CN114418920A - Endoscope multi-focus image fusion method - Google Patents
Endoscope multi-focus image fusion method Download PDFInfo
- Publication number
- CN114418920A CN114418920A CN202210320964.9A CN202210320964A CN114418920A CN 114418920 A CN114418920 A CN 114418920A CN 202210320964 A CN202210320964 A CN 202210320964A CN 114418920 A CN114418920 A CN 114418920A
- Authority
- CN
- China
- Prior art keywords
- image
- images
- corrected
- color
- light
- 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
- 238000007500 overflow downdraw method Methods 0.000 title claims abstract description 24
- 230000004927 fusion Effects 0.000 claims abstract description 11
- 238000007499 fusion processing Methods 0.000 claims abstract description 8
- 238000000605 extraction Methods 0.000 claims abstract description 7
- 238000000034 method Methods 0.000 claims description 27
- 230000009466 transformation Effects 0.000 claims description 15
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000000844 transformation Methods 0.000 claims 2
- 238000003702 image correction Methods 0.000 claims 1
- 238000011426 transformation method Methods 0.000 claims 1
- 238000003384 imaging method Methods 0.000 description 9
- 230000000694 effects Effects 0.000 description 5
- 230000001575 pathological effect Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 210000001035 gastrointestinal tract Anatomy 0.000 description 2
- 230000008855 peristalsis Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration by the use of more than one image, e.g. averaging, subtraction
-
- G06T5/90—
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10068—Endoscopic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Abstract
The invention discloses an endoscope multi-focus image fusion method, which belongs to the technical field of image data processing and comprises the following steps: an image acquisition step, wherein images are acquired under different focal lengths respectively to obtain a plurality of original images; a characteristic point extraction step, which is to respectively extract the characteristic points of each original image; a characteristic matching step, namely obtaining matching point pairs between every two original images; determining a reference image, and searching the reference image; correcting all original images to the plane of the reference image to obtain a plurality of corrected images; a step of light and color homogenizing treatment, namely obtaining a corrected image after light and color homogenizing; and an image fusion step, namely performing fusion processing on the corrected images after light and color uniformization to obtain a fused image. According to the endoscope multi-focus image fusion method, the collected images with different focus distances are fused to obtain an image containing clear positions of all the collected images, and a user can observe the image conveniently.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to an endoscope multi-focus image fusion method.
Background
The endoscope is used as a detection instrument based on an image sensor, can see lesions which cannot be displayed by X-rays, and is widely applied to the field of medical treatment. Most of the existing endoscope photosensitive devices are CCD or CMOS, the imaging principle is pinhole imaging, and the imaging clear range is limited by the depth of field.
In addition, when an endoscope is used for imaging in some organs of a human body, a specific pathological area needs to be found, but because the intestinal tract in the body of the human body is always in a peristalsis state, the resolution of the shot pathological area is extremely low, and even along with the peristalsis of the intestinal tract, the pathological area cannot be kept in the middle of the image, so that the imaging is inconvenient for a doctor to view.
Due to the defects of the image acquisition equipment or poor image acquisition environment, the imaging quality of the endoscope is poor, and under the condition that the image acquisition equipment and the image acquisition environment cannot be changed, the obtained image can be further processed, so that the original image which is not clear enough is restored, or certain characteristic information in the image is highlighted for processing.
Based on the above, the invention provides an endoscope multi-focus image fusion method, which mainly solves the technical problems of small imaging definition range and poor definition of medical images in an image processing mode.
The above information disclosed in this background section is only for enhancement of understanding of the background of the application and therefore it may comprise prior art that does not constitute known to a person of ordinary skill in the art.
Disclosure of Invention
The invention provides an endoscope multi-focus image fusion method for improving the quality of an image acquired and output by an endoscope, aiming at the technical problem of poor imaging quality of the endoscope caused by the defects of image acquisition equipment or poor image acquisition environment in the prior art.
In order to realize the purpose of the invention, the invention is realized by adopting the following technical scheme:
an endoscopic multifocal image fusion method comprising:
an image acquisition step, wherein images are acquired under different focal lengths respectively to obtain a plurality of original images;
a characteristic point extraction step, which is to respectively extract the characteristic points of each original image;
a characteristic matching step, namely performing characteristic matching between every two original images according to the characteristic point pairs to obtain matching point pairs between every two original images;
determining a reference image, and finding out an original image with the most matching points with other original images as the reference image;
correcting all original images to the plane of the reference image to obtain a plurality of corrected images;
a step of light and color homogenizing treatment, which is to respectively carry out light and color homogenizing treatment on the corrected image to obtain a corrected image after light and color homogenizing;
and an image fusion step, namely performing fusion processing on the corrected images after light and color uniformization to obtain a fused image.
Further, in the image correcting step, a homography transformation from other original images except the reference image to the reference image is calculated by adopting a method robust to error matching, and the other original images except the reference image are corrected to the plane of the reference image by using the homography transformation.
Further, the step of light and color homogenizing treatment comprises the following steps:
dividing the plane of the reference image into a plurality of image blocks;
calculating the gray level histogram of each corrected image in the image block in r, g and b channels,Wherein k is the number of the corrected image, and p and q represent the row number and the column number of the image block;
calculating the sum of the gray level histograms of all corrected images in the image block at r, g and b channels respectively, and recording the sum as,;
and respectively carrying out light and color evening calculation on each corrected image according to the transformation coefficient.
a and b are transform coefficients, respectively.
Further, the method for homogenizing light and color of each corrected image comprises the following steps:
which represents the correction of the image or images,representing the corrected image after the dodging and the evening, (i, j) representing the pixel coordinates,representing image blocksOf (2) centerThe coordinates of the pixels are then compared to the coordinates of the pixels,as a weight value, the weight value,representing a center-distance pixel of a corrected image block (p, q)The distance of (a) to (b),is the size of the image block or blocks,is the number of rows of the image block corresponding to the ith row of the corrected image,and correcting the number of columns of the image block corresponding to the jth column of the image.
Further, in the image fusion step, the corrected images after light and color homogenizing are fused by adopting a weighted sum method.
Further, the fusion processing method comprises the following steps:
and (3) calculating the contrast of each pixel in the corrected image after the light and color are homogenized by using a Laplace operator:
and (3) calculating the saturation of each pixel by using the standard deviation of three channels of r, g and b after dodging and color evening:
Further, any one of a SIFT algorithm, an HOG algorithm and an ORB algorithm is adopted in the feature point extraction step.
Further, in the step of feature matching, a correlation coefficient feature matching method or a Euclidean distance feature matching method is adopted for feature matching.
Compared with the prior art, the invention has the advantages and positive effects that:
according to the endoscope multi-focus image fusion method, the collected images with different focus distances are fused to obtain an image containing clear positions of all the collected images, and a user can observe the image conveniently.
Compared with a method for directly weighting and summing a plurality of images, the method effectively eliminates the plaque effect, enables the hue and the brightness of the fused image to be more consistent, and enables the whole image to be more natural.
Compared with a pyramid or wavelet transform fusion method, the method has the advantages that the fusion image plaque effect is eliminated, and meanwhile, the calculation amount is remarkably reduced.
Other features and advantages of the present invention will become more apparent from the following detailed description of the invention when taken in conjunction with the accompanying drawings.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a flowchart of an embodiment of an endoscopic multi-focus image fusion method according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present 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.
Example one
The present embodiment proposes an endoscope multi-focus image fusion method, as shown in fig. 1, including:
and an image acquisition step, wherein images are acquired under different focal lengths respectively to obtain a plurality of original images.
Firstly, an endoscope is used for collecting images at different focal distances, and a point feature extraction method is used for extracting feature points of each image to obtain feature description of the feature points.
And a characteristic point extraction step, which is to respectively extract the characteristic points of each original image.
Methods that can be used to extract features include, but are not limited to, SIFT algorithm, HOG algorithm, ORB algorithm, and the like.
And a characteristic matching step, namely performing characteristic matching between every two original images according to the characteristic point pairs to obtain matching point pairs between every two original images.
In the feature matching step, a correlation coefficient feature matching method or a euclidean distance feature matching method can be adopted for feature matching.
And determining a reference image, and finding out the original image with the most matching points with other original images as the reference image.
And correcting all the original images to the plane of the reference image to obtain a plurality of corrected images.
And a step of light and color homogenizing treatment, which is to respectively carry out light and color homogenizing treatment on the corrected image to obtain the corrected image after light and color homogenizing.
And an image fusion step, namely performing fusion processing on the corrected images after light and color uniformization to obtain a fused image.
According to the endoscope multi-focus image fusion method, the condition that the endoscope collects a plurality of images with different focus distances is improved, a plurality of clear images at different positions are obtained, and the observation of a user is facilitated.
Compared with a method for directly weighting and summing a plurality of images, the method effectively eliminates the plaque effect, enables the hue and the brightness of the fused image to be more consistent, and enables the whole image to be more natural.
Compared with a pyramid or wavelet transform fusion method, the method has the advantages that the fusion image plaque effect is eliminated, and meanwhile, the calculation amount is remarkably reduced.
In some embodiments, in the image correcting step, homographic transformation of other original images than the reference image to the reference image is calculated by using a method robust to error matching, and the other original images than the reference image are corrected to the plane of the reference image by using the homographic transformation. Through the step, the corrected image homonymous points have the same texture coordinates, and the homonymous points or homonymous areas can be conveniently and quickly found through uniform light and uniform color histogram statistics and final image fusion.
In some embodiments, the RANSAC algorithm may be used to calculate a homographic transformation of the other image to the reference image, and the homographic transformation may be used to correct the other image to the plane of the reference image.
In some embodiments, the step of homogenizing comprises:
the plane of the reference image is divided into a plurality of image blocks.
In order to increase the calculation speed, in some embodiments, the plane of the reference image is divided into a plurality of image blocks with the same size.
Calculating the gray histogram of each corrected image in the image block in r, g and b channels,,Where k is the number of the correction image, and p and q denote the row number and column number of the image block.
Calculating the sum of the gray level histograms of all corrected images in the image block at r, g and b channels, and recording the sum as , 。
And respectively carrying out light and color evening calculation on each corrected image according to the transformation coefficient.
Through linear transformation, the brightness and the tone of the image after light and color uniformization are basically consistent, and the plaque effect caused by the color or the brightness of the fused image is avoided.
a and b are transform coefficients, respectively.
The transform coefficients of the same image in one image block have the same value of a and the same value of b. The r, g and b channel a values in the same image block are more stable than the a value.
In some embodiments, the method of homogenizing the respective correction images is:
which represents the correction of the image or images,representing the corrected image after the dodging and the evening, (i, j) representing the pixel coordinates,representing image blocksThe coordinates of the center pixel of (a),as a weight value, the weight value,representing a center-distance pixel of a corrected image block (p, q)The distance of (a) to (b),is the size of the image block or blocks,is the number of rows of the image block corresponding to the ith row of the corrected image,and correcting the number of columns of the image block corresponding to the jth column of the image.
The transformation parameters representing the image block (p, q) influence the magnitude of the gray value at (i, j). The further an image block (p, q) is from a pixel (i, j), the smaller the impact.
In some embodiments, in the image fusion step, the corrected image after dodging and evening is subjected to fusion processing by adopting a weighted sum method. And obtaining a fused image containing clear positions of the plurality of images.
In some embodiments, the fusion processing method is:
and (3) calculating the contrast of each pixel in the corrected image after the light and color are homogenized by using a Laplace operator:
the image fusion step can distribute higher weight to pixels with high saturation and high contrast, so that the fused image comprises the positions of clear imaging and high saturation of each image.
In some embodiments, the 3 × 3 laplacian operator can be used to calculate the contrast of each pixel in the homogenized corrected image.
And (3) calculating the saturation of each pixel by using the standard deviation of three channels of r, g and b after dodging and color evening:
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that various changes may be made and equivalents may be substituted for elements thereof; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions.
Claims (10)
1. An endoscopic multifocal image fusion method characterized by comprising:
an image acquisition step, wherein images are acquired under different focal lengths respectively to obtain a plurality of original images;
a characteristic point extraction step, which is to respectively extract the characteristic points of each original image;
a characteristic matching step, namely performing characteristic matching between every two original images according to the characteristic point pairs to obtain matching point pairs between every two original images;
determining a reference image, and finding out an original image with the most matching points with other original images as the reference image;
correcting all original images to the plane of the reference image to obtain a plurality of corrected images;
a step of light and color homogenizing treatment, which is to respectively carry out light and color homogenizing treatment on the corrected image to obtain a corrected image after light and color homogenizing;
and an image fusion step, namely performing fusion processing on the corrected images after light and color uniformization to obtain a fused image.
2. The endoscopic multifocal image fusion method according to claim 1, characterized in that in the image correction step, homographic transformations of other original images than the reference image to the reference image are calculated by a method robust to mismatching, and the homographic transformations are used to correct the other original images than the reference image to the plane of the reference image.
3. The endoscopic multifocal image fusion method according to claim 1, characterized in that the dodging and color-homogenizing process step comprises:
dividing the plane of the reference image into a plurality of image blocks;
calculating the gray level histogram of each corrected image in the image block in r, g and b channels,,wherein k is the number of the corrected image, and p and q represent the row number and the column number of the image block;
calculating the sum of the gray level histograms of all corrected images in the image block at r, g and b channels respectively, and recording the sum as,;
and respectively carrying out light and color evening calculation on each corrected image according to the transformation coefficient.
5. An endoscope multifocal image fusion method according to claim 4, characterized in that the method of homogenizing the respective correction images comprises:
which represents the correction of the image or images,representing the corrected image after the dodging and the evening, (i, j) representing the pixel coordinates,representing image blocksThe coordinates of the center pixel of (a),as a weight value, the weight value,representing a center-distance pixel of a corrected image block (p, q)The distance of (a) to (b),is the size of the image block or blocks,is the number of rows of the image block corresponding to the ith row of the corrected image,and correcting the number of columns of the image block corresponding to the jth column of the image.
6. An endoscope multifocal image fusion method according to claim 5, characterized in that in the image fusion step, the corrected image after the dodging and the color evening is fused by a weighted sum method.
7. The endoscopic multifocal image fusion method according to claim 6, characterized in that the fusion processing method is: and (3) calculating the contrast of each pixel in the corrected image after the light and color are homogenized by using a Laplace operator:
and (3) calculating the saturation of each pixel by using the standard deviation of three channels of r, g and b after dodging and color evening:
9. The endoscopic multifocal image fusion method according to any of claims 1 to 8, characterized in that any of a SIFT algorithm, a HOG algorithm, and an ORB algorithm is used in the feature point extraction step.
10. The endoscope multifocal image fusion method according to any of claims 1 to 8, characterized in that in the feature matching step, feature matching is performed by using a correlation coefficient feature matching method or an Euclidean distance feature matching method, and a RANSAC method is used to filter matching results.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210320964.9A CN114418920B (en) | 2022-03-30 | 2022-03-30 | Endoscope multi-focus image fusion method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210320964.9A CN114418920B (en) | 2022-03-30 | 2022-03-30 | Endoscope multi-focus image fusion method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114418920A true CN114418920A (en) | 2022-04-29 |
CN114418920B CN114418920B (en) | 2022-06-28 |
Family
ID=81263403
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210320964.9A Active CN114418920B (en) | 2022-03-30 | 2022-03-30 | Endoscope multi-focus image fusion method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114418920B (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103226822A (en) * | 2013-05-15 | 2013-07-31 | 清华大学 | Medical image stitching method |
CN104182949A (en) * | 2014-08-18 | 2014-12-03 | 武汉大学 | Image inking and fusing method and system based on histogram feature point registration |
US20150055886A1 (en) * | 2013-08-20 | 2015-02-26 | Samsung Techwin Co., Ltd. | Image fusion system and method |
CN106504194A (en) * | 2016-11-03 | 2017-03-15 | 重庆邮电大学 | A kind of image split-joint method based on most preferably splicing plane and local feature |
CN107016647A (en) * | 2017-04-26 | 2017-08-04 | 武汉大学 | Panoramic picture color tone consistency correcting method and system |
CN109190460A (en) * | 2018-07-23 | 2019-01-11 | 南京航空航天大学 | Based on cumulative matches and etc. error rates hand shape arm vein fusion identification method |
KR102235127B1 (en) * | 2019-12-18 | 2021-04-01 | 세종대학교산학협력단 | Adaptive fusion method of multimodal images and apparatus |
CN113570536A (en) * | 2021-07-31 | 2021-10-29 | 中国人民解放军61646部队 | Panchromatic and multispectral image real-time fusion method based on CPU and GPU cooperative processing |
-
2022
- 2022-03-30 CN CN202210320964.9A patent/CN114418920B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103226822A (en) * | 2013-05-15 | 2013-07-31 | 清华大学 | Medical image stitching method |
US20150055886A1 (en) * | 2013-08-20 | 2015-02-26 | Samsung Techwin Co., Ltd. | Image fusion system and method |
CN104182949A (en) * | 2014-08-18 | 2014-12-03 | 武汉大学 | Image inking and fusing method and system based on histogram feature point registration |
CN106504194A (en) * | 2016-11-03 | 2017-03-15 | 重庆邮电大学 | A kind of image split-joint method based on most preferably splicing plane and local feature |
CN107016647A (en) * | 2017-04-26 | 2017-08-04 | 武汉大学 | Panoramic picture color tone consistency correcting method and system |
CN109190460A (en) * | 2018-07-23 | 2019-01-11 | 南京航空航天大学 | Based on cumulative matches and etc. error rates hand shape arm vein fusion identification method |
KR102235127B1 (en) * | 2019-12-18 | 2021-04-01 | 세종대학교산학협력단 | Adaptive fusion method of multimodal images and apparatus |
CN113570536A (en) * | 2021-07-31 | 2021-10-29 | 中国人民解放军61646部队 | Panchromatic and multispectral image real-time fusion method based on CPU and GPU cooperative processing |
Non-Patent Citations (3)
Title |
---|
CODRUTA O. ANCUTI ET AL: "Color Balance and Fusion for Underwater Image Enhancement", 《IEEE TRANSACTIONS ON IMAGE PROCESSING》 * |
俞家勇: "移动测量系统异构数据高精度融合方法研究", 《中国博士学位论文全文数据库(电子期刊)》 * |
柴庆朋: "面向图像拼接的几何校准和颜色校正技术研究", 《中国优秀硕士学位论文全文数据库(电子期刊)》 * |
Also Published As
Publication number | Publication date |
---|---|
CN114418920B (en) | 2022-06-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20220207728A1 (en) | Quality assessment in video endoscopy | |
JP5094036B2 (en) | Endoscope insertion direction detection device | |
US20220157047A1 (en) | Feature Point Detection | |
JP4104054B2 (en) | Image alignment apparatus and image processing apparatus | |
CN112804943B (en) | Learning completion model creation method, brightness adjustment method, and image processing apparatus | |
JP2010512173A (en) | Computer-aided analysis using video from an endoscope | |
US8666135B2 (en) | Image processing apparatus | |
CN112446880B (en) | Image processing method, electronic device and readable storage medium | |
CN110751605A (en) | Image processing method and device, electronic equipment and readable storage medium | |
US20220006981A1 (en) | Method of automatic image freezing of digestive endoscopy | |
JP4274400B2 (en) | Image registration method and apparatus | |
CN114004969A (en) | Endoscope image focal zone detection method, device, equipment and storage medium | |
Huang et al. | Deep unsupervised endoscopic image enhancement based on multi-image fusion | |
CN111161852B (en) | Endoscope image processing method, electronic equipment and endoscope system | |
CN114693682B (en) | Spine feature identification method based on image processing | |
CN114418920B (en) | Endoscope multi-focus image fusion method | |
CN112215878A (en) | X-ray image registration method based on SURF feature points | |
CN115223193B (en) | Capsule endoscope image focus identification method based on focus feature importance | |
CN116029934A (en) | Low-dose DR image and CT image denoising method | |
CN114387191B (en) | Endoscope image enhancement method and endoscope device | |
Jimenez-Martin et al. | Specular reflections removal in colposcopic images based on neural networks: Supervised training with no ground truth previous knowledge | |
AU2022230997B2 (en) | Method of image enhancement for distraction deduction in endoscopic procedures | |
Flores et al. | Identifying precursory cancer lesions using temporal texture analysis | |
KR102383495B1 (en) | Medical image data extraction method | |
CN117764847A (en) | Method and system for focusing white light image and fluorescent image of 3D endoscope |
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 |