CN107621655A - Three-dimensional data tomography Enhancement Method based on DoS filtering - Google Patents
Three-dimensional data tomography Enhancement Method based on DoS filtering Download PDFInfo
- Publication number
- CN107621655A CN107621655A CN201710757287.6A CN201710757287A CN107621655A CN 107621655 A CN107621655 A CN 107621655A CN 201710757287 A CN201710757287 A CN 201710757287A CN 107621655 A CN107621655 A CN 107621655A
- Authority
- CN
- China
- Prior art keywords
- plane
- dimensional
- parallel
- pixel
- window
- 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
Landscapes
- Image Processing (AREA)
Abstract
The present invention discloses a kind of three-dimensional data tomography Enhancement Method based on DoS filtering, and field is explained applied to earthquake fault, detects the planar structure of different directions by introducing the Plane of rotation core of three dimensions, generates monoplane three-dimensional operator；It is on the basis of each plane, different according to the direction of plane, a parallel plane is introduced respectively in the upper and lower or left and right of plane, thus defines the nonlinear difference in three-dimensional；The method of mathematical morphology has been used to carry out subsequent treatment.
Description
Technical field
The invention belongs to earthquake fault data processing field, more particularly to a kind of earthquake fault interpretation technique.
Background technology
The origin cause of formation of tomography, refer to rock by after the effect of power, if stress exceedes the maximum upper limit that rock can bear,
It will result in rock and fracture phenomena occur, the geologic structure being consequently formed is to be broken.Fracture be most common geological phenomenon it
One.Tomography is exactly the fault structure that rock stratum produces displacement along splitting plane.Accurate portray of tomography is run through in oil-gas field development
Whole processes in.Mature fault is very wide, is one of the crucial geologic structure in stratum.Tomography comes in every shape, each not phase of scale
Together.In the migration process of oil gas, tomography may play two kinds of different effects, and one kind is channels result, and another kind is envelope
Stifled effect.Nowadays, the development trend of geological prospecting research is to deeper level, as long as being related to oil-gas reservoir research, just from not
The quantitative study to migration path system is opened, and wherein, tomography has larger difficulties in exploration as one of migration path system.Oil gas provides
The Exploratory situation in source, an urgent demand people develop tomography extraction and enhancing technology.
Because the influences of the abnormal geographical environments such as the absorption of seismic image noise, stratum to seismic wave, fracture, earthquake are cutd open
Phenomena such as fracture of adjacent seismic channel set is showed in face layer position and hole so that in the weaker place of seismic signal, tomography leads to
This due direction can often be lost.In the seismic channel that search closes on, if running into the geologic bodies such as fracture, tomography can be caused to mix
Disorderly.In tomography strengthens and identifies, fault information has necessarily uncertain and dispersed, and this can be caused to fault recognizing
The influence of very severe.So although fault recognizing technology achieves unprecedented progress, above-mentioned problem still urgently solves
Certainly.Therefore fault recognizing becomes the link held the balance in geologic information research, while be also very difficult interior
Hold.The identification and enhancing of tomography are directed to entering year, has become the hot issue of geologician and scientist person's research.
Stick wave filters are the wave filters of wire characteristic and non-linear shape characteristic in a kind of differentiation image.Wire in image
Feature is critically important under many circumstances, because these features may include important information.In seismic prospecting, tomography
Feature is usually expressed as linear structure on two dimensional slice.Therefore, the linear structure in earthquake attribute volume is protruded to disconnected
Layer is extracted and strengthened most important.Start with below from wire detection, the stick wave filters on basis are discussed in detail：
1 rotation stick kernel mappings
The subject matter of traditional tomography enhancing is that tiny tomography is generally difficult to identify.Usually used conventional smooth
Wave filter, often without matching linear constraints, cause the crack of weakness very likely as noise filters out together, therefore tradition
Smoothing filter is unsuitable.A practical method is introduced below, and this method there may be by the fitting of some straightways
Fractuer direction, so as to detect the presence across the two-dimentional crack of each section, these be used for be fitted straightways claim
For sticks.In order to adapt to the change in direction, rectangular neighborhood is decomposed into width first and fixed for single pixel, length
Stick, wherein each stick represents potential fractuer direction, and introduce rotation kernel mapping.Generally, by through center
Rotational line carries out spatial sampling and discretization, can decompose neighborhood into strips.
The definition of 2 nonlinear differences
On the basis of the stick wave filters of rotation kernel mapping, the difference based on one group three stick templates arranged side by side
Stick wave filters；Low intensive band is usually expressed as in view of tiny tomography, it is parallel and somewhat spaced using three
Sticks while fracture detection profile and its neighborhood.This be also visually human viewer be used for inferring existing for crack it is former
Reason, especially low-intensity and it is noisy in the case of.
Curvilinear structures can be described using thin (thin), thick (thick) and width (broad) line, its width is from single picture
Element increases to more than 2S (S represents the interval between adjacent stick).The upper left corner is illustrated one with LSAs left stick,
With RSThe situation on a rimala region that three sticks kernels as right stick are covered in faultage image.
Typically, the method for two-dimentional DoS filtering can effectively detect the linear structure in two-dimension earthquake image, so
And fault structure is usually expressed as one piece of continuous plane in practice, therefore, if only carrying out wire to earthquake attribute volume
Structure detection, have significant limitation.
The content of the invention
In order to solve the above technical problems, present applicant proposes it is a kind of based on DoS filtering three-dimensional data tomography Enhancement Method,
Take the filtering algorithm of improvement to handle coherent data volumes, strengthen fault information useful in 3D data volume, remove point-like sheet
The influence of noise and remaining non-faulting response.
The technical scheme that the application uses for：Based on the three-dimensional data tomography Enhancement Method of DoS filtering, including：
S1, input filter the 3D data volume of noise reduction by structure directing；
S2, the Plane of rotation core for introducing three dimensions, generate all monoplane three-dimensional operators corresponding to the 3D data volume；
S3, a parallel plane is respectively introduced in current plane both sides, and the three-dimensional calculation of current plane generated according to step S2
Son, generate the three dimensional anisotropic operator of difference corresponding to current plane；
S4, following processing is done to each pixel in step S1 3D data volume：Choose a untreated pixel
Point, extract the three dimensional window centered on the pixel, by the obtained all difference of three dimensional window and step S3 it is three-dimensional respectively to
Different in nature operator carries out convolution algorithm；
S5, step S4 obtained into maximum convolution value in all convolution values as optimal response value, the optimal response value pair
The direction answered is the position tomography maximum probability direction.
Further, step S2 is specially：
S21, by cube length of window L, obtain 3L2The plane of -6L+4 different directions；
S22, spherical coordinate system is established to cube window, travel through each pixel on sphere and incline corresponding in spherical coordinate system
Angle and azimuth, obtain the normal vector direction of all planes；
S23, each monoplane three-dimensional operator obtained according to the normal vector direction of each plane.
Further, step S23 also includes：Planar process vector direction is：(x⊥,y⊥,z⊥), determined according to following formula corresponding
Plane be：
x⊥(x-m)+y⊥(y-m)+z⊥(z-m)=0
Wherein, x represents x-axis variable, and y represents y-axis variable, and z represents z-axis variable, and (m, m, m) is represented in cube window
Heart point.
Further, step S3 is specially：
S31, in the both sides of current plane a parallel plane is respectively introduced, be designated as the first parallel plane and second parallel respectively
Plane；
S32, calculate two nonlinear differences perpendicular to current plane；
Wherein,The first plane-parallel nonlinear difference is represented,Represent that second is plane-parallel non-linear
Difference, θ represent pixel corresponding inclination angle in spherical coordinate system,Pixel corresponding azimuth in spherical coordinate system is represented,
pMRepresent current plane, pLAnd pRRepresent positioned at current plane both sides and parallel to two planes of current plane；
S33, introducing monoplane standard deviation, two nonlinear differences perpendicular to mid-plane obtained according to step S32,
Two dimensional strengths are calculated to measure；
High response is obtained by adjustment parameter κ planar structures；
Wherein,Represent corresponding monoplane standard deviation.
Beneficial effects of the present invention：The three-dimensional data tomography Enhancement Method based on DoS filtering of the present invention, introduces three-dimensional space
Between Plane of rotation core detect the planar structure of different directions, generate monoplane three-dimensional operator；On the basis of each plane
On, it is according to the direction of plane different, a parallel plane is introduced respectively in the upper and lower or left and right of plane, thus defines three-dimensional
On nonlinear difference；The method of mathematical morphology has been used to carry out subsequent treatment；With advantages below：
(1) anisotropic filter operator is designed according to the actual form of tomography, can more effectively remove unrelated make an uproar
Sound, reach more satisfactory effect；
(2) it is used for inferring that principle existing for crack introduces difference according to visually human viewer, in conjunction with morphology
It operation, can not only effectively retain obvious fault structure, and the tiny tomography in background can also be strengthened.
Brief description of the drawings
Fig. 1 is the protocol procedures figure of the application.
Fig. 2 is spherical coordinates schematic diagram.
Embodiment
For ease of skilled artisan understands that the technology contents of the present invention, enter one to present invention below in conjunction with the accompanying drawings
Step explaination.
It is the protocol procedures figure of the application as shown in Figure 1, the technical scheme of the application is：Three dimensions based on DoS filtering
According to tomography Enhancement Method, including：
S1, input filter the 3D data volume of noise reduction by structure directing；
S2, the Plane of rotation core for introducing three dimensions detect the planar structure of different directions, and generate three dimension
According to all monoplane three-dimensional operators corresponding to body；
S3, a parallel plane is respectively introduced in current plane both sides, and the three-dimensional calculation of current plane generated according to step S2
Son, generate the three dimensional anisotropic operator of difference corresponding to current plane；
S4, following processing is done to each pixel in step S1 3D data volume：Choose a untreated pixel
Point, extract the three dimensional window centered on the pixel, by the obtained all difference of three dimensional window and step S3 it is three-dimensional respectively to
Different in nature operator carries out convolution algorithm；
S5, step S4 obtained into maximum convolution value in all convolution values as optimal response value, the optimal response value pair
The direction answered is the position tomography maximum probability direction.
Step S2 is specially：
S21, by cube length of window L, obtain 3L2The plane of -6L+4 different directions；By taking length of window 5 as an example,
For the DoS of three-dimensional, a pair of Central Symmetry pictures in corresponding one 5 × 5 × 5 cube window surface in each possible direction
Element, the direction using each direction as the normal vector of plane, and then obtain in-plane all in cube window；In window
In the case that mouth length is 5, one shares 49 kinds of possible normal vector directions.
S22, spherical coordinate system is established to cube window, travel through each pixel on sphere and incline corresponding in spherical coordinate system
Angle and azimuth, obtain the normal vector direction of all planes；
By establishing spherical coordinate system to cube window, each pixel corresponding inclination angle in spherical coordinate system on surface is considered
And azimuth, obtain the normal vector direction of all possible plane by traveling through the method for dip and azimuth.
Spherical coordinates schematic diagram such as Fig. 2, any vector have θ,Tri- attributes of r, wherein θ represent vector and z-axis positive direction
Angle be inclination angle,Expression vector is azimuth in the projection in xoy faces and the angle of x-axis positive direction, and r represents the length of vector
Degree, any vector can all be represented with these three attributes.
Below by taking L=5 as an example, 5 × 5 × 5 cube window is taken, spherical coordinate system is established by origin of window center, takes
Vector length permanent is 1, by travel through inclination angle and azimuthal mode travel through all different directions in three-dimensional operator to
Amount, the direction for obtaining all possible normal vector in operator is (x⊥,y⊥,z⊥)
S23, each monoplane three-dimensional operator obtained according to the normal vector direction of each plane.
Step S23 is specially：Planar process vector direction is：(x⊥,y⊥,z⊥), plane is according to corresponding to determining following formula：
x⊥(x-m)+y⊥(y-m)+z⊥(z-m)=0 (1)
Wherein, x represents x-axis variable, and y represents y-axis variable, and z represents z-axis variable, and (m, m, m) is represented in cube window
Heart point, and m=(L+1)/2.
Step S3 is specially：Generally, the pixel value put on tomography can be stronger than background area, therefore traditional way is to find
Operator with maximum mean intensity.However, but tend to close to the background pixel of bright structures, this standard for those
Selection is approximately perpendicular to the operator on border.In order to avoid this shortcoming, similar to the three sticks in two dimension, the application
3 parallel planes are introduced in three dimensions, i.e., on the basis of each plane during step S2 determines operator, according to the direction of plane
Difference, a parallel plane p is introduced respectively in the upper and lower or left and right of planeLAnd pR, and it is thus non-linear in three-dimensional to define
Difference.Specifically include it is following step by step：
S31, the monoplane determined according to step S2, a parallel plane p is respectively introduced in the both sides of current planeLAnd pR；
Pass through pMSpace move and obtain pLAnd pRWith simplified operation, it is necessary to consider plane pMRelative position in a coordinate system, can
To be converted into the inclination angle of plane normal vector and azimuth to consider, between S is represented in step S2 between two neighboring plane in figure
Every.
Consider to carry out different approximations for the inclination angle of different normal vectors and azimuth, when the direction almost horizontal of plane
When, such as formula (2), it is only necessary to which the coordinate for changing its third dimension can be obtained by two planes in parallel；And when the direction of plane
During near vertical, if formula (3) and (4) are, it is necessary to change the coordinate of its preceding bidimensional.
S32, calculate two nonlinear differences perpendicular to current plane；
Wherein,The first plane-parallel nonlinear difference is represented,Represent that second is plane-parallel non-linear
Difference, θ represent pixel corresponding inclination angle in spherical coordinate system,Pixel corresponding azimuth in spherical coordinate system is represented,
pMRepresent current plane, pLAnd pRRepresent positioned at current plane both sides and parallel to two planes of current plane；
S33, monoplane standard deviation is introduced,
Wherein, E represents expected value operator, Ix,y,zRepresent the intensity along (x, y, z) place pixel of mid-plane.
Two nonlinear differences perpendicular to mid-plane obtained according to step S32, two dimensional strengths are calculated
Measure；
Wherein,Represent corresponding monoplane standard deviation, κ be sensitivity of the adjustment to axial strength inhomogeneities just
Coefficient, use suitable κ parameters will be so that planar structure all obtains high response.
Step S4 is specially：The pixel included in the 3D data volume that step S1 is obtained, these pixels are extracted each
Convolution fortune is carried out from residing three dimensional window, and with regard to respective three dimensional window and the three dimensional anisotropic operator of all difference
Calculate.Three dimensional window residing for extraction pixel uses nonlinear filtering technique well known in the art, and the application does not do explain in detail herein
State.
Step S5 is specially：Each convolution value obtained according to step S4, selects maximum of which convolution value as optimal phase
It should be worth, direction corresponding to the optimal response value is the position tomography maximum probability direction.Due to difference while convolutional calculation
Characteristic, the position fault attributes are also strengthened.
By the data volume of three-dimensional filtering, the spotted noise in background, and the wire in data volume have been eliminated substantially
Structure has also been remarkably reinforced.But the tomography that generally this step obtains is thicker, and interference is had on tomography, it is clear that this
A little results need to obtain finer processing, the specifically used method of mathematical morphology of the three-dimensional DoS filtering of the application.
Morphology processing is basic using algebra of sets, and with the method for set theory come the structure of description collections.It is this
Shape and structure of the mathematical method commonly used to analysis set.In mathematical morphology and Digital Image Processing, top cap conversion and
Bottom cap conversion is that the operation of small element and details is extracted from given image.Opening operation is subtracted with original image, as a result as top cap
Conversion；Original image is subtracted with closed operation, as a result as bottom cap converts.Top cap is converted for various image processing tasks, such as
Feature extraction, background equalization, image enhaucament etc..
Top cap converts as follows with the mathematical definition of bottom cap conversion：
Make I:E → R is a secondary bianry image, will point from Euclidean space or discrete grid block domain mapping to solid line.With S '
A structural elements are represented, then top cap operation can be represented such as formula (8)：
Wherein,Represent opening operation, Represent erosion operation,Represent dilation operation.
Similarly, bottom cap operation can be represented such as formula (9)：
Bhat(I)=(IS ')-I (9)
Wherein, closed operation is represented,
The deleted object of fitting is compared to, these conversion are more likely to use and open operation or closed operation, utilize some
Structural elements Delete Objects from original image.Next, carrying out poor operation, as a result exactly a width comprises only and has deleted component
Image.When brighter wisp in than dark overall situation be present, converted usually using top cap；If in contrast, use bottom
Cap converts.For this reason, when speaking of the two conversion, white top cap conversion and the conversion of black matrix cap are usually respectively referred to as.
The image of final output is：
Fp=I+That(I)-Bhat(I) (10)
Consider below using the intensity level in gray level image as height above sea level, to be specifically described the basic of top bottom cap conversion
Principle.Then now scene is made up of mountain top (most bright spot) and the lowest point (most dark areas).Uneven contrast in image can usually drop
The threshold value isolation on adjacent mountain top between low ebb.For this problem, filtered using the top cap for opening operationIt is bright to strengthen
Point, strengthen mountain valley using bottom cap filtering (IS ')-I of closed operation.
One of ordinary skill in the art will be appreciated that embodiment described here is to aid in reader and understands this hair
Bright principle, it should be understood that protection scope of the present invention is not limited to such especially statement and embodiment.For ability
For the technical staff in domain, the present invention can have various modifications and variations.Within the spirit and principles of the invention, made
Any modification, equivalent substitution and improvements etc., should be included within scope of the presently claimed invention.
Claims (4)
1. the three-dimensional data tomography Enhancement Method based on DoS filtering, it is characterised in that including：
S1, input filter the 3D data volume of noise reduction by structure directing；
S2, the Plane of rotation core for introducing three dimensions, generate all monoplane three-dimensional operators corresponding to the 3D data volume；
S3, a parallel plane, and the current plane three-dimensional operator generated according to step S2 are respectively introduced in current plane both sides, it is raw
Into the three dimensional anisotropic operator of difference corresponding to current plane；
S4, following processing is done to each pixel in step S1 3D data volume：A untreated pixel is chosen, is carried
The three dimensional window centered on the pixel is taken, the three dimensional anisotropic of the obtained all difference of three dimensional window and step S3 is calculated
Son carries out convolution algorithm；
S5, step S4 is obtained to maximum convolution value in all convolution values as optimal response value, corresponding to the optimal response value
Direction is the position tomography maximum probability direction.
2. the three-dimensional data tomography Enhancement Method according to claim 1 based on DoS filtering, it is characterised in that step S2
Specially：
S21, by cube length of window L, obtain 3L2The plane of -6L+4 different directions；
S22, establish spherical coordinate system to cube window, travel through on sphere each pixel in spherical coordinate system corresponding inclination angle and
Azimuth, obtain the normal vector direction of all planes；
S23, each monoplane three-dimensional operator obtained according to the normal vector direction of each plane.
3. the three-dimensional data tomography Enhancement Method according to claim 2 based on DoS filtering, it is characterised in that step S23
Also include：Planar process vector direction is：(x⊥,y⊥,z⊥), plane is according to corresponding to determining following formula：
x⊥(x-m)+y⊥(y-m)+z⊥(z-m)=0
Wherein, x represents x-axis variable, and y represents y-axis variable, and z represents z-axis variable, and (m, m, m) represents the central point of cube window.
4. the three-dimensional data tomography Enhancement Method according to claim 3 based on DoS filtering, it is characterised in that step S3
Specially：
S31, in the both sides of current plane a parallel plane is respectively introduced, be designated as the first parallel plane and second parallel flat respectively
Face；
S32, calculate two nonlinear differences perpendicular to current plane；
Wherein,The first plane-parallel nonlinear difference is represented,Represent the second plane-parallel non-linear difference
Point, θ represents pixel corresponding inclination angle in spherical coordinate system,Represent pixel corresponding azimuth, p in spherical coordinate systemMTable
Show current plane, pLAnd pRRepresent positioned at current plane both sides and parallel to two planes of current plane；
S33, monoplane standard deviation is introduced, two nonlinear differences perpendicular to mid-plane obtained according to step S32, calculated
Two dimensional strengths are obtained to measure；
High response is obtained by adjustment parameter κ planar structures；
Wherein,Represent corresponding monoplane standard deviation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710757287.6A CN107621655B (en) | 2017-08-29 | 2017-08-29 | Three-dimensional data fault enhancement method based on DoS filtering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710757287.6A CN107621655B (en) | 2017-08-29 | 2017-08-29 | Three-dimensional data fault enhancement method based on DoS filtering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107621655A true CN107621655A (en) | 2018-01-23 |
| CN107621655B CN107621655B (en) | 2020-06-02 |
Family
ID=61089167
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710757287.6A Active CN107621655B (en) | 2017-08-29 | 2017-08-29 | Three-dimensional data fault enhancement method based on DoS filtering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107621655B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112130200A (en) * | 2020-09-23 | 2020-12-25 | 电子科技大学 | Fault identification method based on grad-CAM attention guidance |
| CN113702257A (en) * | 2021-08-09 | 2021-11-26 | 西南石油大学 | Conglomerate pore structure characterization method based on CT three-dimensional data volume |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2372402A1 (en) * | 2010-03-22 | 2011-10-05 | PGS Geophysical AS | Self-Positioning Nodal Geophysical Recorder |
| CN102681029A (en) * | 2012-05-11 | 2012-09-19 | 威海双丰物探设备股份有限公司 | Detector parameter testing device and method |
| CN103135136A (en) * | 2011-11-25 | 2013-06-05 | 中国石油化工股份有限公司 | Automatic fault interpretation device for three-dimensional seismic data body |
| CN104020492A (en) * | 2013-07-01 | 2014-09-03 | 西安交通大学 | Edge-preserving filtering method of three-dimensional earthquake data |
| WO2015199727A1 (en) * | 2014-06-27 | 2015-12-30 | Hewlett-Packard Development Company, L.P. | Finding faults in geological data |
-
2017
- 2017-08-29 CN CN201710757287.6A patent/CN107621655B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2372402A1 (en) * | 2010-03-22 | 2011-10-05 | PGS Geophysical AS | Self-Positioning Nodal Geophysical Recorder |
| CN103135136A (en) * | 2011-11-25 | 2013-06-05 | 中国石油化工股份有限公司 | Automatic fault interpretation device for three-dimensional seismic data body |
| CN102681029A (en) * | 2012-05-11 | 2012-09-19 | 威海双丰物探设备股份有限公司 | Detector parameter testing device and method |
| CN104020492A (en) * | 2013-07-01 | 2014-09-03 | 西安交通大学 | Edge-preserving filtering method of three-dimensional earthquake data |
| WO2015199727A1 (en) * | 2014-06-27 | 2015-12-30 | Hewlett-Packard Development Company, L.P. | Finding faults in geological data |
Non-Patent Citations (2)
| Title |
|---|
| 彭双 等: ""结合线性形状特征与模糊连接度的三维CT图像肺裂检测"", 《计算机应用与软件》 * |
| 杨培杰 等: ""方向性边界保持断层增强技术"", 《地球物理学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112130200A (en) * | 2020-09-23 | 2020-12-25 | 电子科技大学 | Fault identification method based on grad-CAM attention guidance |
| CN112130200B (en) * | 2020-09-23 | 2021-07-20 | 电子科技大学 | Fault identification method based on grad-CAM attention guidance |
| CN113702257A (en) * | 2021-08-09 | 2021-11-26 | 西南石油大学 | Conglomerate pore structure characterization method based on CT three-dimensional data volume |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107621655B (en) | 2020-06-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kim et al. | A dense stereo matching using two-pass dynamic programming with generalized ground control points | |
| CN106408513B (en) | Depth map super resolution ratio reconstruction method | |
| NagaRaju et al. | Morphological edge detection algorithm based on multi-structure elements of different directions | |
| CN105913407B (en) | A method of poly focal power image co-registration is optimized based on differential chart | |
| AU2005308450A1 (en) | System and method for fault identification | |
| CN110287780B (en) | Method for extracting facial image features under illumination | |
| CN110544300B (en) | Method for automatically generating three-dimensional model based on two-dimensional hand-drawn image characteristics | |
| CN111242864A (en) | Finger vein image restoration method based on Gabor texture constraint | |
| CN111310688A (en) | Finger vein identification method based on multi-angle imaging | |
| CN107621655A (en) | Three-dimensional data tomography Enhancement Method based on DoS filtering | |
| CN107451595A (en) | Infrared image salient region detection method based on hybrid algorithm | |
| Condorelli et al. | A comparison between 3D reconstruction using nerf neural networks and mvs algorithms on cultural heritage images | |
| CN106650663A (en) | Building true/false change judgement method and false change removal method comprising building true/false change judgement method | |
| CN114862710A (en) | Infrared and visible light image fusion method and device | |
| JP2014106713A (en) | Program, method, and information processor | |
| CN108090460A (en) | Expression recognition feature extraction algorithm based on multi-direction description of weber | |
| Megyesi et al. | Affine propagation for surface reconstruction in wide baseline stereo | |
| CN109948291A (en) | A kind of discontinuous boundary line direction-adaptive recognition methods of sand body | |
| Eshaghzadeh et al. | Canny edge detection algorithm application for analysis of the potential field map | |
| Jindal et al. | An ensemble mosaicing and ridgelet based fusion technique for underwater panoramic image reconstruction and its refinement | |
| Ratner et al. | Image enhancement using elastic manifolds | |
| CN115841618A (en) | Remote sensing image coastline extraction method based on clustering and edge detection | |
| Ma | Morphological Structural Element Optimization for Automatic Line Characterization of Ink Drawings | |
| Wang et al. | Accurate extraction method for structural features of building facades through texture fusion | |
| Zhang et al. | Fingerprint orientation field interpolation based on the constrained delaunay triangulation |
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 |