CN104881854B - High dynamic range images fusion method based on gradient and monochrome information - Google Patents
High dynamic range images fusion method based on gradient and monochrome information Download PDFInfo
- Publication number
- CN104881854B CN104881854B CN201510262164.6A CN201510262164A CN104881854B CN 104881854 B CN104881854 B CN 104881854B CN 201510262164 A CN201510262164 A CN 201510262164A CN 104881854 B CN104881854 B CN 104881854B
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- mover
- image
- sigma
- 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.)
- Expired - Fee Related
Links
Landscapes
- Image Processing (AREA)
Abstract
It is a kind of calculating fast and effectively many exposure image fusion methods the present invention relates to digital image processing techniques.This method is not required to picture breakdown and the restructuring procedure of complexity, completes to merge using pixel weighting scheme, image is shown more detailed information, improve picture quality.Therefore, the present invention is adopted the technical scheme that, the high dynamic range images fusion method based on gradient and monochrome information comprises the following steps:1) to many exposure images of every width of input, its local contrast weight factor L is calculatedi(x, y);2) original image is transformed into tone H, color saturation S and brightness I color spaces, extract light intensity level I calculates its luminance weights factor Hi(x, y);3) 4) weight estimation is improved using recursion filter to weight map, obtains the final weighting function of every width input picture:5) fusion is weighted to many exposure images.Present invention is mainly applied to Digital Image Processing.
Description
Technical field
The present invention relates to digital image processing techniques, high dynamic range images treatment technology.Specifically, it is related to based on gradient
With the high dynamic range images fusion method of monochrome information.
Background technology
Natural scene has a very huge dynamic range, and common picture pick-up device is difficult to catch all bright under Same Scene
Spend grade.Under the strong scene of light, the image that common camera is shot often occurs under-exposure or overexposure
Phenomenon, therefore region excessively dark or excessively bright in the picture can lose detailed information.In order to solve this problem, people generally clap
A series of image of the different Same Scene of exposures is taken the photograph, by carrying out processing generation HDR (High to image sequence
Dynamic Range, HDR) image.High dynamic range images can provide more dynamic ranges and image detail, preferably
The visual effect reflected in true environment.
There are some many exposure image blending algorithms to be suggested at present.Goshtasby [1] proposes a kind of image block
Many exposure image integration technologies.The many exposure images of every width are divided into a series of size identical blocks, use information by this method
To find, each region is corresponding to expose optimal image block to the quality of entropy index measurement image block.Afterwards, using a list
Selected piece is fused together by the tune fusion function that successively decreases.In order that obtaining fused images has maximum information content, it is optimal
The width of piecemeal size and fusion function is iterated optimization by gradient ascent algorithm.Due to needing substantial amounts of complicated calculations,
This algorithm it is inefficient.In addition, this method can not obtain good syncretizing effect at object boundary.Mertens etc.
[2] index is instructed to generate weight map as fusion using contrast, saturation degree and exposure in rgb space, using Laplce
Pyramid completes image co-registration.Many exposure images are carried out laplacian decomposition by algorithm first, and weight map is carried out into Gauss point
Solution.Afterwards, each tomographic image of laplacian pyramid is multiplied with each tomographic image of corresponding gaussian pyramid.Finally,
Fused images can be generated by Pyramid Reconstruction.The result of this blending algorithm shows unintelligible in the dark space of image, under-exposure
Pixel ratio is more.Moreover, with the increase and the increase of picture size of the pyramid decomposition number of plies, the run time of algorithm can be bright
Aobvious increase.Wei Zhang etc. [3] carry out many exposure image fusions using the gradient information of input picture.Ladder is respectively adopted in algorithm
Spend size and Orientation and build the compliance evaluation factor and the visibility evaluation factor, two fac-tors are obtained into weight map afterwards.
Final fused images are combined by original sequence with weight map to be obtained.Shen Jianbing etc. [4] propose a kind of new
The enhancing laplacian pyramid blending algorithm of grain husk.The algorithm is based on the partial weight factor and global weight factor is combined conduct
Weight measures image exposure quality, and enhancing process is guided by the weight map.The image of this method generation has more preferable color
And texture information.But, the algorithm has equally used pyramid decomposition and restructuring procedure, and computational efficiency is not high.
As can be seen that existing many exposure image blending algorithms generally require progress to obtain relatively good syncretizing effect
Filtering or picture breakdown restructuring procedure, amount of calculation is than larger.The result that the simple algorithm of fusion process is obtained is again unsatisfactory.
Therefore, the present invention proposes a kind of easy direct fusion method of many exposure images, is weighed using gradient magnitude and luminance component guiding
Multigraph completes fusion process, the dynamic range of expanded images.
Bibliography
[1]Goshtasby A A.Fusion of multi-exposure images[J].Image and Vision
Computing,2005,23(6):611-618.
[2]Mertens T,Kautz J,Van Reeth F.Exposure fusion[C]//Computer
Graphics and Applications,2007.PG'07.15th Pacific Conference on.IEEE,2007:
382-390.
[3]Zhang W,Cham W K.Gradient-directed composition of multi-exposure
images[C]//Computer Vision and Pattern Recognition(CVPR),2010 IEEE Conference
on.IEEE,2010:530-536.
[4]Shen J,Zhao Y,Yan S,et al.Exposure fusion using boosting Laplacian
pyramid[J].IEEE Trans.Cybern,2014,44(9):1579-1590.
The content of the invention
To overcome the shortcomings of technology, it is contemplated that exploitation is a kind of to calculate fast and effectively many exposure image fusion methods.
This method is not required to picture breakdown and the restructuring procedure of complexity, completes to merge using pixel weighting scheme, image is shown more
Detailed information, improves picture quality.Therefore, the present invention is adopted the technical scheme that, the high dynamic based on gradient and monochrome information
Range image fusion method, comprises the following steps:
1) to many exposure images of every width of input, its local contrast weight factor L is calculatedi(x,y);
2) original image is transformed into tone H, color saturation S and brightness I color spaces, extract light intensity level I calculates its bright
Spend weight factor Hi(x,y);
3) weight is estimated:Two image quality measurement factors are combined to calculate weighting function during image co-registrationNormalized weighting function W is obtained after being normalizedi′(x,y):
ε<10-12;
4) weight map is improved using recursion filter, obtains the final weighting function W of every width input picturei(x,
y):
Wi(x, y)=RF (Wi′(x,y),sigma_s,sigma_r) (9)
Wherein, RF represents recursion filter operator, and sigma_s and sigma_r are filter parameter;
5) fusion is weighted to many exposure images, uses Wi(x, y) as weight, obtain final fused images F (x,
y):
Ii(x, y) is the intensity for the pixel that the i-th width input picture is located at (x, y) position.
Local contrast weight factor Li(x, y) is calculated by below equation:
Wherein, gi(x, y) is the image obtained to the gray level image application Sobel gradient operator of the i-th width image of input
Gradient size values,It is not normalized local contrast weight factor.
The calculating of the luminance weights factor, the bad pixel of these exposures is rejected with following formula, obtains good with exposure
The input picture B of good pixeli(x,y):
Wherein, li(x, y) is the brightness value for the pixel that the i-th width input picture is located at (x, y) position, and t is a definition
The threshold value of exposure scope, the weight by brightness value for 0.5 pixel is set to 1, and the weighted value of other pixels is according to brightness
The increase of value or decrease are successively decreased successively, obtain initial luminance weights figure
To avoid the occurrence of singular point, luminance weights are added with a very small value, the luminance weights after being improved because
Son
Normalization, obtains final luminance weights factor Hi(x,y):
Compared with the prior art, technical characterstic of the invention and effect:
The fused images obtained using technical scheme see overall chiaroscuro effect preferably from subjective, particularly with
The fusion in dark portion region achieves preferable effect in image, and with preferable color rendition degree.As a result of weight
The mode for scheming directly to be multiplied with former many exposure image sequences is merged, and is calculated without complicated picture breakdown, inventive algorithm
Computation complexity is low.In terms of objective indicator experiment.
Brief description of the drawings
In many exposure image sequences of Fig. 1 " Garage " and fusion results, figure, (a) " Garage " LDR image sequence;(b)
Mertens et al. algorithms;(c) Wei Zhang et al. algorithms;(d) inventive algorithm.
The flow chart of Fig. 2 the present invention program.
Embodiment
The dynamic range of imaging sensor is far smaller than the excursion of nature brightness, causes and changes acutely in brightness
Occasion can not obtain clearly image by single shot.Many exposure image fusions are a kind of dynamic models of effective expanded images
The mode enclosed, is changed greatly to brightness and the scene without fast-moving target can significantly improve image definition and contrast,
Show more image detail informations.HDR technology can be widely applied to traffic video, biologic medical, satellite
Remote sensing, game etc. some need show high dynamic detail pictures industry.Therefore, the method tool of exploitation expanded images dynamic range
There are very high research and application value.
The present invention is based on two direct Weighted Fusions of many exposure images of image quality measurement factor pair.Referred to by two fusions
The weighting function constituted is marked, every many exposure images of width can calculate a width weight map.Two weight factors are as follows:
1st, local contrast weight factor.For many exposure images of input, its gradient magnitude can reflect the clear of image
It is clear degree and comprising information content.Under-exposed or over-exposed region in the picture, the Grad of pixel can be very
It is small.On the contrary, in well-exposed region, the Grad of pixel can be than larger.It therefore, it can as measurement be schemed with gradient magnitude
As the index of exposure quality.The local contrast weight factor L of many exposure imagesi(x, y) can be calculated by below equation:
Wherein, ε is a very small value, to avoid the occurrence of singular point.gi(x, y) is the ash to the i-th width image of input
The image gradient sizes values that degree image application Sobel gradient operator is obtained,It is not normalized local contrast power
Repeated factor.
2nd, the luminance weights factor.The brightness value of pixel can often reflect the quality of image exposure degree.Brightness in the picture
Being worth suitable pixel has more preferable color effect and apparent detailed information.Therefore, the present invention is made from luminance component
For another image quality measurement factor., can in order to avoid the too small or excessive pixel of brightness value is interfered in fusion
To reject the bad pixel of these exposures with following formula, the input picture B with the good pixel of exposure is obtainedi(x,
y):
Wherein, li(x, y) is the brightness value for the pixel that the i-th width input picture is located at (x, y) position, and t is a definition
The threshold value of exposure scope.The present invention is when constructing the luminance weights factor, and the weight by brightness value for 0.5 pixel is set to 1.
The weighted value of other pixels successively decreases successively according to the increase or decrease of brightness value, obtains initial luminance weights figure
To avoid the occurrence of singular point, luminance weights are added with a very small value, the luminance weights after being improved because
Son
Normalization, obtains final luminance weights factor Hi(x,y):
Many exposure image sequence blending algorithm steps are as follows:
1) to many exposure images of every width of input, its local contrast weight factor is calculated.
2) original image is transformed into tone, color saturation and brightness (HIS) color space, extract light intensity level I, calculates it
The luminance weights factor.
3) weight is estimated.Two image quality measurement factors are combined to calculate weighting function during image co-registrationNormalized weighting function W is obtained after being normalizedi′(x,y):
4) weight map is improved using recursion filter so that in its weighted value of the similar region of image exposure degree
It is close, it is to avoid to occur gap artifact phenomenon in fused images, obtain the final weighting function W of every width input picturei(x,y):
Wi(x, y)=RF (Wi′(x,y),sigma_s,sigma_r) (9)
Wherein, RF represents recursion filter operator, and sigma_s and sigma_r are filter parameter.
5) fusion is weighted to many exposure images, uses Wi(x, y) as weight, obtain final fused images F (x,
y):
For verification algorithm effect, many exposure image sequences are merged and to syncretizing effect using algorithm as described above
Analyzed.For coloured image, tri- passages of R, G, B are respectively calculated in fusion, and are calculating luminance factor
When only luminance component is operated.In order to compare the quality of fused images, by the result of inventive algorithm and Mertens et al.
Algorithm [2], Wei Zhang et al. algorithm [3] compares.Respectively from it is subjective and objective two in terms of experimental result is divided
Analysis.
Image average, colored entropy, average gradient, five picture qualities of comentropy and standard deviation are respectively adopted in objective experiment
Measurement index is measured to experimental result.Image average reflects the average bright-dark degree of image.Colored entropy reflects image
The summation for the information content that tri- passages of R, G, B are included.Average gradient is the definition of image, table of the reflection image to Detail contrast
Danone power.Comentropy reflects the size that image includes information content, is to weigh the important indicator that image information enriches degree.
Standard deviation reflects the dispersion degree of gray average.
The experimental result of " Garage " image sequence is as shown in Fig. 1, table 1.
The objective indicator experimental result of table 1
Can be seen that Mertens et al. algorithms from Fig. 1 experimental result has preferable effect in terms of color rendition degree,
But the dark space part details in garage shows unintelligible, under exposed pixel ratio is more, and general image is partially dark.It is additionally, since
Need to carry out the decomposition and reconstruction of laplacian pyramid, Mertens et al. method amount of calculation is too big.Equally, Wei Zhang
Et al. algorithm show dark in garage region, lack detailed information.The fused images obtained using technical scheme
Overall chiaroscuro effect is seen from subjective preferably, and the fusion particularly with dark portion region in image achieves preferable effect, and
With preferable color rendition degree.Melted as a result of weight map with the mode that former many exposure image sequences are directly multiplied
Close, calculated without complicated picture breakdown, inventive algorithm computation complexity is low., can by table 1 in terms of objective indicator experiment
See, the fused images of inventive algorithm generation, its average, colored entropy, comentropy, standard deviation requirement are all than other two kinds of algorithm numbers
Value is high, consistent with the conclusion of subjective vision effect assessment.This explanation inventive algorithm is in the side such as extraction and processing of detailed information
Face is superior to other two kinds of algorithms.
In actual applications, in order to obtain optimal fusion results, the parameter being related in inventive algorithm is carried out as follows
Set:ε spans are less than or equal to 10-12, t span is 0.05 to 0.2, sigma_s=60, sigma_r=2.
ε=10 in above-mentioned experiment-12, t=0.1.Using many exposure image sequences as experimental subjects, the step of using described in the present invention, press
Above-mentioned parameter value is set, and can obtain the preferable fused images of visual effect.Test result indicates that, inventive algorithm is regarded in subjectivity
Feel and have preferable effect in terms of objective quantitative index.
Claims (3)
1. a kind of high dynamic range images fusion method based on gradient and monochrome information, it is characterized in that, comprise the following steps:
1) to many exposure images of every width of input, its local contrast weight factor L is calculatedi(x,y);
2) original image is transformed into tone H, color saturation S and brightness I color spaces, extract light intensity level I calculates its brightness power
Repeated factor Hi(x,y);
3) weight is estimated:Two image quality measurement factors are combined to calculate weighting function during image co-registrationNormalized weighting function W is obtained after being normalizedi′(x,y):
<mrow>
<msubsup>
<mi>W</mi>
<mi>i</mi>
<mo>&prime;</mo>
</msubsup>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mover>
<mi>W</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<msub>
<mover>
<mi>W</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>8</mn>
<mo>)</mo>
</mrow>
</mrow>
ε<10-12;
4) weight map is improved using recursion filter, obtains the final weighting function W of every width input picturei(x,y):
Wi(x, y)=RF (Wi′(x,y),sigma_s,sigma_r) (9)
Wherein, RF represents recursion filter operator, and sigma_s and sigma_r are filter parameter;
5) fusion is weighted to many exposure images, uses Wi(x, y) obtains final fused images F (x, y) as weight:
<mrow>
<mi>F</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</msubsup>
<msub>
<mi>W</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
<mo>&times;</mo>
<msub>
<mi>I</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>10</mn>
<mo>)</mo>
</mrow>
</mrow>
Ii(x, y) is the intensity for the pixel that the i-th width input picture is located at (x, y) position.
2. the high dynamic range images fusion method as claimed in claim 1 based on gradient and monochrome information, it is characterized in that, office
Portion contrast weight factor Li(x, y) is calculated by below equation:
<mrow>
<msub>
<mover>
<mi>L</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
</mrow>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mi>g</mi>
<mrow>
<mi>i</mi>
<mrow>
<mo>(</mo>
<mrow>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mo>+</mo>
<mi>&epsiv;</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msub>
<mi>L</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
</mrow>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mover>
<mi>L</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</msubsup>
<msub>
<mover>
<mi>L</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, gi(x, y) is the image gradient obtained to the gray level image application Sobel gradient operator of the i-th width image of input
Sizes values,It is not normalized local contrast weight factor.
3. the high dynamic range images fusion method as claimed in claim 1 based on gradient and monochrome information, it is characterized in that, it is bright
The calculating of weight factor is spent, the bad pixel of these exposures is rejected with following formula, is obtained with the good pixel of exposure
Input picture Bi(x,y):
Wherein, li(x, y) is the brightness value for the pixel that the i-th width input picture is located at (x, y) position, and t is one and defines exposure
The threshold value of scope, the weight by brightness value for 0.5 pixel is set to 1, and the weighted values of other pixels is according to the increasing of brightness value
Plus or weaken successively decrease successively, obtain initial luminance weights figure
To avoid the occurrence of singular point, luminance weights are added with a very small value, the luminance weights factor after being improved
<mrow>
<msub>
<mover>
<mi>H</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mover>
<mi>H</mi>
<mo>~</mo>
</mover>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>&epsiv;</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
</mrow>
1
Normalization, obtains final luminance weights factor Hi(x,y):
<mrow>
<msub>
<mi>H</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mover>
<mi>H</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</msubsup>
<msub>
<mover>
<mi>H</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>6</mn>
<mo>)</mo>
</mrow>
<mo>.</mo>
</mrow>
2
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510262164.6A CN104881854B (en) | 2015-05-20 | 2015-05-20 | High dynamic range images fusion method based on gradient and monochrome information |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510262164.6A CN104881854B (en) | 2015-05-20 | 2015-05-20 | High dynamic range images fusion method based on gradient and monochrome information |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104881854A CN104881854A (en) | 2015-09-02 |
CN104881854B true CN104881854B (en) | 2017-10-31 |
Family
ID=53949339
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510262164.6A Expired - Fee Related CN104881854B (en) | 2015-05-20 | 2015-05-20 | High dynamic range images fusion method based on gradient and monochrome information |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104881854B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111212241A (en) * | 2020-01-13 | 2020-05-29 | 禾多科技(北京)有限公司 | High-speed automatic driving automatic exposure control method based on image gradient and entropy fusion |
Families Citing this family (33)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105744159B (en) * | 2016-02-15 | 2019-05-24 | 努比亚技术有限公司 | A kind of image composition method and device |
CN106169182B (en) * | 2016-05-25 | 2019-08-09 | 西安邮电大学 | A method of synthesizing several different exposure images |
CN106251365A (en) * | 2016-07-22 | 2016-12-21 | 北京邮电大学 | Many exposure video fusion method and device |
CN106355569A (en) * | 2016-08-29 | 2017-01-25 | 努比亚技术有限公司 | Image generating device and method thereof |
CN107993189B (en) * | 2016-10-27 | 2021-06-18 | 瑞芯微电子股份有限公司 | Image tone dynamic adjustment method and device based on local blocking |
CN108713318A (en) * | 2016-10-31 | 2018-10-26 | 华为技术有限公司 | A kind of processing method and equipment of video frame |
CN106488150B (en) * | 2016-11-25 | 2019-04-19 | 阿依瓦(北京)技术有限公司 | The system for generating high dynamic range images based on Heterogeneous Computing |
CN106780463B (en) * | 2016-12-15 | 2019-07-05 | 华侨大学 | It is a kind of to expose fused image quality appraisal procedures with reference to entirely more |
CN108205796B (en) * | 2016-12-16 | 2021-08-10 | 大唐电信科技股份有限公司 | Multi-exposure image fusion method and device |
CN106920221B (en) * | 2017-03-10 | 2019-03-26 | 重庆邮电大学 | Take into account the exposure fusion method that Luminance Distribution and details are presented |
CN107424124B (en) * | 2017-03-31 | 2020-03-17 | 北京臻迪科技股份有限公司 | Image enhancement method and device |
CN107220956A (en) * | 2017-04-18 | 2017-09-29 | 天津大学 | A kind of HDR image fusion method of the LDR image based on several with different exposures |
EP3418972A1 (en) * | 2017-06-23 | 2018-12-26 | Thomson Licensing | Method for tone adapting an image to a target peak luminance lt of a target display device |
CN108288253B (en) * | 2018-01-08 | 2020-11-27 | 厦门美图之家科技有限公司 | HDR image generation method and device |
CN108090886B (en) * | 2018-01-11 | 2022-04-22 | 南京大学 | High dynamic range infrared image display and detail enhancement method |
CN108717690B (en) * | 2018-05-21 | 2022-03-04 | 电子科技大学 | Method for synthesizing high dynamic range picture |
CN109001230A (en) * | 2018-05-28 | 2018-12-14 | 中兵国铁(广东)科技有限公司 | Welding point defect detection method based on machine vision |
CN109146798A (en) * | 2018-07-10 | 2019-01-04 | 西安天盈光电科技有限公司 | image detail enhancement method |
CN109166076B (en) * | 2018-08-10 | 2020-08-07 | 影石创新科技股份有限公司 | Multi-camera splicing brightness adjusting method and device and portable terminal |
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 |
CN110211077B (en) * | 2019-05-13 | 2021-03-09 | 杭州电子科技大学上虞科学与工程研究院有限公司 | Multi-exposure image fusion method based on high-order singular value decomposition |
CN110355465B (en) * | 2019-08-21 | 2023-07-25 | 广东福维德焊接股份有限公司 | High dynamic vision system based on industrial camera and synthesis method |
CN110580696A (en) * | 2019-08-30 | 2019-12-17 | 金陵科技学院 | Multi-exposure image fast fusion method for detail preservation |
CN110738627B (en) * | 2019-09-04 | 2022-04-26 | Tcl华星光电技术有限公司 | Multi-exposure image fusion device and multi-exposure image fusion method |
CN111080565B (en) * | 2019-12-11 | 2023-07-18 | 九江学院 | Exposure fusion method, device and storage medium based on image quality change rule |
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 |
CN113129391B (en) * | 2021-04-27 | 2023-01-31 | 西安邮电大学 | Multi-exposure fusion method based on multi-exposure image feature distribution weight |
CN113222869B (en) * | 2021-05-06 | 2024-03-01 | 杭州海康威视数字技术股份有限公司 | Image processing method |
CN113327534B (en) * | 2021-05-20 | 2023-02-28 | Tcl华星光电技术有限公司 | Data compensation method and data compensation device of display panel |
CN113240609A (en) * | 2021-05-26 | 2021-08-10 | Oppo广东移动通信有限公司 | Image denoising method and device and storage medium |
CN115278090B (en) * | 2022-09-28 | 2022-12-06 | 江苏游隼微电子有限公司 | Single-frame four-exposure WDR processing method based on line exposure |
CN115760663B (en) * | 2022-11-14 | 2023-09-22 | 辉羲智能科技(上海)有限公司 | Method for synthesizing high dynamic range image based on multi-frame multi-exposure low dynamic range image |
CN115731146B (en) * | 2022-12-26 | 2023-05-12 | 中国人民解放军战略支援部队航天工程大学 | Multi-exposure image fusion method based on color gradient histogram feature optical flow estimation |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102063712A (en) * | 2010-11-04 | 2011-05-18 | 北京理工大学 | Multi-exposure image fusion method based on sub-band structure |
CN103247036A (en) * | 2012-02-10 | 2013-08-14 | 株式会社理光 | Multiple-exposure image fusion method and device |
CN103473749A (en) * | 2013-01-09 | 2013-12-25 | 深圳信息职业技术学院 | Method and apparatus based on total variation image fusion |
CN104077759A (en) * | 2014-02-28 | 2014-10-01 | 西安电子科技大学 | Multi-exposure image fusion method based on color perception and local quality factors |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20130031574A (en) * | 2011-09-21 | 2013-03-29 | 삼성전자주식회사 | Image processing method and image processing apparatus |
-
2015
- 2015-05-20 CN CN201510262164.6A patent/CN104881854B/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102063712A (en) * | 2010-11-04 | 2011-05-18 | 北京理工大学 | Multi-exposure image fusion method based on sub-band structure |
CN103247036A (en) * | 2012-02-10 | 2013-08-14 | 株式会社理光 | Multiple-exposure image fusion method and device |
CN103473749A (en) * | 2013-01-09 | 2013-12-25 | 深圳信息职业技术学院 | Method and apparatus based on total variation image fusion |
CN104077759A (en) * | 2014-02-28 | 2014-10-01 | 西安电子科技大学 | Multi-exposure image fusion method based on color perception and local quality factors |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111212241A (en) * | 2020-01-13 | 2020-05-29 | 禾多科技(北京)有限公司 | High-speed automatic driving automatic exposure control method based on image gradient and entropy fusion |
CN111212241B (en) * | 2020-01-13 | 2021-06-18 | 禾多科技(北京)有限公司 | High-speed automatic driving automatic exposure control method based on image gradient and entropy fusion |
Also Published As
Publication number | Publication date |
---|---|
CN104881854A (en) | 2015-09-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104881854B (en) | High dynamic range images fusion method based on gradient and monochrome information | |
Ram Prabhakar et al. | Deepfuse: A deep unsupervised approach for exposure fusion with extreme exposure image pairs | |
CN110046673B (en) | No-reference tone mapping image quality evaluation method based on multi-feature fusion | |
Yang et al. | A biological vision inspired framework for image enhancement in poor visibility conditions | |
CN106127718B (en) | A kind of more exposure image fusion methods based on wavelet transformation | |
Vonikakis et al. | Multi-exposure image fusion based on illumination estimation | |
US8648942B2 (en) | Method of and apparatus for processing images using saturation functions | |
Wang et al. | Exposure fusion based on sparse representation using approximate K-SVD | |
WO2004075535A1 (en) | Method of and device for modulating a dynamic range of still and video images | |
Tian et al. | Global and local contrast adaptive enhancement for non-uniform illumination color images | |
CN112070692B (en) | Single backlight image enhancement method based on virtual exposure | |
Fang et al. | A method to improve the image enhancement result based on image fusion | |
Parihar et al. | A comprehensive analysis of fusion-based image enhancement techniques | |
CN109427041B (en) | Image white balance method and system, storage medium and terminal equipment | |
CN110910347B (en) | Tone mapping image non-reference quality evaluation method based on image segmentation | |
Vanmali et al. | Low complexity detail preserving multi-exposure image fusion for images with balanced exposure | |
Al-Ameen | Contrast Enhancement of Digital Images Using an Improved Type-II Fuzzy Set-Based Algorithm. | |
Lee et al. | Ramp distribution-based contrast enhancement techniques and over-contrast measure | |
Merianos et al. | A hybrid multiple exposure image fusion approach for HDR image synthesis | |
CN112651945A (en) | Multi-feature-based multi-exposure image perception quality evaluation method | |
Lee et al. | Image fusion-based tone mapping using gaussian mixture model clustering | |
Toh et al. | Implementation of high dynamic range rendering on acute leukemia slide images using contrast stretching | |
Wang et al. | Retinex-based color correction for displaying high dynamic range images | |
Kumari et al. | Image fusion techniques based on pyramid decomposition | |
Lin et al. | Level-base compounded logarithmic curve function for colour image enhancement |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
EXSB | Decision made by sipo to initiate substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171031 Termination date: 20210520 |
|
CF01 | Termination of patent right due to non-payment of annual fee |