average calculating operation obtains translating parallax, so that the common portion in two images be extracted;Specifically do
Method is:The characteristic value of each pixel is obtained first with Hessian matrixes, Hessian matrixes are
Wherein, Lxx(x, σ) is the image g (σ) that are obtained after gaussian filtering of original image I in the second dervative in x directions, Lxy
(x,σ)、Lyy(x, σ) is also all the second dervative of the g (σ) in all directions.
Calculate characteristic value formula be
det(Happrox)=DxxDyy-(0.9Dxy)2
Wherein Dxx、Dyy、DxyThe respectively second dervative of the approximate template of Hesse matrices in the corresponding direction.
If characteristic value of certain point is maximum in the point of 27, its field, it is believed that the point is characterized a little.As shown in Figure 5
For the Partial Feature point in the image of extraction;
The characteristic vector of characteristic point is obtained, the principal direction of characteristic point is calculated first, detailed process is as follows:
1) statistics is proportional to some numerical digit radius of feature point scale centered on characteristic point, and subtended angle is 60 ° of fan section
(x directional wavelet transforms ring sumX=(response of y directional wavelet transforms) * (Gaussian function), sumY=of all pixels point in domain
Should) * (Gaussian function), calculate composite vector angle, θ=arctan (sumY/sumX), the long sqrt (sumy*sumy+sumx* of mould
sumx)。
2) sector is in kind calculated into composite vector along rotate counterclockwise (it is 0.1 radian typically to take step-length).
3) the fan-shaped long maximum of composite vector mould of all directions is obtained, its corresponding angle is characteristic point principal direction.
Obtain characteristic vector detailed process as follows:
1) one piece of square area centered on characteristic point is selected, is rotated and is alignd with principal direction.
2) square is divided into 4 × 4 16 sub-regions, wavelet transformation is carried out to each region, 4 coefficients are obtained.
3) by above-mentioned two step, 4 × 4 × 4=64 dimensional vectors are generated.
Two inner product of vectors are calculated, maximum is the point most matched with the point, a specific threshold value is set, only when maximum is big
Two Feature Points Matchings are believed that in this threshold value.
Step (3), fusion generation HDR image:There is the intersection of length difference exposure to adjacent two images with right
Estimate factor generation weight map for three than degree, saturation degree and appropriate light exposure, carry out multi-resolution pyramid fusion, generation tool
There is the image of HDR effects;The specific formula of fusion is as follows:
Wherein RijTo merge pixel value of the result images of generation at (i, j) position, Lyy(x, σ) is that kth width inputs figure
As correspondence position pixel value,For the weighted value after normalization.
Contrast computing formula is:
C=| h*I |
Wherein C represents contrast, and I is the image of contrast to be asked, and h is Laplace filter.
Saturation degree is obtained by calculating the standard deviation of three chrominance channels, and specific formula for calculation is as follows:
Wherein S represents saturation degree, wC、IG、IBThe respectively pixel value of tri- color channels of R, G, B, μ is equal for its three's
Value.
Appropriate light exposure estimates that specific formula for calculation is by Gaussian curve:
E=ER×EG×EB
Wherein E is the overall appropriate light exposure of image, ER、EG、EBThe appropriate light exposure of respectively each passage, here I
Provide σ=0.2.
Weight map computing formula is:
wC=wS=wE=1
Wherein Cij,k、Sij,k、Eij,kThe contrast of pixel, saturation degree at (i, j) position respectively in kth width image
With appropriate light exposure.Final weighted value is obtained by estimating fac-tor to three.wC、wS、wEFor represent three estimate because
" influence " size of son in generation weight map.
Step (4), the histogram for obtaining each Color Channel of HDR image of fusion generation in reference picture i.e. step (3),
The HDR image generated is merged as reference using intersection, to two original LDR images of acquisition, Histogram Matching is carried out, passes through
Mapping adjustment original image pixel value is sized such that the histogram and the histogram approximately equal of reference picture of image after adjustment, with
Make there is no the part overlapped to obtain the tone similar to HDR image in image, that is, obtain the two width figures with similar HDR effects
Picture;
Step (5), generation will be merged with intersection by the two images with similar HDR effects of Histogram Matching
HDR image, for each passage of pixel in picture registration region chromatic value Pixel by corresponding points in two images
Gray value Pixel_L and Pixel_R weighted average is obtained, i.e.,:
Pixel=k × Pixel_L+ (1-k) × Pixel_R
Wherein, k=h/R, overlapping region overall width is accounted for for represent current pixel point and overlapping region left margin apart from h
R ratio;K values are bigger, illustrate that the pixel value of left-side images occupies bigger proportion in fusion.I.e. in overlapping region, edge
The direction of left-side images image to the right, k fades to 0 by 1, so as to realize the smooth registration of overlapping region;Pass through Weighted Fusion
Stitching algorithm is spliced, and is seamlessly transitted in adjacent, so as to generate high-resolution HDR image.
Compared with prior art, the present invention efficiently utilizes the intersection of different visual angles image, dexterously will fusion
Combined with splicing, reduce the number of image frames obtained needed for a panel height resolution ratio HDR image;Efficiency is improved, is reduced
Algorithm complex;The real-time of several splicings and the frame per second in HDR video systems are improved with improving image picture quality
Visual effect.
Brief description of the drawings
Fig. 1 is the schematic diagram for the generation HDR high-definition pictures commonly used;
Fig. 2 is acquisition image process schematic diagram of the invention;
Fig. 3 illustrates for a kind of overall flow of the method for various visual angles LDR image generation high-resolution HDR image of the present invention
Figure;
Fig. 4 is the original image obtained;
Fig. 5 is to carry out the result of Feature point recognition and matching in the picture by surf operators;
Fig. 6 is the HDR image that generation is merged by the intersection in two images;
Fig. 7 obtains similar to generation HDR image for initial LDR image is done into Histogram Matching with reference to upper figure fusion results
The image of tone;
Fig. 8 is the high-resolution HDR ultimately produced that a few width images are stitched together by way of Weighted Fusion
Image.
Embodiment
The present invention is described in further detail below in conjunction with the accompanying drawings.
Step (1), the image for obtaining the different visual angles of two width on the same line, and two neighboring camera is set to difference
Time for exposure, have the long and short two kind time for exposure, in adjacent cameras alternately, and require the visual field content of an image with it is left
Right adjacent two image has intersection;
Step (2), identification and matching using the adjacent two images characteristic point of surf operators progress, pass through the spy to matching
Levy and a little do the occurrence that average calculating operation obtains translating parallax, so that the common portion in two images be extracted;Specifically do
Method is:The characteristic value of each pixel is obtained first with Hessian matrixes, Hessian matrixes are
Wherein, Lxx(x, σ) is the image g (σ) that are obtained after gaussian filtering of original image I in the second dervative in x directions, Lxy
(x,σ)、Lyy(x, σ) is also all the second dervative of the g (σ) in all directions.
Calculate characteristic value formula be
det(Happrox)=DxxDyy-(0.9Dxy)2
Wherein Dxx、Dyy、DxyThe respectively second dervative of the approximate template of Hesse matrices in the corresponding direction.
If characteristic value of certain point is maximum in the point of 27, its field, it is believed that the point is characterized a little.As shown in Figure 5
For the Partial Feature point in the image of extraction;
The characteristic vector of characteristic point is obtained, the principal direction of characteristic point is calculated first, detailed process is as follows:
1) statistics is proportional to some numerical digit radius of feature point scale centered on characteristic point, and subtended angle is 60 ° of fan section
(x directional wavelet transforms ring sumX=(response of y directional wavelet transforms) * (Gaussian function), sumY=of all pixels point in domain
Should) * (Gaussian function), calculate composite vector angle, θ=arctan (sumY/sumX), the long sqrt (sumy*sumy+sumx* of mould
sumx)。
2) sector is in kind calculated into composite vector along rotate counterclockwise (it is 0.1 radian typically to take step-length).
3) the fan-shaped long maximum of composite vector mould of all directions is obtained, its corresponding angle is characteristic point principal direction.
Obtain characteristic vector detailed process as follows:
1) one piece of square area centered on characteristic point is selected, is rotated and is alignd with principal direction.
2) square is divided into 4 × 4 16 sub-regions, wavelet transformation is carried out to each region, 4 coefficients are obtained.
3) by above-mentioned two step, 4 × 4 × 4=64 dimensional vectors are generated.
Two inner product of vectors are calculated, maximum is the point most matched with the point, a specific threshold value is set, only when maximum is big
Two Feature Points Matchings are believed that in this threshold value.
Step (3), fusion generation HDR image:There is the intersection of length difference exposure to adjacent two images with right
Estimate factor generation weight map for three than degree, saturation degree and appropriate light exposure, carry out multi-resolution pyramid fusion, generation tool
There is the image of HDR effects;The specific formula of fusion is as follows:
Wherein RijTo merge pixel value of the result images of generation at (i, j) position, Lyy(x, σ) is that kth width inputs figure
As correspondence position pixel value,For the weighted value after normalization.
Contrast computing formula is:
C=| h*I |
Wherein C represents contrast, and I is the image of contrast to be asked, and h is Laplace filter.
Saturation degree is obtained by calculating the standard deviation of three chrominance channels, and specific formula for calculation is as follows:
Wherein S represents saturation degree, wC、IG、IBThe respectively pixel value of tri- color channels of R, G, B, μ is equal for its three's
Value.
Appropriate light exposure estimates that specific formula for calculation is by Gaussian curve:
E=ER×EG×EB
Wherein E is the overall appropriate light exposure of image, ER、EG、EBThe appropriate light exposure of respectively each passage, here I
Provide σ=0.2.
Weight map computing formula is:
wC=wS=wE=1
Wherein Cij,k、Sij,k、Eij,kThe contrast of pixel, saturation degree at (i, j) position respectively in kth width image
With appropriate light exposure.Final weighted value is obtained by estimating fac-tor to three.wC、wS、wEFor represent three estimate because
" influence " size of son in generation weight map.
Step (4), the histogram for obtaining each Color Channel of HDR image of fusion generation in reference picture i.e. step (3),
The HDR image generated is merged as reference using intersection, to two original LDR images of acquisition, Histogram Matching is carried out, passes through
Mapping adjustment original image pixel value is sized such that the histogram and the histogram approximately equal of reference picture of image after adjustment, with
Make there is no the part overlapped to obtain the tone similar to HDR image in image, that is, obtain the two width figures with similar HDR effects
Picture;
Step (5), generation will be merged with intersection by the two images with similar HDR effects of Histogram Matching
HDR image, for each passage of pixel in picture registration region chromatic value Pixel by corresponding points in two images
Gray value Pixel_L and Pixel_R weighted average is obtained, i.e.,:
Pixel=k × Pixel_L+ (1-k) × Pixel_R
Wherein, k=h/R, overlapping region overall width is accounted for for represent current pixel point and overlapping region left margin apart from h
R ratio;K values are bigger, illustrate that the pixel value of left-side images occupies bigger proportion in fusion.I.e. in overlapping region, edge
The direction of left-side images image to the right, k fades to 0 by 1, so as to realize the smooth registration of overlapping region;Pass through Weighted Fusion
Stitching algorithm is spliced, and is seamlessly transitted in adjacent, so as to generate high-resolution HDR image.
The each chrominance channel of the LDR image of exposure more than original as Figure 4-8, is generated La Pula by the specific embodiment of the present invention
This pyramid, then be provided with the information of original image multiresolution;Then it is index according to contrast, saturation degree, appropriate light exposure
Calculate the weight map of original image, with this, that is, by this pixel information contained number weigh the figure after merging where it
The shared ratio as in;The weight maps of different exposure images is normalized, i.e., the weight of same pixel different images and be 1;Again
Obtained weight map is generated into gaussian pyramid;Finally the laplacian pyramid of same piece image and their weight maps are generated
Gaussian pyramid correspondence be multiplied, be added the laplacian pyramid after being merged in the pyramid that obtains different images,
This laplacian pyramid is restored to the HDR image that can obtain last fusion generation.
Claims (1)
1. a kind of method that various visual angles LDR image generates high-resolution HDR image, it is characterised in that this method includes following step
Suddenly:
Step (1), the image for obtaining the different visual angles of two width on the same line, and two neighboring camera is set to different exposures
Between light time, the long and short two kind time for exposure is had, in adjacent cameras alternately, and the visual field content and left and right phase of an image is required
Adjacent two images have intersection;
Step (2), identification and matching using the adjacent two images characteristic point of surf operators progress, pass through the characteristic point to matching
The occurrence that average calculating operation obtains translating parallax is done, so that the common portion in two images be extracted;Specific practice is:
The characteristic value of each pixel is obtained first with Hessian matrixes, Hessian matrixes are
Wherein, Lxx(x, σ) is the image g (σ) that are obtained after gaussian filtering of original image I in the second dervative in x directions, Lxy(x,σ)、
Lyy(x, σ) is also all the second dervative of the g (σ) in all directions;
Calculate characteristic value formula be
det(Happrox)=DxxDyy-(0.9Dxy)2
Wherein Dxx、Dyy、DxyThe respectively second dervative of the approximate template of Hesse matrices in the corresponding direction;
If characteristic value of certain point is maximum in the point of 27, its field, it is believed that the point is characterized a little;
The characteristic vector of characteristic point is obtained, the principal direction of characteristic point is calculated first, detailed process is as follows:
1) statistics is proportional to some numerical digit radius of feature point scale, subtended angle is in 60 ° of sector region centered on characteristic point
SumX=(response of y directional wavelet transforms) * (Gaussian function), sumY=(response of x directional wavelet transforms) * of all pixels point
(Gaussian function), calculates composite vector angle, θ=arctan (sumY/sumX), the long sqrt (sumy*sumy+sumx* of mould
sumx);
2) sector is in kind calculated into composite vector along rotate counterclockwise (it is 0.1 radian typically to take step-length);
3) the fan-shaped long maximum of composite vector mould of all directions is obtained, its corresponding angle is characteristic point principal direction, obtains feature
Vectorial detailed process is as follows:
31) one piece of square area centered on characteristic point is selected, is rotated and is alignd with principal direction;
32) square is divided into 4 × 4 16 sub-regions, wavelet transformation is carried out to each region, 4 coefficients are obtained;
33) by above-mentioned two step, 4 × 4 × 4=64 dimensional vectors are generated;
Two inner product of vectors are calculated, maximum is the point most matched with the point, sets a specific threshold value, only when maximum is more than this
Individual threshold value is believed that two Feature Points Matchings;
Step (3), fusion generation HDR image:Have to adjacent two images length difference exposure intersection with contrast,
Saturation degree and appropriate light exposure estimate factor generation weight map for three, carry out multi-resolution pyramid fusion, generation has HDR
The image of effect;The specific formula of fusion is as follows:
Wherein RijTo merge pixel value of the result images of generation at (i, j) position, Lyy(x, σ) is kth width input picture pair
Position pixel value is answered,For the weighted value after normalization;
Contrast computing formula is:
C=| h*I |
Wherein C represents contrast, and I is the image of contrast to be asked, and h is Laplace filter;
Saturation degree is obtained by calculating the standard deviation of three chrominance channels, and specific formula for calculation is as follows:
Wherein S represents saturation degree, wC、IG、IBThe respectively pixel value of tri- color channels of R, G, B, μ is the average of its three;
Appropriate light exposure estimates that specific formula for calculation is by Gaussian curve:
E=ER×EG×EB
Wherein E is the overall appropriate light exposure of image, ER、EG、EGThe appropriate light exposure of respectively each passage, here provide σ=
0.2;
Weight map computing formula is:
wC=wS=wE=1
Wherein Cij,k、Sij,k、Eij,kThe contrast of pixel at (i, j) position, saturation degree and suitable respectively in kth width image
Spend light exposure;Final weighted value, w are obtained by estimating fac-tor to threeC、wS、wEFor representing that estimating the factor for three exists
Generate " influence " size in weight map;
Step (4), the histogram for obtaining each Color Channel of HDR image of fusion generation in reference picture i.e. step (3), with weight
The HDR image for closing partial fusion generation is reference, to two original LDR images of acquisition, carries out Histogram Matching, passes through mapping
Adjustment original image pixel value is sized such that the histogram and the histogram approximately equal of reference picture of image after adjustment, so that figure
There is no the part overlapped to obtain the tone similar to HDR image as in, that is, obtain the two images with similar HDR effects;
Step (5), by by Histogram Matching the two images with similar HDR effects merged with intersection generate
HDR image, for each passage of pixel in picture registration region chromatic value Pixel by corresponding points in two images ash
Angle value Pixel_L and Pixel_R weighted average is obtained, i.e.,:
Pixel=k × Pixel_L+ (1-k) × Pixel_R
Wherein, k=h/R, accounts for overlapping region overall width R's for represent current pixel point and overlapping region left margin apart from h
Ratio;K values are bigger, illustrate that the pixel value of left-side images occupies bigger proportion in fusion;I.e. in overlapping region, along left side
The direction of image image to the right, k fades to 0 by 1, so as to realize the smooth registration of overlapping region;Spliced by Weighted Fusion
Algorithm is spliced, and is seamlessly transitted in adjacent, so as to generate high-resolution HDR image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611139687.2A CN106960414B (en) | 2016-12-12 | 2016-12-12 | Method for generating high-resolution HDR image from multi-view LDR image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611139687.2A CN106960414B (en) | 2016-12-12 | 2016-12-12 | Method for generating high-resolution HDR image from multi-view LDR image |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106960414A true CN106960414A (en) | 2017-07-18 |
CN106960414B CN106960414B (en) | 2020-01-03 |
Family
ID=59481515
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611139687.2A Active CN106960414B (en) | 2016-12-12 | 2016-12-12 | Method for generating high-resolution HDR image from multi-view LDR image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106960414B (en) |
Cited By (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107277387A (en) * | 2017-07-26 | 2017-10-20 | 维沃移动通信有限公司 | High dynamic range images image pickup method, terminal and computer-readable recording medium |
CN107465882A (en) * | 2017-09-22 | 2017-12-12 | 维沃移动通信有限公司 | A kind of image capturing method and mobile terminal |
CN107566739A (en) * | 2017-10-18 | 2018-01-09 | 维沃移动通信有限公司 | A kind of photographic method and mobile terminal |
CN107845128A (en) * | 2017-11-03 | 2018-03-27 | 安康学院 | A kind of more exposure high-dynamics image method for reconstructing of multiple dimensioned details fusion |
CN107885787A (en) * | 2017-10-18 | 2018-04-06 | 大连理工大学 | Image search method based on the embedded various visual angles Fusion Features of spectrum |
CN107895350A (en) * | 2017-10-27 | 2018-04-10 | 天津大学 | A kind of HDR image generation method based on adaptive double gamma conversion |
CN107945148A (en) * | 2017-12-15 | 2018-04-20 | 电子科技大学 | A kind of more exposure image interfusion methods based on MRF regional choices |
CN107977931A (en) * | 2017-12-14 | 2018-05-01 | 元橡科技(北京)有限公司 | Utilize the method for calibrated more mesh cameras generation super-resolution image |
CN108335272A (en) * | 2018-01-31 | 2018-07-27 | 青岛海信移动通信技术股份有限公司 | A kind of method and apparatus of shooting picture |
CN108848323A (en) * | 2018-06-27 | 2018-11-20 | 西安输变电工程环境影响控制技术中心有限公司 | A kind of image processing method of substation's noise synthesis cloud atlas |
CN109151334A (en) * | 2018-09-21 | 2019-01-04 | 中国计量大学 | A kind of unmanned vehicle camera system |
CN109712091A (en) * | 2018-12-19 | 2019-05-03 | 深圳市华星光电技术有限公司 | Image processing method, device and electronic equipment |
CN110460747A (en) * | 2018-05-08 | 2019-11-15 | 宁波舜宇光电信息有限公司 | Array camera module and electronic equipment and image processing method with array camera module |
CN110933327A (en) * | 2019-12-16 | 2020-03-27 | 合肥师范学院 | High-dynamic visual monitoring method based on oxygen-enriched side-blown molten pool melt surface |
CN112163584A (en) * | 2020-10-13 | 2021-01-01 | 安谋科技(中国)有限公司 | Electronic device, and method and medium for extracting image features based on wide dynamic range |
CN112241953A (en) * | 2020-10-22 | 2021-01-19 | 江苏美克医学技术有限公司 | Sample image fusion method and device based on multi-focus image fusion and HDR algorithm |
CN112351212A (en) * | 2019-08-06 | 2021-02-09 | 三星电子株式会社 | Local histogram matching with global regularization and motion exclusion for multi-exposure image fusion |
CN113808059A (en) * | 2021-09-16 | 2021-12-17 | 北京拙河科技有限公司 | Array image fusion method, device, medium and equipment |
CN116055659A (en) * | 2023-01-10 | 2023-05-02 | 如你所视(北京)科技有限公司 | Original image processing method and device, electronic equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101963751A (en) * | 2010-08-19 | 2011-02-02 | 西北工业大学 | Device and method for acquiring high-resolution full-scene image in high dynamic range in real time |
CN104349066A (en) * | 2013-07-31 | 2015-02-11 | 华为终端有限公司 | Method and device for generating images with high dynamic ranges |
CN105430298A (en) * | 2015-12-08 | 2016-03-23 | 天津大学 | Method for simultaneously exposing and synthesizing HDR image via stereo camera system |
US20160104270A1 (en) * | 2014-10-10 | 2016-04-14 | Thomson Licensing | Method for obtaining at least one high dynamic range image, and corresponding computer program product, and electronic device |
-
2016
- 2016-12-12 CN CN201611139687.2A patent/CN106960414B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101963751A (en) * | 2010-08-19 | 2011-02-02 | 西北工业大学 | Device and method for acquiring high-resolution full-scene image in high dynamic range in real time |
CN104349066A (en) * | 2013-07-31 | 2015-02-11 | 华为终端有限公司 | Method and device for generating images with high dynamic ranges |
US20160104270A1 (en) * | 2014-10-10 | 2016-04-14 | Thomson Licensing | Method for obtaining at least one high dynamic range image, and corresponding computer program product, and electronic device |
CN105430298A (en) * | 2015-12-08 | 2016-03-23 | 天津大学 | Method for simultaneously exposing and synthesizing HDR image via stereo camera system |
Non-Patent Citations (3)
Title |
---|
T. MERTENS等: "Exposure Fusion: A Simple and Practical Alternative to High Dynamic Range Photography", 《COMPUTER GRAPHICS FORUM》 * |
孙婧: "高动态范围(HDR) 技术综述", 《信息技术》 * |
朱恩弘: "单幅图像的高动态范围图像生成方法", 《计算机辅助设计与图形学学报》 * |
Cited By (30)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107277387B (en) * | 2017-07-26 | 2019-11-05 | 维沃移动通信有限公司 | High dynamic range images image pickup method, terminal and computer readable storage medium |
CN107277387A (en) * | 2017-07-26 | 2017-10-20 | 维沃移动通信有限公司 | High dynamic range images image pickup method, terminal and computer-readable recording medium |
CN107465882A (en) * | 2017-09-22 | 2017-12-12 | 维沃移动通信有限公司 | A kind of image capturing method and mobile terminal |
CN107566739A (en) * | 2017-10-18 | 2018-01-09 | 维沃移动通信有限公司 | A kind of photographic method and mobile terminal |
CN107885787A (en) * | 2017-10-18 | 2018-04-06 | 大连理工大学 | Image search method based on the embedded various visual angles Fusion Features of spectrum |
CN107885787B (en) * | 2017-10-18 | 2021-05-14 | 大连理工大学 | Multi-view feature fusion image retrieval method based on spectrum embedding |
CN107566739B (en) * | 2017-10-18 | 2019-12-06 | 维沃移动通信有限公司 | photographing method and mobile terminal |
CN107895350A (en) * | 2017-10-27 | 2018-04-10 | 天津大学 | A kind of HDR image generation method based on adaptive double gamma conversion |
CN107895350B (en) * | 2017-10-27 | 2020-01-03 | 天津大学 | HDR image generation method based on self-adaptive double gamma transformation |
CN107845128A (en) * | 2017-11-03 | 2018-03-27 | 安康学院 | A kind of more exposure high-dynamics image method for reconstructing of multiple dimensioned details fusion |
CN107977931A (en) * | 2017-12-14 | 2018-05-01 | 元橡科技(北京)有限公司 | Utilize the method for calibrated more mesh cameras generation super-resolution image |
CN107945148A (en) * | 2017-12-15 | 2018-04-20 | 电子科技大学 | A kind of more exposure image interfusion methods based on MRF regional choices |
CN107945148B (en) * | 2017-12-15 | 2021-06-01 | 电子科技大学 | Multi-exposure image fusion method based on MRF (Markov random field) region selection |
CN108335272A (en) * | 2018-01-31 | 2018-07-27 | 青岛海信移动通信技术股份有限公司 | A kind of method and apparatus of shooting picture |
CN108335272B (en) * | 2018-01-31 | 2021-10-08 | 青岛海信移动通信技术股份有限公司 | Method and device for shooting picture |
CN110460747A (en) * | 2018-05-08 | 2019-11-15 | 宁波舜宇光电信息有限公司 | Array camera module and electronic equipment and image processing method with array camera module |
CN108848323A (en) * | 2018-06-27 | 2018-11-20 | 西安输变电工程环境影响控制技术中心有限公司 | A kind of image processing method of substation's noise synthesis cloud atlas |
CN109151334B (en) * | 2018-09-21 | 2020-12-22 | 中国计量大学 | Unmanned vehicle camera system |
CN109151334A (en) * | 2018-09-21 | 2019-01-04 | 中国计量大学 | A kind of unmanned vehicle camera system |
CN109712091A (en) * | 2018-12-19 | 2019-05-03 | 深圳市华星光电技术有限公司 | Image processing method, device and electronic equipment |
CN109712091B (en) * | 2018-12-19 | 2021-03-23 | Tcl华星光电技术有限公司 | Picture processing method and device and electronic equipment |
CN112351212A (en) * | 2019-08-06 | 2021-02-09 | 三星电子株式会社 | Local histogram matching with global regularization and motion exclusion for multi-exposure image fusion |
CN110933327B (en) * | 2019-12-16 | 2021-07-30 | 合肥师范学院 | High-dynamic visual monitoring method based on oxygen-enriched side-blown molten pool melt surface |
CN110933327A (en) * | 2019-12-16 | 2020-03-27 | 合肥师范学院 | High-dynamic visual monitoring method based on oxygen-enriched side-blown molten pool melt surface |
CN112163584A (en) * | 2020-10-13 | 2021-01-01 | 安谋科技(中国)有限公司 | Electronic device, and method and medium for extracting image features based on wide dynamic range |
CN112241953A (en) * | 2020-10-22 | 2021-01-19 | 江苏美克医学技术有限公司 | Sample image fusion method and device based on multi-focus image fusion and HDR algorithm |
CN112241953B (en) * | 2020-10-22 | 2023-07-21 | 江苏美克医学技术有限公司 | Sample image fusion method and device based on multi-focus image fusion and HDR algorithm |
CN113808059A (en) * | 2021-09-16 | 2021-12-17 | 北京拙河科技有限公司 | Array image fusion method, device, medium and equipment |
CN116055659A (en) * | 2023-01-10 | 2023-05-02 | 如你所视(北京)科技有限公司 | Original image processing method and device, electronic equipment and storage medium |
CN116055659B (en) * | 2023-01-10 | 2024-02-20 | 如你所视(北京)科技有限公司 | Original image processing method and device, electronic equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN106960414B (en) | 2020-01-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106960414A (en) | A kind of method that various visual angles LDR image generates high-resolution HDR image | |
CN106462944B (en) | High-resolution panorama VR generator and method | |
CN107301620B (en) | Method for panoramic imaging based on camera array | |
US9142021B1 (en) | Aligning ground based images and aerial imagery | |
CN104732482B (en) | A kind of multi-resolution image joining method based on control point | |
Yang et al. | A content-aware metric for stitched panoramic image quality assessment | |
CN109314753A (en) | Medial view is generated using light stream | |
CN109934772B (en) | Image fusion method and device and portable terminal | |
CN106060493A (en) | Multi-source projection seamless edge stitching method and system | |
WO2015165222A1 (en) | Method and device for acquiring panoramic image | |
CN104392416A (en) | Video stitching method for sports scene | |
CN105023260A (en) | Panorama image fusion method and fusion apparatus | |
Chen et al. | Deep exposure fusion with deghosting via homography estimation and attention learning | |
CN105488777A (en) | System and method for generating panoramic picture in real time based on moving foreground | |
CN112862683B (en) | Adjacent image splicing method based on elastic registration and grid optimization | |
CN107767339A (en) | A kind of binocular stereo image joining method | |
CN105005964A (en) | Video sequence image based method for rapidly generating panorama of geographic scene | |
CN109166076B (en) | Multi-camera splicing brightness adjusting method and device and portable terminal | |
CN105894448A (en) | Mask matrix generation method and device and parking image synthesis method and device | |
Zia et al. | 360 panorama generation using drone mounted fisheye cameras | |
CN107330856B (en) | Panoramic imaging method based on projective transformation and thin plate spline | |
Weibel et al. | Contrast-enhancing seam detection and blending using graph cuts | |
EP2879090B1 (en) | Aligning ground based images and aerial imagery | |
Cao et al. | Boosting image super-resolution via fusion of complementary information captured by multi-modal sensors | |
CN113688680B (en) | Intelligent recognition and tracking system |
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 |