CN115760663A - Method for synthesizing high dynamic range image from low dynamic range image based on multi-frame multi-exposure - Google Patents
Method for synthesizing high dynamic range image from low dynamic range image based on multi-frame multi-exposure Download PDFInfo
- Publication number
- CN115760663A CN115760663A CN202211419101.3A CN202211419101A CN115760663A CN 115760663 A CN115760663 A CN 115760663A CN 202211419101 A CN202211419101 A CN 202211419101A CN 115760663 A CN115760663 A CN 115760663A
- Authority
- CN
- China
- Prior art keywords
- pyramid
- weight
- layer
- frame
- 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
Images
Landscapes
- Studio Devices (AREA)
- Image Processing (AREA)
Abstract
A method for synthesizing a high dynamic range image based on multi-frame multi-exposure low dynamic range images comprises the steps of carrying out pyramid decomposition on N frames of LDR images with different exposures; acquiring a pyramid weight map; determining pyramid layers of each pixel point according to the image gradient information and the similarity weight of each pyramid layer; and (5) carrying out pyramid synthesis and reconstruction from the moment of determining the pyramid layer number of the current pixel point to generate the HDR image. According to the method, the ghost can be removed while HDR synthesis is carried out through the weight, an independent ghost removing process is not needed, the dynamic range of an output image can be improved, the motion ghost can be effectively removed, the pyramid layer number of each pixel position is dynamically determined, the motion area is divided more finely, the ghost is judged more accurately, and edge noise is smaller.
Description
Technical Field
The invention belongs to the technical field of digital image processing, and particularly relates to a method for synthesizing a high dynamic range image from a low dynamic range image based on multi-frame multi-exposure.
Background
Dynamic range is an important index for vehicle-mounted camera applications, and a shot image is required to cover high-brightness and low-brightness details in a real scene, so that various High Dynamic Range (HDR) imaging technologies are widely researched. The multiple exposure HDR technology respectively shoots multiple frames of images by controlling the exposure duration to collect information of different brightness of a shot scene, and reconstructs an HDR image by multi-frame synthesis.
When a moving scene is shot, the moving area of the multi-frame images has displacement, so that the HDR images have ghost images in the moving area. A general method is to select a frame reference frame, calculate a synthesis weight based on the similarity between each synthesized frame and the reference frame, and reduce the ghost by reducing the synthesis weight of the difference region. The HDR image reconstruction technique based on pyramid decomposition can improve the ghost detection effect by calculating the combining weights of the multi-exposure images at different scales. However, after the number of pyramid decomposition layers is determined, the pyramid structure is fixed. Local differences diffuse to neighboring regions when weights are computed at a pyramid level to increase de-ghosting range, resulting in increased noise near edges. And (3) calculating the synthesis weight at the pyramid lower layer, wherein when the image noise is serious, the noise and the weak motion difference are difficult to distinguish, and the noise and the ghost removing effect are difficult to balance.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a method for synthesizing a high dynamic range image by a multi-frame multi-exposure low dynamic range image, which enables the image synthesized in a non-motion area to keep high brightness and low brightness details and reduce the noise of the synthesized image through a pyramid decomposition weight method, and enables the motion ghost to be removed in a motion area, thereby improving the ghost problem and the low brightness noise problem which are difficult to solve in multi-exposure synthesis and improving the HDR image effect.
The technical solution of the invention is as follows:
a method for synthesizing a high dynamic range image from low dynamic range images based on multi-frame and multi-exposure is characterized by comprising the following steps:
carrying out pyramid decomposition on the LDR images with different exposures of N frames;
acquiring a pyramid weight map, wherein the weight in a non-motion area is determined by the brightness value and the exposure value of an input frame, and the weight in a motion area is determined by the similarity between the input frame and a reference frame;
determining pyramid layer number of each pixel point according to the image gradient information and the similarity weight of each pyramid layer;
and carrying out pyramid synthesis and reconstruction from the pyramid layer number of the current pixel point to generate the HDR image.
The method comprises the following specific steps:
step 1: receiving N frames of images with different exposures, and establishing a Gaussian pyramid and a Laplacian pyramid with the number of layers being L for each frame of image.
And 2, step: and calculating the synthetic weight of the N frames of images at each pixel position of each layer of the pyramid to obtain a pyramid weight map of each frame of image. The weight calculation formula of the pixel point (i, j) on the ith layer of the pyramid is as follows:
wherein the content of the first and second substances,
is the exposure weight of each frame of image, and is calculated according to the exposure parameters of the current frame. The signal-to-noise ratio of the long exposure frame is generally better, the improvement of the signal-to-noise ratio is proportional to the exposure duration, and the short exposure frame is generally noisier due to the shorter exposure time. The long exposure frame is given a larger exposure weight and the short exposure frame is given a smaller exposure weight. For example:
E n representing exposure parameters of the nth frameAnd (4) counting.
Along with the increase of the pyramid layers, the influence of noise on the image subjected to multiple times of Gaussian filtering is gradually reduced, and different distance thresholds th are set according to the noise level and the pixel point brightness of each pyramid layer l (l)。
When the two images are similar, w n (i, j, l) is primarily related to the input frame exposure duration. When the difference between the two frames of images is larger, more reference frames are reserved for the synthesis result to reduce the ghost.
And step 3: and determining the pyramid layer number of the current pixel point according to the gradient information and the weight of each pyramid layer.
Step 3.1: calculating gradient information G = { G ] of each layer of the pyramid of the reference frame 0 ,G 1 ,...G L And determining that the current area belongs to a flat area or an edge according to the gradient.
Step 3.2: g l (L ≦ 0) when the current region is less than the threshold Th1, the current region is considered to be a flat region, at which time the weight calculated for the high level pyramid is considered more reliable, using a higher number of pyramid levels, e.g., L n (i, j) = L, and the weight of the next layer is updated using the weight of the previous layer.
Step 3.3: in the case of regions adjacent to the edge, a slight displacement of the edge will generally result in an increase in the similarity distance. When G is l+1 Greater than Th1, the gradient of adjacent layers increases by more than a certain threshold G l+1 -G l >Th2, and the similarity weight reduction is larger than a certain threshold
When using L as the number of HDR synthesis layers for the current point, L n (i,j)＝l。
Step 3.4: g l (0. Ltoreq. L. Ltoreq.L) points each larger than the threshold Th1 are regarded as points located on the edge. To prevent edge differences from being averaged, the similarity weight is increased significantly
When making L n (i,j)＝l。
And 4, step 4: number of layers L determined from step 3 n (i, j) begin with pyramid synthesis and reconstruction to generate the HDR image.
Step 4.1: from L n And (5) fusing N frames of input images by Gaussian images of the (i, j) layer, wherein the synthetic calculation formula of the position (i, j) is as follows:
wherein L is n (i, j, L) is the value of the pyramid coefficient of the L-th layer at the position (i, j) of the LDR image of the n-th frame, and h () represents L n (i, j, l) to correspond with HDR image luminance, w n (i, j, L) is the weight calculated by step 2, the starting pyramid level number L of location (i, j) n (i, j) is determined by step 3.
Step 4.2: at L n (i, j) -1 layer calculates the laplace coefficient synthesis result, and combines L n Pyramid upsampling of the (i, j) layer results, and matching with the L n (i, j) -1 layer results are added to give L n (i, j) -1 layer Gaussian composition coefficients.
Step 4.3: steps 4.1 and 4.2 are repeated until layer 0 composite results are calculated as output of the HDR image.
Compared with the prior art, the invention has the beneficial effects that
1) A method for generating an HDR image by a multi-exposure LDR image is provided, so that the dynamic range of an output image is improved, and motion ghosting is effectively removed.
2) The weight calculation method of the scheme can remove ghosting while HDR synthesis, and a separate ghost removing process is not needed.
3) By dynamically determining the pyramid layer number of each pixel position, the division of the motion area is finer, the ghost judgment is more accurate, and the edge noise is smaller.
Drawings
FIG. 1 is a flow chart of a method of synthesizing a high dynamic range image based on a low dynamic range image of a multi-frame multi-exposure in accordance with the present invention.
FIG. 2 is a schematic diagram of pyramid synthesis and reconstruction in the method for synthesizing a high dynamic range image based on a low dynamic range image of multi-frame multi-exposure according to the present invention.
Detailed Description
In order that those skilled in the art will better understand the disclosure, reference will now be made in detail to the embodiments of the disclosure as illustrated in the accompanying drawings. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
A method for synthesizing a high dynamic range image based on a multi-frame multi-exposure low dynamic range image is characterized in that an LDR image with N frames of different exposures is expanded to be mapped to an HDR domain, and the HDR image is generated through a linear synthesis method. The method comprises the following specific steps:
step 1: receiving N frames of images with different exposures, and establishing a Gaussian pyramid and a Laplacian pyramid with the number of layers being L for each frame of image.
And 2, step: and calculating the synthesis weight of the N frames of images at each pixel position of each layer of the pyramid to obtain a pyramid weight image of each frame of image. The weight calculation formula of the pixel point (i, j) on the ith layer of the pyramid is as follows:
wherein, the first and the second end of the pipe are connected with each other,
the exposure weight of each frame of image is calculated according to the exposure parameters of the current frame. The signal-to-noise ratio of the long exposure frame is generally better, the improvement of the signal-to-noise ratio is proportional to the exposure duration, and the short exposure frame is generally noisier due to the shorter exposure time. The long exposure frame is given a larger exposure weight and the short exposure frame is given a smaller exposure weight. For example:
E n indicating the exposure parameters for the nth frame.
Along with the increase of the pyramid layer number, the influence of noise on the image subjected to multiple times of Gaussian filtering is gradually reduced, and different distance thresholds th are set according to the noise level of each pyramid layer and the pixel point brightness l (l)。
When the two images are similar, w n (i, j, l) is primarily related to the input frame exposure duration. When the difference between the two frames of images is larger, more reference frames are reserved for the synthesis result to reduce the ghost.
And step 3: and determining the pyramid layer number of the current pixel point according to the gradient information and the weight of each pyramid layer.
Step 3.1: calculating gradient information G = { G) of each layer of reference frame pyramid 0 ,G 1 ,...G L And determining that the current area belongs to a flat area or an edge according to the gradient.
Step 3.2: g l (L ≦ 0 ≦ L) less than threshold Th1, the current region is considered to be flat, and the higher pyramid computed weights are considered more reliable, using a higher number of pyramid layers, e.g., L n (i, j) = L, and the weight of the next layer is updated using the weight of the previous layer.
Step 3.3: in the case of regions adjacent to the edge, a slight displacement of the edge will generally result in an increase in the similarity distance. When G is l+1 Greater than Th1, the gradient of adjacent layers increases by more than a certain threshold G l+1 -G l >Th2, and the similarity weight reduction is larger than a certain threshold
When using L as the number of HDR synthesis layers for the current point, L n (i,j)＝l。
Step 3.4: g l Points where (0. Ltoreq. L. Ltoreq.L) are each larger than the threshold Th1 are regarded as points located on the edge. To prevent edge differences from being averaged, the similarity weight is increased significantly
While making L n (i,j)＝l。
And 4, step 4: number of layers L determined from step 3 n (i, j) begin with pyramid synthesis and reconstruction to generate the HDR image.
Step 4.1: from L n The Gaussian images of the (i, j) layer start to be fused with N frames of input images, and the synthetic calculation formula of the position (i, j) is as follows:
wherein L is n (i, j, L) is the value of the pyramid coefficient of the L-th layer at the position (i, j) of the LDR image of the n-th frame, and h () represents L n (i, j, l) to correspond with HDR image luminance, w n (i, j, L) is the weight calculated by step 2, the starting pyramid level number L of location (i, j) n (i, j) is determined by step 3.
And 4.2: at L n (i, j) -1-layer calculation laplacian coefficient synthesisAs a result, L n Pyramid upsampling of the (i, j) layer results, and matching with Lth n (i, j) -1 layer results are added to give L n (i, j) -1 layer Gaussian composition coefficients.
Step 4.3: steps 4.1 and 4.2 are repeated until layer 0 composite results are calculated as output of the HDR image.
The invention provides a weight calculation method based on pyramid decomposition. The non-motion area weight calculation is determined by the input frame brightness value and the exposure value, so that the synthesized image can keep highlight and low-highlight details and reduce the noise of the synthesized image, and the motion area weight calculation is determined by the similarity of the input frame and the reference frame and is used for removing motion ghost. And adaptively determining the pyramid synthesis layer number of each pixel point according to the image gradient information and the similarity weight of each pyramid layer, so as to avoid the diffusion of noise near the edge when the pyramid layer number is fixed.
The above examples are only intended to illustrate the technical solution of the present invention, and not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (2)
1. A method for synthesizing a high dynamic range image based on multi-frame multi-exposure low dynamic range images is characterized by comprising the following steps:
carrying out pyramid decomposition on the LDR images with different exposures of N frames;
acquiring a pyramid weight map, wherein the weight in a non-motion area is determined by the brightness value and the exposure value of an input frame, and the weight in a motion area is determined by the similarity between the input frame and a reference frame;
determining pyramid layers of each pixel point according to the image gradient information and the similarity weight of each pyramid layer;
and (5) carrying out pyramid synthesis and reconstruction from the moment of determining the pyramid layer number of the current pixel point to generate the HDR image.
2. The method for synthesizing a high dynamic range image based on a multi-frame multi-exposure low dynamic range image according to claim 1, characterized by comprising the following specific steps:
step 1: receiving N frames of LDR images with different exposures, and establishing a Gaussian pyramid and a Laplacian pyramid with the number of layers being L for each frame of image;
step 2: calculating the synthetic weight of each pixel position of each pyramid layer of the N frames of images to obtain a pyramid weight map of each frame of image;
weight w of pixel point (I, j) on pyramid level I n (i, j, l), the formula is as follows:
in the formula (I), the compound is shown in the specification,
is the exposure weight for each frame of the image,
E n representing an exposure parameter for the nth frame;
is the luminance weight of the pixel location (i, j). Calculating the brightness value l of the current pixel point position n (i, j), the brightness value is the maximum value of a certain neighborhood taking (i, j) as the center. The luminance values are given less weight when they are close to overexposure and less weight when they are close to 0, the luminance weight being greatest in the middle segment, which ensures that the best exposed part of the LDR image per frame is used for HDR image reconstruction.
Is pixel point (i, j) at the ith layer of pyramidThe weight of the degree of similarity is given,
d n (i, j, l) n- 1 Distance of input frame from reference frame, th l (l) Is a distance threshold;
and 3, step 3: determining the pyramid layer number of the current pixel point according to the gradient information and the weight of each pyramid layer:
step 3.1: calculating gradient information G = { G) of each layer of reference frame pyramid 0 ，G 1 ，...，G L }；
Step 3.2: determining that the current region belongs to a flat region or an edge according to the gradient:
if G is 1 (L is more than or equal to 0 and less than or equal to L) is less than the threshold value Thl, then the current area is a flat area, namely L n (i, j) = L, and updates the weight of the next layer using the weight of the previous layer,
if G is 1+1 Greater than Th1, the gradient of adjacent layers increases by more than a certain threshold G 1+1 -G 1 > Th2, and the similarity weight reduction is also greater than a certain threshold
When, 1 is used as the HDR synthesis layer number of the current point, ln (i, j) = l;
if G1 (0 ≦ 1 ≦ L) are both greater than the threshold Th1, then the current region is the point located on the edge, i.e., L n (i，j)＝l；
And 4, step 4: number of layers L determined from step 3 n (i, j) starting, performing pyramid synthesis and reconstruction, and generating an HDR image:
step 4.1: from L n The Gaussian images of the (i, j) layer start to be fused with N frames of input images, and the synthesis formula of the position (i, j) is as follows:
wherein L is n (i, j, L) is the value of the L-th layer pyramid coefficient of the n-th frame LDR image at position (i, j), and h () represents L n (i, j, 1) transform to be consistent with HDR image luminance;
step 4.2: at L n (i, j) -1 layer calculating the Laplace coefficient synthesis result, and combining L n Pyramid upsampling of the (i, j) layer results, and matching with the L n (i, j) -1 layer results are added to give L n (i, j) -1 layer of Gaussian composition coefficients;
step 4.3: steps 4.1 and 4.2 are repeated until layer 0 composite results are calculated as output of the HDR image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211419101.3A CN115760663B (en) | 2022-11-14 | 2022-11-14 | Method for synthesizing high dynamic range image based on multi-frame multi-exposure low dynamic range image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211419101.3A CN115760663B (en) | 2022-11-14 | 2022-11-14 | Method for synthesizing high dynamic range image based on multi-frame multi-exposure low dynamic range image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115760663A true CN115760663A (en) | 2023-03-07 |
| CN115760663B CN115760663B (en) | 2023-09-22 |
Family
ID=85370150
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211419101.3A Active CN115760663B (en) | 2022-11-14 | 2022-11-14 | Method for synthesizing high dynamic range image based on multi-frame multi-exposure low dynamic range image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115760663B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117876252A (en) * | 2024-03-11 | 2024-04-12 | 上海玄戒技术有限公司 | Image noise reduction method, device, equipment, storage medium and chip |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130028509A1 (en) * | 2011-07-28 | 2013-01-31 | Samsung Electronics Co., Ltd. | Apparatus and method for generating high dynamic range image from which ghost blur is removed using multi-exposure fusion |
| CN104881854A (en) * | 2015-05-20 | 2015-09-02 | 天津大学 | High-dynamic-range image fusion method based on gradient and brightness information |
| US20150350513A1 (en) * | 2014-05-30 | 2015-12-03 | Apple Inc. | Constant Bracket High Dynamic Range (cHDR) Operations |
| CN107302668A (en) * | 2017-08-17 | 2017-10-27 | 中国人民解放军国防科技大学 | High dynamic range imaging module based on runner dynamic light splitting |
| 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 |
| CN109754377A (en) * | 2018-12-29 | 2019-05-14 | 重庆邮电大学 | A kind of more exposure image fusion methods |
| WO2021082580A1 (en) * | 2019-10-31 | 2021-05-06 | 北京迈格威科技有限公司 | Night scene high dynamic range image generation method, device, and electronic apparatus |
| CN113222954A (en) * | 2021-05-21 | 2021-08-06 | 大连海事大学 | Multi-exposure image ghost-free fusion method based on patch alignment under global gradient |
| CN113888447A (en) * | 2021-11-11 | 2022-01-04 | 努比亚技术有限公司 | Image processing method, terminal and storage medium |
| CN114998141A (en) * | 2022-06-07 | 2022-09-02 | 西北工业大学 | Space environment high dynamic range imaging method based on multi-branch network |
| CN115063331A (en) * | 2022-06-14 | 2022-09-16 | 安徽大学 | No-ghost multi-exposure image fusion algorithm based on multi-scale block LBP operator |
-
2022
- 2022-11-14 CN CN202211419101.3A patent/CN115760663B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130028509A1 (en) * | 2011-07-28 | 2013-01-31 | Samsung Electronics Co., Ltd. | Apparatus and method for generating high dynamic range image from which ghost blur is removed using multi-exposure fusion |
| US20150350513A1 (en) * | 2014-05-30 | 2015-12-03 | Apple Inc. | Constant Bracket High Dynamic Range (cHDR) Operations |
| CN104881854A (en) * | 2015-05-20 | 2015-09-02 | 天津大学 | High-dynamic-range image fusion method based on gradient and brightness information |
| CN107302668A (en) * | 2017-08-17 | 2017-10-27 | 中国人民解放军国防科技大学 | High dynamic range imaging module based on runner dynamic light splitting |
| 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 |
| CN109754377A (en) * | 2018-12-29 | 2019-05-14 | 重庆邮电大学 | A kind of more exposure image fusion methods |
| WO2021082580A1 (en) * | 2019-10-31 | 2021-05-06 | 北京迈格威科技有限公司 | Night scene high dynamic range image generation method, device, and electronic apparatus |
| CN113222954A (en) * | 2021-05-21 | 2021-08-06 | 大连海事大学 | Multi-exposure image ghost-free fusion method based on patch alignment under global gradient |
| CN113888447A (en) * | 2021-11-11 | 2022-01-04 | 努比亚技术有限公司 | Image processing method, terminal and storage medium |
| CN114998141A (en) * | 2022-06-07 | 2022-09-02 | 西北工业大学 | Space environment high dynamic range imaging method based on multi-branch network |
| CN115063331A (en) * | 2022-06-14 | 2022-09-16 | 安徽大学 | No-ghost multi-exposure image fusion algorithm based on multi-scale block LBP operator |
Non-Patent Citations (2)
| Title |
|---|
| LINGFENG WU ET.AL: "Details-preserving multi-exposure image fusion based on dual-pyramid using improved exposure evaluation", 《RESULTS IN OPTICS》, pages 1 - 10 * |
| 郝福得 等: "基于图像内容评价因子的动态场景曝光融合算法", 光学技术, vol. 44, no. 04, pages 480 - 486 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117876252A (en) * | 2024-03-11 | 2024-04-12 | 上海玄戒技术有限公司 | Image noise reduction method, device, equipment, storage medium and chip |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115760663B (en) | 2023-09-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112734650B (en) | Virtual multi-exposure fusion based uneven illumination image enhancement method | |
| CN107845128B (en) | Multi-exposure high-dynamic image reconstruction method with multi-scale detail fusion | |
| CN108492262B (en) | No-ghost high-dynamic-range imaging method based on gradient structure similarity | |
| TWI697867B (en) | Metering compensation method and related monitoring camera apparatus | |
| JP4598507B2 (en) | System and method for image noise reduction using minimum error space-time recursive filter | |
| WO2020143257A1 (en) | Motion artifact resistant hdr method and portable terminal | |
| CN108833775B (en) | HDR method and device for resisting motion ghost and portable terminal | |
| KR101664123B1 (en) | Apparatus and method of creating high dynamic range image empty ghost image by using filtering | |
| CN110503617B (en) | Underwater image enhancement method based on high-frequency and low-frequency information fusion | |
| JP6097588B2 (en) | Image processing apparatus and image processing method | |
| CN110852982A (en) | Self-adaptive exposure adjustment multi-scale entropy fusion underwater image enhancement method | |
| WO2021139635A1 (en) | Method and apparatus for generating super night scene image, and electronic device and storage medium | |
| CN105900138B (en) | Device, electronic equipment and method for enhancing image contrast | |
| CN105931213A (en) | Edge detection and frame difference method-based high-dynamic range video de-ghosting method | |
| CN110225260B (en) | Three-dimensional high dynamic range imaging method based on generation countermeasure network | |
| CN107911599B (en) | Infrared image global automatic focusing method and device | |
| CN115760663A (en) | Method for synthesizing high dynamic range image from low dynamic range image based on multi-frame multi-exposure | |
| CN113632134A (en) | High dynamic range image generation using pre-combination denoising | |
| JP6600216B2 (en) | Image processing apparatus, image processing method, program, and storage medium | |
| Zhang et al. | Non-uniform illumination video enhancement based on zone system and fusion | |
| JP2010130150A (en) | Gray level correction apparatus and imaging apparatus | |
| CN106375675B (en) | A kind of more exposure image fusion methods of aerial camera | |
| WO2014088259A1 (en) | Apparatus and method for generating hdr image using single image | |
| CN116416175A (en) | Image fusion method based on self-adaptive edge-preserving smooth pyramid | |
| CN108629739B (en) | HDR image generation method and device and mobile terminal |
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 |