CN117455766A - Image fusion method based on improved spelling line optimizing and smooth transition - Google Patents
Image fusion method based on improved spelling line optimizing and smooth transition Download PDFInfo
- Publication number
- CN117455766A CN117455766A CN202311743908.7A CN202311743908A CN117455766A CN 117455766 A CN117455766 A CN 117455766A CN 202311743908 A CN202311743908 A CN 202311743908A CN 117455766 A CN117455766 A CN 117455766A
- Authority
- CN
- China
- Prior art keywords
- image
- node
- fused
- line
- wavelet
- 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.)
- Pending
Links
- 230000007704 transition Effects 0.000 title claims abstract description 24
- 238000007500 overflow downdraw method Methods 0.000 title claims abstract description 16
- 230000004927 fusion Effects 0.000 claims abstract description 20
- 238000005192 partition Methods 0.000 claims abstract description 18
- 238000000034 method Methods 0.000 claims abstract description 14
- 238000004458 analytical method Methods 0.000 claims abstract description 6
- 230000000007 visual effect Effects 0.000 claims abstract description 5
- 230000009466 transformation Effects 0.000 claims description 17
- 238000009499 grossing Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000009191 jumping Effects 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 abstract 1
- 238000005516 engineering process Methods 0.000 description 7
- 230000000694 effects Effects 0.000 description 4
- 230000006872 improvement Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
Abstract
The invention provides an image fusion method based on improved splice line optimization and smooth transition, which comprises the following steps: searching an optimal partition in the image to be spliced by using an optimal partition searching algorithm with minimum path cost; determining an optimal splice line by comparing the influence of the pixel difference degree and the length of the splice line on human vision according to the structural relationship and the color relationship of the pixel point and other pixels in the neighborhood of the pixel point; decomposing and reconstructing the image through wavelet multi-resolution analysis, and realizing transition fusion through smooth function transition to eliminate the splice joint so as to obtain a panoramic image; by the method, the problems of color and brightness difference and obvious stitching line in the traditional image fusion method can be effectively solved, so that the fused image is clearer and more natural, and has a better visual effect.
Description
Technical Field
The invention relates to an image fusion technology, in particular to an image fusion method based on improved splice line optimizing and smooth transition.
Background
The image stitching refers to stitching a plurality of images respectively shot into a complete image, and the application scene of the image stitching includes panoramic shooting, virtual reality and the like. The transition region fusion in the image stitching refers to that edges and discontinuities are eliminated through various fusion technologies in the overlapping region between different images, so that the stitched images are more natural and smooth and have good visual effects. In the image splicing process, problems such as inconsistent color, illumination difference, motion object residual shadows and the like can occur due to errors and differences caused by various factors, the problems need to be treated through fusion technologies, the current common fusion methods comprise linear fusion, gaussian fusion, laplacian pyramid fusion and the like, and researchers have made certain progress through constantly improved image fusion technologies; for example, in 2008, shenJB et al propose an improved LP transform method, which can achieve a better fusion effect on a multi-exposure image, and retain more texture detail information; in 2015, naveen et al successfully realized 2D depth image fusion using discrete wavelet transform techniques; in recent years, the continuous improvement of the image fusion technology has led to an improvement in the effect of eliminating the splice line gap and the ghost phenomenon. However, the conventional image fusion method mainly performs fusion processing on images with overlapping areas, is not completely consistent with the sequential image stitching and fusion requirements required by the image stitching technology, has some problems in the application process, and particularly has strict restriction on the optimizing direction of stitching lines.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides an image fusion method based on improved splice line optimizing and smooth transition, which is further improved and combined with a new technology to meet the requirements of image splice fusion, eliminate splice line gaps and realize smooth natural processing of images.
An image fusion method based on improved splice line optimizing and smooth transition comprises the following steps:
step 1: searching an optimal partition in the image to be spliced by using an optimal partition searching algorithm with minimum path cost;
step 2: determining an optimal splice line by comparing the influence of the pixel difference degree and the length of the splice line on human vision according to the structural relationship and the color relationship of the pixel point and other pixels in the neighborhood of the pixel point;
step 3: decomposing and reconstructing the image through wavelet multi-resolution analysis, and realizing transition fusion through smooth function transition to eliminate the splice joint, so as to obtain the panoramic image.
The method further comprises the following steps: step 1 comprises the following steps:
step 1.1: the normalized cross-correlation NCC is used to evaluate the degree of matching of pixels on the stitching line, and the normalized cross-correlation NCC of a pixel point (x, y) between two images is calculated by a 5×5 sub-graph and according to the following formula:
wherein,and->Is the average of 25 pixel values in a 5 x 5 window over the two images; />Andis +.>Pixel value of location +.>；
Step 1.2: by usingRepresenting pixel +.>Degree of mismatch, ++>；
Step 1.3: statisticsNumber d of->For maximum mismatch of human visual acceptance, when the number d is greater than 1, then +.>Subtracting 1, and re-executing 1.3; when the number d is 1, jumping to the step 1.4;
step 1.4: order theWherein->；
Step 1.5: each pixel point in the overlapping area of the two images to be fused is respectively marked as a node, and the pixel points in the overlapping area are respectively markedIs marked as node->Node->For adjacent nodes of (a)Representation of->，/>The method comprises the steps of carrying out a first treatment on the surface of the Is provided with->Representation->Is provided with all neighboring nodes ofRepresenting node->Global minimum path to end node, set +.>Representing nodesA minimum path to an end node;
the calculation formula of (2) is shown as follows:
wherein,all represent->And->A path therebetween;
step 1.6: marking the initial node in the subgraph as a visited node, and marking other nodes as unviewed nodes;
step 1.7: finding a node that is not visited and that is visited can form a path that is near or equal to the global minimum, bothMarking the non-access node with the minimum value as the accessed node, marking the path formed by connecting the accessed nodes in sequence as the minimum cost path, and executing the step 1.8 when the minimum cost path reaches the end node in the subgraph; otherwise, repeating the step 1.7;
step 1.8: and extracting pixels corresponding to the minimum cost route and forming an optimal partition.
The method further comprises the following steps: step 2 comprises the following steps:
step 2.1: all on the least cost path in the best partitionThe values are ordered in descending order and are taken as the original cost set, which is noted +.>The descending function is->The result after ordering is +.>And (2) and，/>the measurement mode of (2) is shown as follows:
wherein,representing the length of the splice line>Is weight(s)>Is a fixed value (+)>Is determined according to the cost, if the gap between the cost is large and the interval exceeds 1, there are several +.>Just a few, if there is no gap, then +.>Number of costs);
step 2.2: and selecting the least cost route with the minimum VI as the optimal splicing line.
The method further comprises the following steps: step 3 comprises the following steps:
step 3.1: decompressing the two images to be fused, acquiring data of R, G, B channels, constructing a wavelet filter bank through Daubechies wavelets, convoluting and downsampling the current color channel;
step 3.2: if all the R, G, B color channels are processed, normalizing the wavelet coefficients obtained after transformation to obtain intermediate coefficients, and executing the step 3.3; otherwise, executing the step 3.1;
step 3.3: performing inverse transformation on the intermediate coefficient, performing row reconstruction and column reconstruction on the current color channel, and obtaining two registered images A to be fused, an image B to be fused and a mask image C containing the optimal splicing line;
step 3.4: carrying out k layers of Daubechies wavelet transformation on the image A to be fused and the image B to be fused to obtain 3k+1 band-pass images A 'and 3k+1 band-pass images B', wherein k is the number of layers of the wavelet transformation;
step 3.5: downsampling the mask image C to obtain a reference image C';
step 3.6: and fusing the overlapping area of the band-pass image A 'and the band-pass image B' by using a cosine weighted smoothing function to obtain a fused image, and performing k-layer wavelet inverse transformation on the fused image to obtain a panoramic image.
The method further comprises the following steps: the cosine weighted smoothing function is:
in the middle of，/>Representing the width of the overlap area of each row of image stitching line in the k-th wavelet transform band-pass image,/and>represents the abscissa on the optimal stitching line in the kth layer wavelet transformed band-pass image, +.>Represents the abscissa of the pixel point in the kth layer wavelet transform band-pass image, +.>Represents the firstThe height of the k-layer bandpass image.
The invention has the beneficial effects that: firstly, searching an optimal partition in a spliced image by using an optimal partition searching and improving algorithm with minimum path cost, considering the structural relation and the color relation of pixel points and other pixels in the neighborhood of the pixel points, and determining an optimal splicing line by comparing the pixel difference degree and the influence of the length of the splicing line on human vision, thereby not only improving the accuracy of the algorithm, but also taking the efficiency into consideration; then, combining a wavelet multi-resolution analysis and smooth function transition fusion algorithm to eliminate 'ghosts' and reduce 'splice seams', so as to obtain a panoramic image with better effect, decomposing and reconstructing the image by utilizing the wavelet multi-resolution analysis, and realizing transition fusion by smooth function transition, so that the image splicing quality can be greatly improved, the flexible searching direction and the capability of giving priority can be realized, and the high efficiency and the accuracy can be considered, so that the operation efficiency of the algorithm can be improved.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The present invention will be described in detail with reference to the accompanying drawings. Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative only and are not to be construed as limiting the invention. The terms left, middle, right, upper, lower, etc. in the embodiments of the present invention are merely relative concepts or references to the normal use state of the product, and should not be construed as limiting.
An image fusion method based on improved splice line optimizing and smooth transition comprises the following steps:
step 1: searching an optimal partition in the image to be spliced by using an optimal partition searching algorithm with minimum path cost; the step of obtaining the optimal partition includes:
step 1.1: the normalized cross-correlation NCC is used to evaluate the degree of matching of pixels on the stitching line, and the normalized cross-correlation NCC of a pixel point (x, y) between two images is calculated by a 5×5 sub-graph and according to the following formula:
wherein,and->Is the average of 25 pixel values in a 5 x 5 window over the two images; />Andis +.>Pixel value of location +.>；
Step 1.2: by usingRepresenting pixel +.>Degree of mismatch, ++>；
Step 1.3: statisticsNumber d of->The maximum mismatch degree acceptable by human vision is set to 7 by presetting, and the maximum mismatch degree is takenThe value ranges from 1 to 10, most preferably from 5 to 10, when the number d is greater than 1, then +.>Subtracting 1, and re-executing 1.3; when the number d is 1, jumping to the step 1.4;
step 1.4: order theWherein->；
Step 1.5: each pixel point in the overlapping area of the two images to be fused is respectively marked as a node, and the pixel points in the overlapping area are respectively markedIs marked as node->Node->For adjacent nodes of (a)>Representation of->，/>The method comprises the steps of carrying out a first treatment on the surface of the Is provided with->Representation->Is set to->Representing node->Global best to end nodeSmall path, set->Representing node->A minimum path to an end node;
the calculation formula of (2) is shown as follows:
wherein,all represent->And->A path therebetween;
step 1.6: marking the initial node in the subgraph as a visited node, and marking other nodes as unviewed nodes;
step 1.7: finding a node that is not visited and that is visited can form a path that is near or equal to the global minimum, bothMarking the non-access node with the minimum value as the accessed node, marking the path formed by connecting the accessed nodes in sequence as the minimum cost path, and executing the step 1.8 when the minimum cost path reaches the end node in the subgraph; otherwise, repeating the step 1.7;
step 1.8: extracting pixels corresponding to the minimum cost route and forming an optimal partition;
step 2: determining an optimal splice line by comparing the influence of the pixel difference degree and the length of the splice line on human vision according to the structural relationship and the color relationship of the pixel point and other pixels in the neighborhood of the pixel point; the step of obtaining the optimal partition includes:
step 2.1: all on the least cost path in the best partitionThe values are ordered in descending order and are taken as the original cost set, which is noted +.>The descending function is->The result after ordering is +.>And (2) and，/>the measurement mode of (2) is shown as follows:
wherein,representing the length of the splice line>Is weight(s)>Is a fixed value, i.e. a preset value +.>Is determined according to the cost, if the gap between the cost is large and the interval exceeds 1, there are several +.>Just a few, if there is no gap, then +.>The number of the cost;
step 2.2: selectingThe smallest minimum cost route is used as the optimal splicing line;
step 3: decomposing and reconstructing the image through wavelet multi-resolution analysis, and realizing transition fusion through smooth function transition to eliminate the splice joint, so as to obtain the panoramic image, which is specifically as follows:
step 3.1: decompressing the two images to be fused, acquiring data of R, G, B channels, constructing a wavelet filter bank through Daubechies wavelets, convoluting and downsampling the current color channel;
step 3.2: if all the R, G, B color channels are processed, normalizing the wavelet coefficients obtained after transformation to obtain intermediate coefficients, and executing the step 3.3; otherwise, executing the step 3.1;
step 3.3: performing inverse transformation on the intermediate coefficient, performing row reconstruction and column reconstruction on the current color channel, and obtaining two registered images A to be fused, an image B to be fused and a mask image C containing the optimal splicing line;
step 3.4: carrying out k layers of Daubechies wavelet transformation on the image A to be fused and the image B to be fused to obtain 3k+1 band-pass images A 'and 3k+1 band-pass images B', wherein k is the number of layers of the wavelet transformation, the more the number of layers of the wavelet transformation is, the more the transformation is complicated, the more the time is consumed, and the image fusion can be satisfied when the value of k is 3; the value of k can be 2, but the effect is poor, and when the value of k is more than 3, the time consumption is high;
step 3.5: since each layer of band-pass sub-image is the upper layer k/4, the mask image C is downsampled to obtain a reference image；
Step 3.6: fusing the overlapping area of the band-pass image A 'and the band-pass image B' by using a cosine weighted smoothing function to obtain a fused image, and performing k-layer wavelet inverse transformation on the fused image to obtain a panoramic image;
the cosine weighted smoothing function is:
in the middle of，/>Representing the width of the overlap area of each row of image stitching line in the k-th wavelet transform band-pass image,/and>represents the abscissa on the optimal stitching line in the kth layer wavelet transformed band-pass image, +.>Represents the abscissa of the pixel point in the kth layer wavelet transform band-pass image, +.>Representing the height of the kth layer bandpass image.
By the method, the problems of color and brightness difference and obvious stitching line in the traditional image fusion method can be effectively solved, so that the fused image is clearer and more natural, and has a better visual effect.
The foregoing has shown and described the basic principles, principal features and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, and that the above embodiments and descriptions are merely illustrative of the principles of the present invention, and various changes and modifications may be made without departing from the spirit and scope of the invention, which is defined in the appended claims. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (5)
1. The image fusion method based on improved splice line optimizing and smooth transition is characterized by comprising the following steps of: the method comprises the following steps:
step 1: searching an optimal partition in the image to be spliced by using an optimal partition searching algorithm with minimum path cost;
step 2: determining an optimal splice line by comparing the influence of the pixel difference degree and the length of the splice line on human vision according to the structural relationship and the color relationship of the pixel point and other pixels in the neighborhood of the pixel point;
step 3: decomposing and reconstructing the image through wavelet multi-resolution analysis, and realizing transition fusion through smooth function transition to eliminate the splice joint, so as to obtain the panoramic image.
2. The improved stitching line optimizing and smooth transition based image fusion method of claim 1, wherein: step 1 comprises the following steps:
step 1.1: the normalized cross-correlation NCC is used to evaluate the degree of matching of pixels on the stitching line, and the normalized cross-correlation NCC of a pixel point (x, y) between two images is calculated by a 5×5 sub-graph and according to the following formula:
；
wherein,and->Is the average of 25 pixel values in a 5 x 5 window over the two images; />And->Is two-webImage +.>Pixel value of location +.>；
Step 1.2: by usingRepresenting pixel +.>Degree of mismatch, ++>；
Step 1.3: statisticsThe number of (2)>Maximum degree of mismatch for human visual acceptance, when number +.>When the number is greater than 1, the number is +.>Subtracting 1, and re-executing 1.3; when number->If the value is 1, jumping to the step 1.4;
step 1.4: order theWherein->；
Step 1.5: two webs are combinedEach pixel point in the overlapping area of the images to be fused is respectively marked as a node, and the pixel points in the overlapping area are respectively markedIs marked as node->Node->For adjacent nodes of (a)>Representation of->The method comprises the steps of carrying out a first treatment on the surface of the Is provided with->Representation->Is set to->Representing nodesGlobal minimum path to end node, set +.>Representing node->A minimum path to an end node;
the calculation formula of (2) is shown as follows:
；
wherein,all represent->And->A path therebetween;
step 1.6: marking the initial node in the subgraph as a visited node, and marking other nodes as unviewed nodes;
step 1.7: finding a node that is not visited and that is visited can form a path that is near or equal to the global minimum, bothMarking the non-access node with the minimum value as the accessed node, marking the path formed by connecting the accessed nodes in sequence as the minimum cost path, and executing the step 1.8 when the minimum cost path reaches the end node in the subgraph; otherwise, repeating the step 1.7;
step 1.8: and extracting pixels corresponding to the minimum cost route and forming an optimal partition.
3. The improved stitching line optimizing and smooth transition based image fusion method of claim 2, wherein: step 2 comprises the following steps:
step 2.1: all on the least cost path in the best partitionThe values are ordered in descending order and are taken as the original cost set, which is noted +.>Descending order ofThe function is->The result after ordering is +.>And (2) and，/>the measurement mode of (2) is shown as follows:
；
wherein,representing the length of the splice line>Is weight(s)>Is a fixed value;
step 2.2: and selecting the least cost route with the minimum VI as the optimal splicing line.
4. The improved stitching line optimizing and smoothing transition based image fusion method of claim 3, wherein: step 3 comprises the following steps:
step 3.1: decompressing the two images to be fused, acquiring data of R, G, B channels, constructing a wavelet filter bank through Daubechies wavelets, convoluting and downsampling the current color channel;
step 3.2: if all the R, G, B color channels are processed, normalizing the wavelet coefficients obtained after transformation to obtain intermediate coefficients, and executing the step 3.3; otherwise, executing the step 3.1;
step 3.3: performing inverse transformation on the intermediate coefficient, performing row reconstruction and column reconstruction on the current color channel, and obtaining registered image A to be fused, registered image B to be fused and mask image C containing the optimal splicing line;
step 3.4: carrying out k layers of Daubechies wavelet transformation on the image A to be fused and the image B to be fused to obtain 3k+1 band-pass images A 'and 3k+1 band-pass images B', wherein k is the number of layers of the wavelet transformation;
step 3.5: downsampling the mask image C to obtain a reference image C';
step 3.6: and fusing the overlapping area of the band-pass image A 'and the band-pass image B' by using a cosine weighted smoothing function to obtain a fused image, and performing k-layer wavelet inverse transformation on the fused image to obtain a panoramic image.
5. The improved stitching line optimizing and smoothing transition based image fusion method of claim 4, wherein:
the cosine weighted smoothing function is:
；
in the middle of，/>Representing the width of the overlap area of each row of image stitching line in the k-th wavelet transform band-pass image,/and>represents the abscissa on the optimal stitching line in the kth layer wavelet transformed band-pass image, +.>Represents the abscissa of the pixel point in the kth layer wavelet transform band-pass image, +.>Representing the height of the kth layer bandpass image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311743908.7A CN117455766A (en) | 2023-12-19 | 2023-12-19 | Image fusion method based on improved spelling line optimizing and smooth transition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311743908.7A CN117455766A (en) | 2023-12-19 | 2023-12-19 | Image fusion method based on improved spelling line optimizing and smooth transition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117455766A true CN117455766A (en) | 2024-01-26 |
Family
ID=89591223
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311743908.7A Pending CN117455766A (en) | 2023-12-19 | 2023-12-19 | Image fusion method based on improved spelling line optimizing and smooth transition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117455766A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101110122A (en) * | 2007-08-31 | 2008-01-23 | 北京工业大学 | Large cultural heritage picture pattern split-joint method based on characteristic |
| CN103279939A (en) * | 2013-04-27 | 2013-09-04 | 北京工业大学 | Image stitching processing system |
| US20150154776A1 (en) * | 2013-12-03 | 2015-06-04 | Huawei Technologies Co., Ltd. | Image splicing method and apparatus |
| CN107240082A (en) * | 2017-06-23 | 2017-10-10 | 微鲸科技有限公司 | A kind of splicing line optimization method and equipment |
| US20180101931A1 (en) * | 2016-10-10 | 2018-04-12 | Gopro, Inc. | Apparatus and methods for the optimal stitch zone calculation of a generated projection of a spherical image |
| CN108198135A (en) * | 2018-01-02 | 2018-06-22 | 佛山科学技术学院 | A kind of optimal suture line search method of Panorama Mosaic |
| CN115619636A (en) * | 2022-09-15 | 2023-01-17 | 浙江大华技术股份有限公司 | Image stitching method, electronic device and storage medium |
-
2023
- 2023-12-19 CN CN202311743908.7A patent/CN117455766A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101110122A (en) * | 2007-08-31 | 2008-01-23 | 北京工业大学 | Large cultural heritage picture pattern split-joint method based on characteristic |
| CN103279939A (en) * | 2013-04-27 | 2013-09-04 | 北京工业大学 | Image stitching processing system |
| US20150154776A1 (en) * | 2013-12-03 | 2015-06-04 | Huawei Technologies Co., Ltd. | Image splicing method and apparatus |
| US20180101931A1 (en) * | 2016-10-10 | 2018-04-12 | Gopro, Inc. | Apparatus and methods for the optimal stitch zone calculation of a generated projection of a spherical image |
| CN107240082A (en) * | 2017-06-23 | 2017-10-10 | 微鲸科技有限公司 | A kind of splicing line optimization method and equipment |
| CN108198135A (en) * | 2018-01-02 | 2018-06-22 | 佛山科学技术学院 | A kind of optimal suture line search method of Panorama Mosaic |
| CN115619636A (en) * | 2022-09-15 | 2023-01-17 | 浙江大华技术股份有限公司 | Image stitching method, electronic device and storage medium |
Non-Patent Citations (4)
| Title |
|---|
| 何玲: "基于运动视频的全景图像拼接系统设计与实现", 中国优秀硕士学位论文全文数据库 信息科技辑, 15 January 2023 (2023-01-15), pages 3 * |
| 宋宝森: "全景图像拼接方法研究与实现", 中国博士学位论文全文数据库信息科技辑, 15 June 2013 (2013-06-15), pages 3 * |
| 李婉;赵双明;张卫龙;刘骁;喻国荣;: "一种无人机影像滤波分频拼接算法", 武汉大学学报(信息科学版), no. 06, 14 April 2017 (2017-04-14) * |
| 王军;朱宝山;朱述龙;晋克宇;汤志强;: "基于相关系数的遥感图像拼接线检测算法", 测绘与空间地理信息, no. 03, 25 June 2011 (2011-06-25) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111062872B (en) | Image super-resolution reconstruction method and system based on edge detection | |
| Yang et al. | Unsupervised learning of geometry with edge-aware depth-normal consistency | |
| Gilbert et al. | Volumetric performance capture from minimal camera viewpoints | |
| Chi et al. | Single image reflection removal using deep encoder-decoder network | |
| Liu et al. | Switchable temporal propagation network | |
| CN102982520B (en) | Robustness face super-resolution processing method based on contour inspection | |
| Liu et al. | Unpaired learning for deep image deraining with rain direction regularizer | |
| CN111091503A (en) | Image out-of-focus blur removing method based on deep learning | |
| CN112991231B (en) | Single-image super-image and perception image enhancement joint task learning system | |
| CN111553841B (en) | Real-time video splicing method based on optimal suture line updating | |
| CN113284061B (en) | Underwater image enhancement method based on gradient network | |
| CN105701770B (en) | A kind of human face super-resolution processing method and system based on context linear model | |
| CN105335930B (en) | The robustness human face super-resolution processing method and system of edge data driving | |
| Chen et al. | Remote sensing image super-resolution via residual aggregation and split attentional fusion network | |
| CN109919832A (en) | One kind being used for unpiloted traffic image joining method | |
| Sun et al. | Deep maximum a posterior estimator for video denoising | |
| Chen et al. | Image denoising via deep network based on edge enhancement | |
| Niu et al. | Progressive moire removal and texture complementation for image demoireing | |
| CN116797488A (en) | Low-illumination image enhancement method based on feature fusion and attention embedding | |
| CN116563693A (en) | Underwater image color restoration method based on lightweight attention mechanism | |
| Almasri et al. | Rgb guided thermal super-resolution enhancement | |
| CN116757988B (en) | Infrared and visible light image fusion method based on semantic enrichment and segmentation tasks | |
| CN112598604A (en) | Blind face restoration method and system | |
| CN117455766A (en) | Image fusion method based on improved spelling line optimizing and smooth transition | |
| CN116934592A (en) | Image stitching method, system, equipment and medium based on deep learning |
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 |