CN103093449A - Multi-resolution fusion radial image enhancement method - Google Patents
Multi-resolution fusion radial image enhancement method Download PDFInfo
- Publication number
- CN103093449A CN103093449A CN2013100636046A CN201310063604A CN103093449A CN 103093449 A CN103093449 A CN 103093449A CN 2013100636046 A CN2013100636046 A CN 2013100636046A CN 201310063604 A CN201310063604 A CN 201310063604A CN 103093449 A CN103093449 A CN 103093449A
- Authority
- CN
- China
- Prior art keywords
- fusion
- image
- coefficient
- ray image
- different
- 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
Images
Abstract
The invention discloses a multi-resolution fusion Radial image enhancement method which includes a first step of obtaining multiple images containing different flaw features in tested workpiece Radial images, a second step of conducting multilayer decomposition on the multiple images by respectively adopting a multi-resolution method and decomposing the images to frequency coefficients in different layers, a third step of conducting fusion processing on the frequency coefficients and obtaining a multi-resolution frequency coefficient pyramid after fusion, and a fourth step of reconfiguring the multi-resolution frequency coefficient pyramid obtained after the fusion, and obtaining reconfigured images namely fusion enhancement images. The fusion Radial image enhancement method with different speeds and different accuracy is provided. Due to the fact that a multi-resolution decomposing and compositing method in the fusion enhancement method is changed, the speed and the accuracy of fusion enhancement are just changed. The multi-resolution fusion Radial image enhancement method has the advantages of being convenient to operate, high in processing efficiency, good in enhancement effect and the like.
Description
Technical field
The present invention relates to image processes and field of non destructive testing, particularly a kind of Ray Image Enhancement Method of multi-resolution Fusion.
Background technology
When carrying out the foundry goods detection with X ray, inevitably image is existed external disturbance and internal interference, unstable as due to the rescattering of X-ray light source and X-ray light source, make the unevenness that exists sensitive element sensitivity in a large amount of random noises and photoelectric conversion process in the X ray detected image, quantizing noise, the error in transmitting procedure and the human factor etc. of digitized process, all can bring to a certain degree noise to image.Therefore generally all must carry out image pre-service or figure image intensifying before defect recognition.
The effective collection signal of digital radial imaging system of most is all between 12～18 bits, and computing machine can only demonstrate 256 grades of gradation datas simultaneously, if therefore directly convert the signal into the loss that the part details must be caused in 0～255 interval.Radiation imaging system adopts two kinds of methods to strengthen image usually: hardware strengthens method and software strengthens method.Generally speaking the image hardware Enhancement Method of radiation imaging system is parameter and characteristics (such as radiographic source type, collimator size etc.) according to system feedback circuit that certain is set in system or wave filter etc., after in a single day hardware system designs and completes, the hardware of image strengthens method just to be determined, is difficult to change.What therefore generally adopt in ray image strengthens is the method that software is processed.
Present usage comparison Ray Image Enhancement Method widely has: the methods such as gray scale stretching, histogram equalization, histogram modification, self-adaptation enhancing, homomorphic filtering and small echo enhancing.Evidence, traditional image enchancing method can inevitably bring noise to cross to strengthen to the weak characteristic image of ray image and seriously reduce the problem of Disposal quality.Thereby at present the method that generally adopts of all big enterprises is for different surveyed areas, does the corresponding overall situation or local gray level linearity or power transform, to improve the resolution characteristic of ROI.
The essence of greyscale transformation method is that less gray space is expanded to larger gray space by linear relationship.Therefore, the greyscale transformation method can make the dynamic range of image strengthen, and picture contrast expansion, clear picture, feature are obvious, are the important means of figure image intensifying.The greyscale transformation method can effectively improve the detectability of workpiece, but repeatedly global change produces many ray images can for same workpiece, with the circumscription of the greyscale transformation of the image left and right in high-density matter institute corresponding grey scale, at this moment materials of low density is lumped together and can't be distinguished, and vice versa.That is to say if image is done overall greyscale transformation can't be differentiated simultaneously the defective of materials of low density and the defective of high-density matter.Therefore at present each manufacturer all provides the local enhancement function, and when namely strengthening the image of high-density matter region, the image of materials of low density is unaffected; When strengthening the image of materials of low density, the image of high-density matter is unaffected, to reach the purpose that shows simultaneously different defectives in a sub-picture.For example, suppose that certain measured workpiece not only contains high-density matter (iron, copper etc.) but also contains materials of low density (rubber etc.), when us when analyzing its ray image, if investigate its high-density matter, whether crackle etc. is arranged, can choose and comprise the high-density matter zone, then the figure image intensifying be carried out in the zone; If investigate its materials of low density whether during defectiveness, choose and comprise the materials of low density zone, then the figure image intensifying is carried out in the zone.This method can all show the various defectives of a sub-picture, and can avoid owing to image being carried out " excessively stretching " phenomenon that overall enhanced brings, can cause the violent conversion of gray scale between regional on image but do like this, artificial has brought a lot of false edges to image, very unfavorable to follow-up defect recognition, further processing is become impossible.Therefore how showing simultaneously all defectives on piece image and don't introduce extra error, is current urgent problem.
Summary of the invention
In view of this, technical matters to be solved by this invention is to provide a kind of Ray Image Enhancement Method of multi-resolution Fusion, the method is for the characteristics of current radial imaging and the deficiency of the present Enhancement Method of generally using, thought according to multiresolution analysis, proposed the scheme that adopts image interfusion method to strengthen as ray image, the result of acquisition and human vision property are more approaching.This method can show simultaneously in piece image that the defective of the different depth of field can not bring false edge to image again, for follow-up defect recognition is laid a good foundation.
The object of the present invention is achieved like this:
The Ray Image Enhancement Method of a kind of multi-resolution Fusion provided by the invention comprises the following steps:
S1: according to testing goal or Flaw display needs, obtain several ray images that comprise the different defect characteristics of tested workpiece;
S2: adopt respectively multiresolution method to carry out multilayer to each ray image that obtains in step S1 and decompose, ray image is decomposed into coefficient of frequency on different layers;
S3: the coefficient of frequency on each decomposition layer is carried out respectively fusion treatment, the multiresolution coefficient of frequency pyramid after being merged;
S4: be reconstructed merging rear gained multiresolution coefficient of frequency pyramid, obtain reconstructed image and be fusion enhancing image.
Further, in described step S1 is according to testing goal or Flaw display needs, and the ray image of measured workpiece is carried out the segmentation stretch processing or gives the different window width of ray image/window position to obtain to comprise several ray images of different defect characteristics.
Further, in described step S3 each decomposition layer coefficient of frequency is carried out fusion treatment respectively the time, different frequency component on each decomposition layer adopts different fusion operators to carry out fusion treatment, thus the multiresolution coefficient of frequency pyramid after being merged.
Further, the different fusion operator that the different frequency component on described each decomposition layer adopts, specifically adopt following methods:
S31: after low frequency coefficient is adopted first filtering, average method merges;
S32: adopt the neighborhood gradient maximization method with consistency check to merge to high frequency coefficient.
Further, the multiresolution method that described ray image adopts carries out multilayer and decomposes multiresolution and fusion treatment employing wavelet transformation enhancing method, and concrete steps are as follows:
S31: each primary ray image is carried out respectively wavelet transformation, and the small echo tower of setting up ray image decomposes;
S32: each decomposition layer is carried out respectively fusion treatment, and the different frequency component on each decomposition layer can adopt different fusion operators to carry out fusion treatment, the wavelet pyramid after finally being merged; After namely low frequency coefficient being adopted first filtering, average method merges; Adopt the neighborhood gradient maximization method with consistency check to merge to high frequency coefficient.
S33: carry out wavelet reconstruction to merging rear gained wavelet pyramid, resulting reconstructed image is fused images.
Further, the multiresolution method that described ray image adopts carries out multilayer and decomposes multiresolution and fusion treatment employing finite ridgelet transform method, and concrete steps are as follows:
S34: to the primary ray Image Segmentation Using, the primary ray image is divided into overlapped sub-image, obtains the subimage of some correspondence positions;
S35: each number of sub images of cutting apart rear relevant position is carried out multistage finite ridgelet transform, obtain the finite ridgelet transform matrix of coefficients in each sub-block different frequency territory;
S36: carry out coefficient according to the finite ridgelet transform matrix of coefficients and merge, select different fusion rules in different frequency fields, namely average method after the first filtering of low frequency coefficient employing is merged; Adopt the neighborhood gradient maximization method with consistency check to merge to high frequency coefficient;
S37: the finite ridgelet transform matrix of coefficients after merging according to each sub-block coefficient carries out Image Reconstruction, the rule of taking when cutting apart makes up each sub-block fused images, the pixel of lap adopts average weighted method to obtain, and resulting reconstructed image is fused images.
The invention has the advantages that: the ray image that the invention provides friction speed, different accuracy merges Enhancement Method.Change the Multiresolution Decomposition, the synthetic method that merge in Enhancement Method and can change speed and the precision that merges enhancing.Adopt wavelet transformation to realize that as the multiresolution function ray image merges enhancing fast in the present invention; Adopt finite ridgelet transform to realize that as the multiresolution function high-precision ray image merges Enhancement Method.That the present invention has is easy to operate, treatment effeciency is high, strengthen the advantages such as effective.
Description of drawings
In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention is described in further detail below in conjunction with accompanying drawing, wherein:
Fig. 1 is the process flow diagram that image co-registration strengthens;
Fig. 2 is the image co-registration process flow diagram based on wavelet transformation.
Embodiment
Below with reference to accompanying drawing, the preferred embodiments of the present invention are described in detail; Should be appreciated that preferred embodiment only for the present invention is described, rather than in order to limit protection scope of the present invention.
Fig. 1 is the process flow diagram that image co-registration strengthens, and Fig. 2 is the image co-registration process flow diagram based on wavelet transformation, as shown in the figure: and the Ray Image Enhancement Method of a kind of multi-resolution Fusion provided by the invention,
Because ray image demonstrates different features at different signal segments, if these features are synthesized together can greatly strengthen understanding to ray image.Therefore can extract these features, then use certain technology that these features are combined, thus convenient follow-up processing.And image co-registration is that (or different time) image or image sequence information about certain concrete scene of obtaining is in addition comprehensive at one time with two or more sensors, redundancy or complementary multi-source data carry out calculation process according to certain rule in the space or on the time those, generate the explanation of new relevant this scene, and this explanation is than more accurate, the abundanter information of any single data.
Comprise the following steps:
S1: according to testing goal or Flaw display needs, obtain several ray images that comprise the different defect characteristics of workpiece; In described step S1 is according to testing goal or Flaw display needs, and ray image is carried out the segmentation stretch processing or gives the different window width of ray image/window position to obtain to comprise several ray images of the different defect characteristics of tested workpiece.
For the characteristics of radial imaging, according to the type of measured workpiece and the grey level histogram of its detection signal distribution, its ray image signal is carried out repeatedly greyscale transformation in the embodiment of the present invention.Ubiquity gray scale inconsistent (being that workpiece of the same race is in signal segment difference corresponding to different detection times) unstable by roentgen dose X and that detector rises and falls and causes in ray detection system, but the histogram shape of its signal distributions remains unchanged substantially, just exist the axial displacement of part, therefore stretching by the gray scale by the histogram distribution shape to reach the purpose of correcting image gray consistency.Used gray-scale transformation method in invention, the greyscale transformation formula is as follows:
, wherein, f (x, y) expression initial ray image, ray image after g (x, y) expression stretches, L1 represents the lower limit that stretches, L2 represents the upper limit that stretches;
S2: adopt multiresolution method to carry out multilayer to each ray image that obtains in step S1 and decompose, ray image is decomposed into coefficient of frequency on different layers;
S3: the coefficient of frequency on each decomposition layer is carried out respectively fusion treatment, the multiresolution coefficient of frequency pyramid after being merged; Described when each decomposition layer coefficient of frequency is carried out fusion treatment respectively, the different frequency component on each decomposition layer adopts different fusion operators to carry out fusion treatment, thus the multiresolution coefficient of frequency pyramid after being merged.The different fusion operator that different frequency component on described each decomposition layer adopts, specifically adopt following methods:
S31: after low frequency coefficient is adopted first filtering, average method merges;
S32: adopt the neighborhood gradient maximization method with consistency check to merge to high frequency coefficient.
S4: be reconstructed merging rear gained multiresolution coefficient of frequency pyramid, obtain reconstructed image and be fused images.
Two kinds of multi-Resolution Image Fusion Enhancement Method that the present embodiment provides: Wavelet Transform Fusion Enhancement Method and finite ridgelet transform Enhancement Method.These two kinds of methods are adapted to respectively different occasions, and the wavelet transformation Enhancement Method is adapted to require at a high speed occasion; Wavelet Transform Fusion strengthens can realize that quick ray image strengthens, and finite ridgelet transform is adapted to the high-precision requirement occasion; The finite ridgelet transform Enhancement Method can realize the high-quality image intensifying.
In the high-speed occasion that requires, the multiresolution method that the present embodiment provides carries out the multilayer decomposition and fusion treatment can adopt wavelet transformation to strengthen method, and concrete steps are as follows:
S31: each primary ray image is carried out respectively wavelet transformation, and the small echo tower of setting up ray image decomposes;
S32: each decomposition layer is carried out respectively fusion treatment, and the different frequency component on each decomposition layer can adopt different fusion operators to carry out fusion treatment, the wavelet pyramid after finally being merged; After namely low frequency coefficient being adopted first filtering, average method merges; Adopt the neighborhood gradient maximization method with consistency check to merge to high frequency coefficient.
S33: carry out wavelet reconstruction to merging rear gained wavelet pyramid, resulting reconstructed image is fused images.
Wavelet transformation strengthens and merges Enhancement Method realization flow such as Fig. 2, take two width image fusions as example.If A, B are two width original images, F is the image after merging.If two dimensional image is carried out the wavelet decomposition of N layer, (3N+1) individual different frequency bands is arranged the most at last, wherein comprise 3N high frequency subimage and 1 low frequency subgraph picture.The basic step of its fusion treatment is as follows:
Each original image is carried out respectively wavelet transformation, and the small echo tower of setting up image decomposes;
Each decomposition layer is carried out respectively fusion treatment.Different frequency component on each decomposition layer can adopt different fusion operators to carry out fusion treatment, the wavelet pyramid after finally being merged;
Carry out wavelet reconstruction to merging rear gained wavelet pyramid, resulting reconstructed image is fused images.
Used following fusion rule in the present invention:
The fusion rule of low frequency part: for low frequency Global Information component, be called again approximation coefficient, it has preserved the main profile of image.Owing to being same object, therefore two width images will be much smaller than the difference between the high-frequency information component through the difference between its low-frequency information after wavelet decomposition.From same source figure, that is to say between these figure it has been accuracy registration due to several greyscale transformation figure.Therefore, the present invention adopts the simple method of average to merge to low frequency part.
The fusion rule of HFS: what high-frequency sub-band characterized is the detailed information of image, the coefficient of high-frequency sub-band is in null value left and right fluctuation, the coefficient that absolute value is larger represents that this place's grey scale change Shaoxing opera is strong, namely comprises the important information of image, as edge, lines and the regional border of image.And these details of paying close attention to the most in ray detection just, therefore the present invention adopts HFS and processes with the neighborhood gradient maximization method of consistency check, and concrete grammar is as follows: at first with neighborhood template window sequential action on the high frequency subimage after decomposition; Then calculate respectively the gradient of corresponding localized mass; Utilize at last the partial gradient criterion to carry out corresponding high frequency subimage and merge (large gradient principle is chosen in employing); Repeat above process, until the whole fusions of the high frequency subimage of three directions of each decomposition layer are complete;
Do consistency check to merging rear high frequency subimage.Its method of adjustment is: after supposing to merge certain position pixel from source images A, if in its neighborhood from the number of pixels of B image greater than the number of pixels from the A image, this position pixel is adjusted into the pixel that comes from B.
In the occasion of high schedule requirement, the multiresolution method that the present embodiment provides carries out the multilayer decomposition and fusion treatment can adopt the finite ridgelet transform method, and concrete steps are as follows:
S34: to the primary ray Image Segmentation Using, the primary ray image is divided into overlapped sub-image, obtains the subimage of some correspondence positions;
Take two width image fusions as example.If A, B are two width original images, F is the image after merging.The basic step of finite ridgelet transform pixel-level image fusion is as follows:
Image segmentation.Finite ridgelet transform can represent effectively that in fact line singularity refers to straight line, and is more the singularity of curve in general pattern.For this reason, can first carry out piecemeal to image, the approximate singularity that is converted into straight line of the singularity of little regional inner curve.Respectively A, B source images are cut apart, obtained the subimage of some correspondence positions.If when cutting apart, in the time of image can not being divided into the sub-block of integer size just, can first carry out cutting apart again after continuation to image.But having a problem here is exactly the block edge effect of the reconstructed image that brought by image segmentation.For digital picture, the distortion that brings because of piecemeal in the reconstructed image need to be divided into image overlapped sub-block.For example, (when the image segmentation of establishing N＜M) became a series of b * b sub-blocks, a kind of typical way was exactly at every row and often lists and mark off respectively 2N/b sub-block, and the lap size between adjacent two sub-blocks is b * b/2 with M * N.Each point of cutting apart in rear image belongs to 4 sub-blocks (except edge pixel) simultaneously, and namely the redundance of each pixel is 4, and this is to exchange the sharpness of reconstructed image for redundance.
S35: each number of sub images of cutting apart rear relevant position is carried out multistage finite ridgelet transform, obtain the finite ridgelet transform matrix of coefficients in each sub-block different frequency territory;
S36: carry out coefficient according to the finite ridgelet transform matrix of coefficients and merge, select different fusion rules in different frequency fields, in the finite ridgelet transform coefficient of piece image, the larger HFS coefficient of absolute value is corresponding to some comparatively significant features such as edges; Low frequency coefficient has determined the general picture of image, namely average method after the first filtering of low frequency coefficient employing is merged; Adopt the neighborhood gradient maximization method with consistency check to merge to high frequency coefficient;
S37: the finite ridgelet transform matrix of coefficients after merging according to each sub-block coefficient carries out Image Reconstruction, the rule of taking when cutting apart makes up each sub-block fused images, the pixel of lap adopts average weighted method to obtain, and resulting reconstructed image is fused images.If carried out boundary extension in the image segmentation process, need corresponding adjustment can obtain final fused images.
The above is only the preferred embodiments of the present invention, is not limited to the present invention, and obviously, those skilled in the art can carry out various changes and modification and not break away from the spirit and scope of the present invention the present invention.Like this, if within of the present invention these are revised and modification belongs to the scope of claim of the present invention and equivalent technologies thereof, the present invention also is intended to comprise these changes and modification interior.
Claims (5)
1. the Ray Image Enhancement Method of a multi-resolution Fusion is characterized in that: comprise the following steps:
S1: according to testing goal or Flaw display needs, the ray image of measured workpiece is carried out the segmentation stretch processing or gives the different window width of ray image/window position to obtain to comprise several ray images of different defect characteristics;
S2: adopt multiresolution method to carry out multilayer to each ray image that obtains in step S1 and decompose, several ray images are decomposed into respectively coefficient of frequency on different layers;
S3: the coefficient of frequency on each decomposition layer is carried out respectively fusion treatment, the multiresolution coefficient of frequency pyramid after being merged;
S4: be reconstructed merging rear gained multiresolution coefficient of frequency pyramid, obtain reconstructed image and be fusion enhancing image.
2. the Ray Image Enhancement Method of multi-resolution Fusion according to claim 1, it is characterized in that: in described step S3 each decomposition layer coefficient of frequency is carried out fusion treatment respectively the time, different frequency component on each decomposition layer adopts different fusion operators to carry out fusion treatment, thus the multiresolution coefficient of frequency pyramid after being merged.
3. the Ray Image Enhancement Method of multi-resolution Fusion according to claim 1 is characterized in that: the different fusion operator that the different frequency component on described each decomposition layer adopts, specifically adopt following methods:
S31: after low frequency coefficient is adopted first filtering, average method merges;
S32: adopt the neighborhood gradient maximization method with consistency check to merge to high frequency coefficient.
4. the Ray Image Enhancement Method of multi-resolution Fusion according to claim 1 is characterized in that: the multiresolution method that described ray image adopts carries out that multilayer is decomposed multiresolution and fusion treatment adopts wavelet transformation to strengthen method, and concrete steps are as follows:
S31: each primary ray image is carried out respectively wavelet transformation, and the small echo tower of setting up ray image decomposes;
S32: each decomposition layer is carried out respectively fusion treatment, and the different frequency component on each decomposition layer can adopt different fusion operators to carry out fusion treatment, the wavelet pyramid after finally being merged; After namely low frequency coefficient being adopted first filtering, average method merges; Adopt the neighborhood gradient maximization method with consistency check to merge to high frequency coefficient;
S33: carry out wavelet reconstruction to merging rear gained wavelet pyramid, resulting reconstructed image is fused images.
5. the Ray Image Enhancement Method of multi-resolution Fusion according to claim 1, it is characterized in that: the multiresolution method that described ray image adopts carries out multilayer and decomposes multiresolution and fusion treatment employing finite ridgelet transform method, and concrete steps are as follows:
S34: to the primary ray Image Segmentation Using, the primary ray image is divided into overlapped sub-image, obtains the subimage of some correspondence positions;
S35: each number of sub images of cutting apart rear relevant position is carried out multistage finite ridgelet transform, obtain the finite ridgelet transform matrix of coefficients in each sub-block different frequency territory;
S36: carry out coefficient according to the finite ridgelet transform matrix of coefficients and merge, select different fusion rules in different frequency fields, namely average method after the first filtering of low frequency coefficient employing is merged; Adopt the neighborhood gradient maximization method with consistency check to merge to high frequency coefficient;
S37: the finite ridgelet transform matrix of coefficients after merging according to each sub-block coefficient carries out Image Reconstruction, the rule of taking when cutting apart makes up each sub-block fused images, the pixel of lap adopts average weighted method to obtain, and resulting reconstructed image is fused images.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013100636046A CN103093449A (en) | 2013-02-28 | 2013-02-28 | Multi-resolution fusion radial image enhancement method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013100636046A CN103093449A (en) | 2013-02-28 | 2013-02-28 | Multi-resolution fusion radial image enhancement method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103093449A true CN103093449A (en) | 2013-05-08 |
Family
ID=48205983
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2013100636046A Pending CN103093449A (en) | 2013-02-28 | 2013-02-28 | Multi-resolution fusion radial image enhancement method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103093449A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104717465A (en) * | 2014-12-23 | 2015-06-17 | 北京亿羽舜海科技有限公司 | Wide-angle video monitoring device and method for multiple IP cameras |
| CN105894486A (en) * | 2016-06-29 | 2016-08-24 | 深圳市优象计算技术有限公司 | Mobile phone night-shot method based on imu information |
| CN107622482A (en) * | 2017-09-13 | 2018-01-23 | 电子科技大学 | A kind of image interfusion method based on leukorrhea micro-imaging |
| CN110675354A (en) * | 2019-09-11 | 2020-01-10 | 北京大学 | Image processing method, system and storage medium for developmental biology |
| CN111353933A (en) * | 2018-12-20 | 2020-06-30 | 重庆金山医疗器械有限公司 | Image splicing and fusing method and system |
| CN111445492A (en) * | 2020-03-26 | 2020-07-24 | 北京易康医疗科技有限公司 | Three-dimensional fusion calibration method for magnetic resonance image and radiotherapy positioning image |
| CN112070664A (en) * | 2020-07-31 | 2020-12-11 | 华为技术有限公司 | Image processing method and device |
| CN112288662A (en) * | 2019-08-12 | 2021-01-29 | 中北大学 | Hub X-ray image enhancement method and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5452367A (en) * | 1993-11-29 | 1995-09-19 | Arch Development Corporation | Automated method and system for the segmentation of medical images |
| US5740267A (en) * | 1992-05-29 | 1998-04-14 | Echerer; Scott J. | Radiographic image enhancement comparison and storage requirement reduction system |
| CN101504766A (en) * | 2009-03-25 | 2009-08-12 | 湖南大学 | Image amalgamation method based on mixed multi-resolution decomposition |
-
2013
- 2013-02-28 CN CN2013100636046A patent/CN103093449A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5740267A (en) * | 1992-05-29 | 1998-04-14 | Echerer; Scott J. | Radiographic image enhancement comparison and storage requirement reduction system |
| US5452367A (en) * | 1993-11-29 | 1995-09-19 | Arch Development Corporation | Automated method and system for the segmentation of medical images |
| CN101504766A (en) * | 2009-03-25 | 2009-08-12 | 湖南大学 | Image amalgamation method based on mixed multi-resolution decomposition |
Non-Patent Citations (4)
| Title |
|---|
| 冈萨雷斯: "《数字图像处理 第二版》", 30 September 2007 * |
| 沈宽,文玉梅,蔡玉芬: "《基于脊波融合的射线图像增强算法》", 《计算机工程与应用》 * |
| 沈宽,蔡玉芬: "《基于小波融合的射线图像增强算法》", 《计算机应用研究》 * |
| 沈宽,蔡玉芬: "《基于小波融合的射线图像增强算法》", 《计算机应用研究》, vol. 26, no. 8, 31 August 2009 (2009-08-31), pages 3137 - 3139 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104717465A (en) * | 2014-12-23 | 2015-06-17 | 北京亿羽舜海科技有限公司 | Wide-angle video monitoring device and method for multiple IP cameras |
| CN104717465B (en) * | 2014-12-23 | 2023-10-03 | 北京亿羽舜海科技有限公司 | Wide-angle video monitoring equipment and video monitoring method for multiple IP cameras |
| CN105894486A (en) * | 2016-06-29 | 2016-08-24 | 深圳市优象计算技术有限公司 | Mobile phone night-shot method based on imu information |
| CN105894486B (en) * | 2016-06-29 | 2018-08-03 | 深圳市优象计算技术有限公司 | A kind of mobile phone night shooting method based on imu information |
| CN107622482A (en) * | 2017-09-13 | 2018-01-23 | 电子科技大学 | A kind of image interfusion method based on leukorrhea micro-imaging |
| CN111353933A (en) * | 2018-12-20 | 2020-06-30 | 重庆金山医疗器械有限公司 | Image splicing and fusing method and system |
| CN112288662A (en) * | 2019-08-12 | 2021-01-29 | 中北大学 | Hub X-ray image enhancement method and system |
| CN110675354B (en) * | 2019-09-11 | 2022-03-22 | 北京大学 | Image processing method, system and storage medium for developmental biology |
| CN110675354A (en) * | 2019-09-11 | 2020-01-10 | 北京大学 | Image processing method, system and storage medium for developmental biology |
| CN111445492A (en) * | 2020-03-26 | 2020-07-24 | 北京易康医疗科技有限公司 | Three-dimensional fusion calibration method for magnetic resonance image and radiotherapy positioning image |
| WO2022022288A1 (en) * | 2020-07-31 | 2022-02-03 | 华为技术有限公司 | Image processing method and apparatus |
| CN112070664A (en) * | 2020-07-31 | 2020-12-11 | 华为技术有限公司 | Image processing method and device |
| CN112070664B (en) * | 2020-07-31 | 2023-11-03 | 华为技术有限公司 | Image processing method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103093449A (en) | Multi-resolution fusion radial image enhancement method | |
| CN101201937B (en) | Digital image enhancement method and device based on wavelet reconstruction and decompose | |
| Xiao et al. | Change detection of built-up land: A framework of combining pixel-based detection and object-based recognition | |
| CN109242888B (en) | Infrared and visible light image fusion method combining image significance and non-subsampled contourlet transformation | |
| CN102789640B (en) | A kind of method of visible ray full-colour image and infrared remote sensing image being carried out merge | |
| CN107301661A (en) | High-resolution remote sensing image method for registering based on edge point feature | |
| CN105261006B (en) | Medical image segmentation algorithm based on Fourier transformation | |
| CN103985110B (en) | Methods for positioning and displaying artifact or intensity inhomogeneity in digital image | |
| CN103049898A (en) | Method for fusing multispectral and full-color images with light cloud | |
| CN103996168A (en) | X-ray safety inspection image enhancing method based on region self-adaptive processing | |
| CN104700411A (en) | Sparse reconstruction-based dual-time phase remote-sensing image change detecting method | |
| Gong et al. | Automatic subway tunnel crack detection system based on line scan camera | |
| CN107103595A (en) | Method, device, storage medium and the equipment of detection image change | |
| CN106651792A (en) | Method and device for removing strip noise of satellite image | |
| CN107194896A (en) | A kind of background suppression method and system based on neighbour structure | |
| Joshi et al. | Multi-focus image fusion using non-local mean filtering and stationary wavelet transform | |
| Hu et al. | An efficient fusion algorithm based on hybrid multiscale decomposition for infrared-visible and multi-type images | |
| CN105608674B (en) | A kind of image enchancing method based on image registration, interpolation and denoising | |
| CN108230365A (en) | SAR image change detection based on multi-source differential image content mergence | |
| Tamilselvan et al. | Survey and analysis of various image fusion techniques for clinical CT and MRI images | |
| CN103455986B (en) | Random noise point detecting method based on fractional order differential gradient | |
| Li et al. | A study of crack detection algorithm | |
| CN115880288B (en) | Detection method, system and computer equipment for electronic element welding | |
| Yagoub et al. | X-ray image denoising for cargo dual energy inspection system | |
| Kim et al. | Learning structure for concrete crack detection using robust super-resolution with generative adversarial network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20130508 |