CN107170000A - The stereopsis dense Stereo Matching method optimized based on global block - Google Patents

The stereopsis dense Stereo Matching method optimized based on global block Download PDF

Info

Publication number
CN107170000A
CN107170000A CN201710254284.0A CN201710254284A CN107170000A CN 107170000 A CN107170000 A CN 107170000A CN 201710254284 A CN201710254284 A CN 201710254284A CN 107170000 A CN107170000 A CN 107170000A
Authority
CN
China
Prior art keywords
mrow
msub
mtd
mover
block
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
Application number
CN201710254284.0A
Other languages
Chinese (zh)
Other versions
CN107170000B (en
Inventor
黄旭
周刚
高其双
胡堃
陆正武
蔡刚山
范超
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan engineering science and technology research institute
Original Assignee
Wuhan engineering science and technology research institute
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Wuhan engineering science and technology research institute filed Critical Wuhan engineering science and technology research institute
Priority to CN201710254284.0A priority Critical patent/CN107170000B/en
Publication of CN107170000A publication Critical patent/CN107170000A/en
Application granted granted Critical
Publication of CN107170000B publication Critical patent/CN107170000B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G06T7/32—Determination of transform parameters for the alignment of images, i.e. image registration using correlation-based methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
    • G06T7/344—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving models

Abstract

The invention discloses a kind of the present invention relates to the stereopsis dense Stereo Matching method optimized based on global block, step is：1st, reference images are selected and image is referred to, just matching disparity map is obtained using traditional stereopsis dense Stereo Matching algorithm；2nd, using super-pixel segmentation technology, reference images are divided into a series of blocks adjoined each other；3rd, the data item in global energy function is built；4th, the smooth item in global energy function is built；5th, according to data item and smooth item, global energy function is set up, and function is resolved using the method for least square, globally optimal solution is obtained, generates accurate, continuous, smooth matching disparity map.The present invention can be eliminated effectively in " parallax ladder " problem of generally existing in conventional stereo image dense Stereo Matching algorithm, the matching disparity map of generation, model surface continuous and derivable, and reconstruction accuracy is high.

Description

The stereopsis dense Stereo Matching method optimized based on global block
Technical field
The present invention relates to stereopsis dense Stereo Matching technical field, in particular to a kind of three-dimensional shadow optimized based on global block As dense Stereo Matching method.
Technical background
Stereopsis dense Stereo Matching refers to according to certain Matching power flow optimization method, match pixel by pixel two width images it Between same place process.According to the difference of optimization method, stereopsis dense Stereo Matching method can be divided into：Local dense Match somebody with somebody, half global dense Stereo Matching and global dense Stereo Matching.Stereopsis dense Stereo Matching is that three-dimensional modeling, virtual reality, computer are regarded One of core technologies in field such as feel, can be used in unmanned automatic driving, unmanned plane automatic cruising, 3D maps, 3D printing, wisdom City, Robot Binocular Vision, virtual electric business etc. are applied.
The dimensional Modeling Technology of current main flow includes：Manually modeling technique, laser scanner technique, structured light technique and Stereopsis dense Stereo Matching technology.Compared with remaining two kinds of technology, the advantage of stereopsis dense Stereo Matching technology is：(1) into This is cheap；(2) plane precision is high；(3) modeling scope is big；(4) colouring information of threedimensional model is reflected.Traditional stereopsis is close Collection matching algorithm includes bilateral filtering algorithm, half overall situation dense Stereo Matching algorithm (Semi-global Matching, SGM), figure and cut Algorithm, belief propagation algorithm etc..These algorithms always assume that the parallax in image between adjacent pixel will be met as far as possible and regarded The consistent constraint of difference.In fact, in the chamfered region of threedimensional model, the parallax of adjacent pixel is impossible consistent.If Chamfered region requires that the parallax of adjacent pixel is consistent, " parallax ladder " problem will be produced in chamfered region, so as to cause three-dimensional Model surface is coarse, influences the reconstruction precision and Three-dimensional Display effect of threedimensional model.
The content of the invention
The purpose of the present invention is that there is provided a kind of intensive of stereopsis optimized based on global block for above-mentioned technical problem Method of completing the square, this method assumes that threedimensional model scene is that piecemeal is continuous, the reference images in stereopsis is divided into a series of The block adjoined each other, builds global energy function (including data item and smooth item), and stereopsis dense Stereo Matching problem is converted For the optimal solution computational problem of global energy function, optimal solution is obtained using the method for least square, it is intensive as stereopsis The result of matching.The present invention can effectively solve the problem that " parallax ladder " problem of generally existing in conventional stereo image dense Stereo Matching, So that the model surface continuous and derivable after matching.
In order to achieve this, the stereopsis dense Stereo Matching method optimized based on global block designed by the present invention, it is special Levy and be, it comprises the following steps：
Step 1：In stereopsis, select reference images and refer to image, using traditional stereopsis dense Stereo Matching Method, obtains the disparity map of just matching；
Step 2：Using SLIC (Simple Linear Iterative Cluster) superpixel segmentation method, by benchmark Image Segmentation uses S into a series of blocks (Patch) adjoined each otheriRepresent i-th piece in reference images；
Step 3：Build the data item E in the stereopsis dense Stereo Matching global energy function optimized based on global blockdata, Data item EdataFor describing each block of reference images and estimating with reference to the non-similarity between the same name block (NAM) on image；
Each described piece is described with a disparity plane equation, i.e.,：
Wherein, ai、bi、ciRepresent block SiDisparity plane equation parameter；P=(px, py)TRepresent block SiAn interior pixel； D represents the parallax corresponding to pixel；Represent the center of gravity coordinate of pixel p；
OrderRepresent block SiThe correction of corresponding disparity plane equation coefficient,Represent that the correction of all disparity plane equation coefficients in reference images is constituted not Know number vector, τ represents the number of block, i ∈ 1 ... τ, the data item of energy function can be expressed as：
In formula, GdataRepresent data item EdataQuadratic term coefficient matrix；HdataRepresent data item EdataFirst order be Matrix number, EdataFor the data item of the global energy function of stereopsis dense Stereo Matching optimized based on global block, T represents transposition Symbol, above-mentioned quadratic term coefficient matrix GdataWith Monomial coefficient matrix HdataBe embodied as：
Gdata=Diag (- gdata(Si))；
In above formula, Diag represents diagonal matrix；gdata(Si)、hdata(Si) G is represented respectivelydata、HdataIn corresponding Si Block matrix；A, b, c represent disparity plane equation coefficient；a0、b0、c0Represent the initial value of disparity plane equation coefficient；r(a0,b0, c0|Si) represent block SiCorresponding coefficient correlation；U=a, b or c representative function r (a0,b0, c0|Si) right Unknown number u single order partial differential；u1=a, b or c, u2=a, b or c representative function r (a0,b0, c0| Si) to unknown number u1、u2Second order partial differential；
Step 4：Build the smooth item in the stereopsis dense Stereo Matching global energy function optimized based on global block Esmooth, smooth item EsmoothFor ensureing continuously smooth between adjacent block；
Same orderRepresent the correction composition of the parallax aspect equation parameter corresponding to all pieces Unknown number vector,The unknown number vector that the correction of the τ disparity plane equation coefficient is constituted is represented, will can be put down Sliding item EsmoothIt is expressed as：
In formula, GsRepresent smooth item EsmoothQuadratic term coefficient matrix, HSRepresent smooth item EsmoothMonomial coefficient square Battle array；
Above-mentioned quadratic term coefficient matrix GsWith Monomial coefficient matrix HSIt can be expressed as respectively：
Wherein, τ represents the number of block；SjRepresent SiAdjacent block；N(Si) represent block SiAdjacent set of blocks；E(Si,Sj) Represent block SiIt is interior, with block SjAdjacent pixel set；|E(Si,Sj) | represent set E (Si,Sj) in number of pixels；ci= (cix,ciy)TRepresent block SiCenter of gravity；Pn1(i, j) is represented according to block SiWith block SjThe connectivity calculated of syntople punish Penalty factor；Pn2(i, j) is represented according to block SiWith block SjThe coplanarity penalty coefficient that calculates of syntople；T represents benchmark A pixel on image, the pixel is located at set E (Si,Sj) in, and be the pixel of adjoiner between block and block；gsr(Si,Sj, T) block S is representediWith block SjBetween correlation matrix on pixel t；hsr(Si,Sj, t) represent block SiWith block SjBetween Monomial coefficient Partitioning of matrix matrix；
Wherein, σ1(t,Si,Sj) represent block SiWith block SjBetween quadratic term matrix in block form on pixel t, T is transposition symbol Number；03×3The null matrix of expression 3 × 3；Represent pixel t on block SiCenter of gravity coordinate；Represent Pixel t is on block SjCenter of gravity coordinate；
Wherein, hi(t,Si,Sj) represent hsr(Si,Sj, it is t) interior, on SiMatrix in block form；hj(t,Si,Sj) represent hsr(Si, Sj, it is t) interior, on SjMatrix in block form；Represent according to block SiDisparity plane equation initial value, the pixel t calculated Parallax；Represent according to block SjDisparity plane equation initial value, the pixel t calculated parallax；Table Show pixel t on block SiCenter of gravity coordinate；Represent pixel t on block SjCenter of gravity coordinate；
Step 5：According to data item EdataWith smooth item Esmooth, build intensive of the stereopsis optimized based on global block The global energy function matched somebody with somebody, wherein, above-mentioned global energy Function Extreme Value optimal solution, the as knot of stereopsis dense Stereo Matching Really；
The disparity map that D represents stereopsis matching is defined, E (D) represents intensive of the stereopsis optimized based on global block The global energy function matched somebody with somebody, then, above-mentioned global energy function is defined as：
Wherein, EdataRepresent the data item of the global energy function of the stereopsis dense Stereo Matching optimized based on global block； EsmoothRepresent the smooth item of the global energy function of the stereopsis dense Stereo Matching optimized based on global block；
The derivation of equation Minimum value, be equivalent to askCan directly it be calculated using least square method Obtain the stereopsis dense Stereo Matching disparity map of global optimum.
Compared with prior art, the invention has the advantages that：
The present invention can effectively eliminate the parallax ladder problem of generally existing in conventional stereo image dense Stereo Matching algorithm, adopt In dense Stereo Matching disparity map with the method generation of the present invention, model surface continuous and derivable, matching precision is higher, Three-dimensional Display effect It is really good.The present invention can be digital photogrammetry and remote sensing, computer vision, virtual reality, public work safety, national defense construction Technological service is provided Deng subject and application.
Brief description of the drawings
Fig. 1 is flow chart of the invention；
Embodiment
Below in conjunction with drawings and examples, the present invention is described in further detail：
The present invention is directed to the parallax ladder problem of generally existing in conventional stereo image dense Stereo Matching method, it is proposed that a kind of The stereopsis dense Stereo Matching method optimized based on global block.This method can effectively solve the problem that above-mentioned parallax ladder problem, obtain The disparity map of model surface continuous and derivable.The workflow of the present invention is as shown in figure 1, it comprises the following steps：
Step 1：In stereopsis, select reference images and refer to image, using traditional stereopsis dense Stereo Matching Method, obtains the disparity map of just matching；
Stereopsis is the two width images for constituting stereoscopic vision.Image can use satellite image, aerial images, unmanned plane Image etc., reference images are selected first in two images and image is referred to, image on the basis of left view image is typically chosen, selects Right seeing image picture is refers to image, then by stereopsis resampling nucleation DNA mitochondrial DNA image, and the method for sampling, which can be used, increases income Code library OpenCV in initUndistortRectifyMap () function, according to the core DNA mitochondrial DNA shadow Jing Guo resampling Picture, using traditional stereopsis dense Stereo Matching method, the disparity map of generation just matching, traditional stereopsis dense Stereo Matching side Method includes local matching method, half global registration method and global registration method, still can be using the code library increased income Stereopsis dense Stereo Matching function in OpenCV, local matching method correspondence cvFindStereoCorrespondenceBM () function, half global registration algorithm correspondence SGBM () function, global registration algorithm correspondence CvFindStereoCorrespondenceGC () function；
Step 2：Using SLIC superpixel segmentation methods, reference images are divided into a series of blocks adjoined each other, S is usedi Represent i-th piece in reference images；Each block can be described using a plane, mathematically can be using a parallax Plane equation is described.Super-pixel segmentation code is referring to network address：
http://ivrl.epfl.ch/supplementary_material/RK_SLICSuperpixels/ index.html；
Step 3：Build the data item E in the stereopsis dense Stereo Matching global energy function optimized based on global blockdata, Data item EdataFor describing each block of reference images and estimating with reference to the non-similarity between the same name block (NAM) on image, if Data item is bigger, illustrates that non-similarity is estimated bigger；Conversely, explanation non-similarity estimates smaller；
Each described piece is described with a disparity plane equation, i.e.,：
Wherein, ai、bi、ciRepresent block SiDisparity plane equation parameter；P=(px, py)TRepresent block SiAn interior pixel； D represents the parallax corresponding to pixel；Represent the center of gravity coordinate of pixel p；
OrderRepresent block SiThe correction of corresponding disparity plane equation coefficient,Represent that the correction of all disparity plane equation coefficients in reference images is constituted not Know number vector, τ represents the number of block, i ∈ 1 ... τ, the data item of energy function can be expressed as：
In formula, GdataRepresent data item EdataQuadratic term coefficient matrix；HdataRepresent data item EdataFirst order be Matrix number, EdataFor the data item of the global energy function of stereopsis dense Stereo Matching optimized based on global block, T represents transposition Symbol, above-mentioned quadratic term coefficient matrix GdataWith Monomial coefficient matrix HdataBe embodied as：
Gdata=Diag (- gdata(Si))；
In above formula, Diag represents diagonal matrix；gdata(Si)、hdata(Si) G is represented respectivelydata、HdataIn corresponding Si Block matrix；A, b, c represent disparity plane equation coefficient；a0、b0、c0Represent the initial value of disparity plane equation coefficient；r(a0,b0, c0|Si) represent block SiCorresponding coefficient correlation；U=a, b or c representative function r (a0,b0, c0|Si) right Unknown number u single order partial differential；u1=a, b or c, u2=a, b or c representative function r (a0,b0, c0| Si) to unknown number u1、u2Second order partial differential；
Step 4：Build the smooth item in the stereopsis dense Stereo Matching global energy function optimized based on global block Esmooth, smooth item EsmoothFor ensureing continuously smooth between adjacent block；
Same orderRepresent the correction composition of the parallax aspect equation parameter corresponding to all pieces Unknown number vector,The unknown number vector that the correction of the τ disparity plane equation coefficient is constituted is represented, will can be put down Sliding item EsmoothIt is expressed as：
In formula, GsRepresent smooth item EsmoothQuadratic term coefficient matrix, HSRepresent smooth item EsmoothMonomial coefficient square Battle array；
Above-mentioned quadratic term coefficient matrix GsWith Monomial coefficient matrix HSIt can be expressed as respectively：
Wherein, τ represents the number of block；SjRepresent SiAdjacent block；N(Si) represent block SiAdjacent set of blocks；E(Si,Sj) Represent block SiInterior and block SjAdjacent pixel set；|E(Si,Sj) | represent set E (Si,Sj) in number of pixels；ci= (cix,ciy)TRepresent block SiCenter of gravity；Pn1(i, j) is represented according to block SiWith block SjThe connectivity calculated of syntople punish Penalty factor；Pn2(i, j) is represented according to block SiWith block SjThe coplanarity penalty coefficient that calculates of syntople；T represents benchmark A pixel on image, the pixel is located at set E (Si,Sj) in, and be the pixel of adjoiner between block and block；gsr(Si,Sj, T) block S is representediWith block SjBetween correlation matrix on pixel t；hsr(Si,Sj, t) represent block SiWith block SjBetween Monomial coefficient Partitioning of matrix matrix；
Wherein, σ1(t,Si,Sj) represent block SiWith block SjBetween quadratic term matrix in block form on pixel t, T is transposition symbol Number；03×3The null matrix of expression 3 × 3；Represent pixel t on block SiCenter of gravity coordinate；Represent Pixel t is on block SjCenter of gravity coordinate；
Wherein, hi(t,Si,Sj) represent hsr(Si,Sj, it is t) interior, on SiMatrix in block form；hj(t,Si,Sj) represent hsr(Si, Sj, it is t) interior, on SjMatrix in block form；Represent according to block SiDisparity plane equation initial value, the pixel t calculated Parallax；Represent according to block SjDisparity plane equation initial value, the pixel t calculated parallax；Table Show pixel t on block SiCenter of gravity coordinate；Represent pixel t on block SjCenter of gravity coordinate；
Step 5：According to data item EdataWith smooth item Esmooth, build intensive of the stereopsis optimized based on global block The global energy function matched somebody with somebody, wherein, above-mentioned global energy Function Extreme Value optimal solution, the as knot of stereopsis dense Stereo Matching Really；
The disparity map that D represents stereopsis matching is defined, E (D) represents intensive of the stereopsis optimized based on global block The global energy function matched somebody with somebody, then, above-mentioned global energy function is defined as：
Wherein, EdataRepresent the data item of the global energy function of the stereopsis dense Stereo Matching optimized based on global block； EsmoothRepresent the smooth item of the global energy function of the stereopsis dense Stereo Matching optimized based on global block；
The derivation of equation Minimum value, be equivalent to askCan directly it be calculated using least square method Obtain the stereopsis dense Stereo Matching disparity map of global optimum." parallax ladder " problem of solution, obtains the model table of continuous and derivable Face, the code of specific least square may refer to the eigen storehouses increased income：
http://eigen.tuxfamily.org/index.phpTitle=Main_Page
In above-mentioned technical proposal, in the step 5, resolveAfterwards, that is, the disparity plane equation parameter of each block is obtained Correction (dai,dbi,dci)T, then according to the disparity plane equation parameter initial value and correction of each block, calculate each block essence True disparity plane equation parameter：
ai=a0+dai,bi=b0+dbi,ci=c0+dci
In formula, ai、bi、ciRepresent block SiDisparity plane equation parameter；a0、b0、c0Represent block SiDisparity plane equation ginseng Several initial values.
According to the image coordinate and disparity plane equation parameter of pixel in each piece, the accurate parallax of each pixel is calculated, It is shown below：
Wherein, ai、bi、ciRepresent block SiDisparity plane equation parameter；Represent block SiInterior one The image coordinate of pixel；D represents the parallax corresponding to image coordinate, and the parallax of all pixels in each block is calculated successively, obtains The disparity map of stereopsis dense Stereo Matching.
In the step 3 of above-mentioned technical proposal, correlation coefficient r (a0,b0, c0|Si) expression formula be：
In above formula, IbRepresent reference images, IrExpression refers to image, and p represents block SiAn interior pixel；Represent Block SiThe average gray of interior all pixels,Represent block SiAll pixels is flat in same name block (NAM) on correspondence reference image Equal gray scale, Ib(p) gray scale of pixel p in reference images, I are representedr(q) gray scale with reference to pixel q on image, pixel p and picture are represented Plain q belongs to matching same place.
Single order partial differentialU=a, b or c are by correlation coefficient r (a0,b0, c0|Si) to the one of unknown number u Rank differential calculation is obtained, second order partial differentialu1=a, b or c, u2=a, b or c are by correlation coefficient r (a0,b0, c0|Si) to unknown number u1、u2Second-order differential calculate obtain.
In the step 3 of above-mentioned technical proposal, block SiWhat correspondingly the position of the same name block (NAM) on reference image can be in step 1 is first Disparity map is matched to obtain.
In the step 4 of above-mentioned technical proposal, center of gravity coordinateCalculation be：
In formula, (tx,ty)TRepresent coordinates of the pixel t in reference images；(cix,ciy)TRepresent block SiBarycentric coodinates.
In the step 4 of above-mentioned technical proposal, connectivity penalty coefficient Pn1The calculation formula of (i, j) is：
In above formula, P represents the penalty coefficient that user specifies, and prevailing value takes 5~10, exp to represent exponential function；σ is represented Pixel grey scale smoothing factor, typically takes 5~10,Represent block SiThe average gray of interior all pixels,Represent block Sj The average gray of interior all pixels.
In the step 4 of above-mentioned technical proposal, coplanarity penalty coefficient Pn2The calculation formula of (i, j) is：
Pn2(i, j)=Pn1(i,j)·(cos<Si,Sj>)m
In above formula, cos represents cosine function,<Si,Sj>Represent block SiWith block SjBetween normal vector angle, m represent power because Son, typically takes 10, Pn1(i, j) is represented according to block SiWith block SjThe connectivity penalty coefficient that calculates of syntople.
The present invention can effectively solve the problem that " parallax " ladder problem of generally existing in conventional stereo image dense Stereo Matching algorithm, Stereopsis dense Stereo Matching problem is converted into the optimal solution computational problem of global energy function, acquired stereopsis is intensive Match in disparity map, model surface continuous and derivable, be that three-dimensional modeling, virtual reality, computer vision, digital photogrammetry etc. are learned Section and application provide technical support.
The content that this specification is not described in detail belongs to prior art known to professional and technical personnel in the field.

Claims (6)

1. a kind of stereopsis dense Stereo Matching method optimized based on global block, it is characterised in that it comprises the following steps：
Step 1：In stereopsis, select reference images and refer to image, using traditional stereopsis dense Stereo Matching method, Obtain the disparity map of just matching；
Step 2：Using SLIC superpixel segmentation methods, reference images are divided into a series of blocks adjoined each other, S is usediRepresent base I-th piece on quasi- image；
Step 3：Build the data item Edata in the stereopsis dense Stereo Matching global energy function optimized based on global block, number According to item Edata for describing each block of reference images and estimating with reference to the non-similarity between the same name block (NAM) on image；
Each described piece is described with a disparity plane equation, i.e.,：
<mrow> <mi>d</mi> <mrow> <mo>(</mo> <mi>p</mi> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mi>a</mi> <mi>i</mi> </msub> <mo>&amp;CenterDot;</mo> <mover> <msub> <mi>p</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>+</mo> <msub> <mi>b</mi> <mi>i</mi> </msub> <mo>&amp;CenterDot;</mo> <mover> <msub> <mi>p</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>+</mo> <msub> <mi>c</mi> <mi>i</mi> </msub> <mo>;</mo> <mi>p</mi> <mo>&amp;Element;</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>.</mo> </mrow>
Wherein, ai、bi、ciRepresent block SiDisparity plane equation parameter；P=(px, py)TRepresent block SiAn interior pixel；D tables Show the parallax corresponding to pixel；Represent the center of gravity coordinate of pixel p；
OrderRepresent block SiThe correction of corresponding disparity plane equation coefficient, The unknown number vector that the correction of all disparity plane equation coefficients in reference images is constituted is represented, τ represents the number of block, i ∈ 1 ... τ, the data item of energy function can be expressed as：
<mrow> <msub> <mi>E</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> </msub> <mo>=</mo> <msup> <mover> <mi>x</mi> <mo>~</mo> </mover> <mi>T</mi> </msup> <msub> <mi>G</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> </msub> <mover> <mi>x</mi> <mo>~</mo> </mover> <mo>-</mo> <mn>2</mn> <msubsup> <mi>H</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> <mi>T</mi> </msubsup> <mover> <mi>x</mi> <mo>~</mo> </mover> </mrow>
In formula, GdataRepresent data item EdataQuadratic term coefficient matrix；HdataRepresent data item EdataFirst order coefficient square Battle array, EdataFor the data item of the global energy function of stereopsis dense Stereo Matching optimized based on global block, T represents that transposition is accorded with Number, above-mentioned quadratic term coefficient matrix GdataWith Monomial coefficient matrix HdataBe embodied as：
Gdata=Diag (- gdata(Si))；
<mrow> <msub> <mi>H</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> </msub> <mo>=</mo> <msup> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mrow> <mo>-</mo> <msubsup> <mi>h</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> <mi>T</mi> </msubsup> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mn>1</mn> </msub> <mo>)</mo> </mrow> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <msubsup> <mi>h</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> <mi>T</mi> </msubsup> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mn>2</mn> </msub> <mo>)</mo> </mrow> </mrow> </mtd> <mtd> <mn>...</mn> </mtd> <mtd> <mrow> <mo>-</mo> <msubsup> <mi>h</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> <mi>T</mi> </msubsup> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>&amp;tau;</mi> </msub> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> </mtable> </mfenced> <mi>T</mi> </msup> <mo>;</mo> </mrow>
<mrow> <mtable> <mtr> <mtd> <mrow> <msub> <mi>g</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> </msub> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>=</mo> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mfrac> <mrow> <msup> <mo>&amp;part;</mo> <mn>2</mn> </msup> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>a</mi> <mo>&amp;part;</mo> <mi>a</mi> </mrow> </mfrac> </mtd> <mtd> <mfrac> <mrow> <msup> <mo>&amp;part;</mo> <mn>2</mn> </msup> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>a</mi> <mo>&amp;part;</mo> <mi>b</mi> </mrow> </mfrac> </mtd> <mtd> <mfrac> <mrow> <msup> <mo>&amp;part;</mo> <mn>2</mn> </msup> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>a</mi> <mo>&amp;part;</mo> <mi>c</mi> </mrow> </mfrac> </mtd> </mtr> <mtr> <mtd> <mfrac> <mrow> <msup> <mo>&amp;part;</mo> <mn>2</mn> </msup> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>a</mi> <mo>&amp;part;</mo> <mi>b</mi> </mrow> </mfrac> </mtd> <mtd> <mfrac> <mrow> <msup> <mo>&amp;part;</mo> <mn>2</mn> </msup> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>SS</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>b</mi> <mo>&amp;part;</mo> <mi>b</mi> </mrow> </mfrac> </mtd> <mtd> <mfrac> <mrow> <msup> <mo>&amp;part;</mo> <mn>2</mn> </msup> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>b</mi> <mo>&amp;part;</mo> <mi>c</mi> </mrow> </mfrac> </mtd> </mtr> <mtr> <mtd> <mfrac> <mrow> <msup> <mo>&amp;part;</mo> <mn>2</mn> </msup> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>a</mi> <mo>&amp;part;</mo> <mi>c</mi> </mrow> </mfrac> </mtd> <mtd> <mfrac> <mrow> <msup> <mo>&amp;part;</mo> <mn>2</mn> </msup> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>b</mi> <mo>&amp;part;</mo> <mi>c</mi> </mrow> </mfrac> </mtd> <mtd> <mfrac> <mrow> <msup> <mo>&amp;part;</mo> <mn>2</mn> </msup> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>c</mi> <mo>&amp;part;</mo> <mi>c</mi> </mrow> </mfrac> </mtd> </mtr> </mtable> </mfenced> </mrow> </mtd> </mtr> </mtable> <mo>;</mo> </mrow>
<mrow> <mtable> <mtr> <mtd> <mrow> <msub> <mi>h</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> </msub> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>=</mo> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mrow> <mo>-</mo> <mfrac> <mrow> <mo>&amp;part;</mo> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>a</mi> </mrow> </mfrac> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mfrac> <mrow> <mo>&amp;part;</mo> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>b</mi> </mrow> </mfrac> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mfrac> <mrow> <mo>&amp;part;</mo> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <mi>c</mi> </mrow> </mfrac> </mrow> </mtd> </mtr> </mtable> </mfenced> </mrow> </mtd> </mtr> </mtable> <mo>;</mo> </mrow>
In above formula, Diag represents diagonal matrix；gdata(Si)、hdata(Si) G is represented respectivelydata、HdataIn corresponding SiBlock square Battle array；A, b, c represent disparity plane equation coefficient；a0、b0、c0Represent the initial value of disparity plane equation coefficient；r(a0,b0,c0|Si) Represent block SiCorresponding coefficient correlation；U=a, b or c representative function r (a0,b0,c0|Si) to unknown number U single order partial differential；u1=a, b or c, u2=a, b or c representative function r (a0,b0,c0|Si) to not Know several u1、u2Second order partial differential；
Step 4：Build the smooth item E in the stereopsis dense Stereo Matching global energy function optimized based on global blocksmooth, put down Sliding item EsmoothFor ensureing continuously smooth between adjacent block；
Same orderRepresent that the correction of the parallax aspect equation parameter corresponding to all pieces is constituted not Know number vector,The unknown number vector that the correction of the τ disparity plane equation coefficient is constituted is represented, can be by smooth item EsmoothIt is expressed as：
<mrow> <msub> <mi>E</mi> <mrow> <mi>s</mi> <mi>m</mi> <mi>o</mi> <mi>o</mi> <mi>t</mi> <mi>h</mi> </mrow> </msub> <mo>=</mo> <msup> <mover> <mi>x</mi> <mo>~</mo> </mover> <mi>T</mi> </msup> <msub> <mi>G</mi> <mi>s</mi> </msub> <mover> <mi>x</mi> <mo>~</mo> </mover> <mo>-</mo> <mn>2</mn> <msubsup> <mi>H</mi> <mi>s</mi> <mi>T</mi> </msubsup> <mover> <mi>x</mi> <mo>~</mo> </mover> </mrow>
In formula, GsRepresent smooth item EsmoothQuadratic term coefficient matrix, HSRepresent smooth item EsmoothMonomial coefficient matrix；
Above-mentioned quadratic term coefficient matrix GsWith Monomial coefficient matrix HSIt can be expressed as respectively：
<mrow> <msub> <mi>G</mi> <mi>s</mi> </msub> <mo>=</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>&amp;tau;</mi> </munderover> <mrow> <mo>(</mo> <munder> <mo>&amp;Sigma;</mo> <mrow> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>&amp;Element;</mo> <mi>N</mi> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> </munder> <mo>&amp;lsqb;</mo> <mfrac> <mrow> <msub> <mi>Pn</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> </mrow> <mrow> <mo>|</mo> <mi>E</mi> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mo>|</mo> </mrow> </mfrac> <munder> <mo>&amp;Sigma;</mo> <mrow> <mi>t</mi> <mo>&amp;Element;</mo> <mi>E</mi> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> </mrow> </munder> <msub> <mi>g</mi> <mrow> <mi>s</mi> <mi>r</mi> </mrow> </msub> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>,</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>Pn</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> <msub> <mi>g</mi> <mrow> <mi>s</mi> <mi>r</mi> </mrow> </msub> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>,</mo> <msub> <mi>c</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> <mo>)</mo> </mrow>
<mrow> <msub> <mi>H</mi> <mi>s</mi> </msub> <mo>=</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>&amp;tau;</mi> </munderover> <mrow> <mo>(</mo> <munder> <mo>&amp;Sigma;</mo> <mrow> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>&amp;Element;</mo> <mi>N</mi> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> </munder> <mo>&amp;lsqb;</mo> <mfrac> <mrow> <msub> <mi>Pn</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> </mrow> <mrow> <mo>|</mo> <mi>E</mi> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mo>|</mo> </mrow> </mfrac> <munder> <mo>&amp;Sigma;</mo> <mrow> <mi>t</mi> <mo>&amp;Element;</mo> <mi>E</mi> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> </mrow> </munder> <msub> <mi>h</mi> <mrow> <mi>s</mi> <mi>r</mi> </mrow> </msub> <msup> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>,</mo> <mi>t</mi> <mo>)</mo> </mrow> <mi>T</mi> </msup> <mo>+</mo> <msub> <mi>Pn</mi> <mn>2</mn> </msub> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> <msub> <mi>h</mi> <mrow> <mi>s</mi> <mi>r</mi> </mrow> </msub> <msup> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>,</mo> <msub> <mi>c</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mi>T</mi> </msup> <mo>&amp;rsqb;</mo> <mo>)</mo> </mrow>
Wherein, τ represents the number of block；SjRepresent SiAdjacent block；N(Si) represent block SiAdjacent set of blocks；E(Si,Sj) represent Block SiInterior and block SjAdjacent pixel set；|E(Si,Sj) | represent set E (Si,Sj) in number of pixels；ci=(cix,ciy )TRepresent block SiCenter of gravity；Pn1(i, j) is represented according to block SiWith block SjThe connectivity penalty coefficient that calculates of syntople； Pn2(i, j) is represented according to block SiWith block SjThe coplanarity penalty coefficient that calculates of syntople；T is represented in reference images One pixel, the pixel is located at set E (Si,Sj) in, and be the pixel of adjoiner between block and block；gsr(Si,Sj, t) represent block SiWith block SjBetween correlation matrix on pixel t；hsr(Si,Sj, t) represent block SiWith block SjBetween Monomial coefficient matrix point Block matrix；
<mrow> <msub> <mi>&amp;sigma;</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>,</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <msup> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mn>2</mn> </msup> </mtd> <mtd> <mrow> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>&amp;CenterDot;</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> </mtr> <mtr> <mtd> <mrow> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>&amp;CenterDot;</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> <mtd> <msup> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mn>2</mn> </msup> </mtd> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> </mtr> <mtr> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> <mtd> <mn>1</mn> </mtd> </mtr> </mtable> </mfenced> </mrow>
<mrow> <msub> <mi>&amp;sigma;</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>,</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <msup> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mn>2</mn> </msup> </mtd> <mtd> <mrow> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>&amp;CenterDot;</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> </mtr> <mtr> <mtd> <mrow> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>&amp;CenterDot;</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> <mtd> <msup> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mn>2</mn> </msup> </mtd> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> </mtr> <mtr> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> <mtd> <mn>1</mn> </mtd> </mtr> </mtable> </mfenced> </mrow> 2
<mrow> <msub> <mi>&amp;sigma;</mi> <mn>3</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>,</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mrow> <mo>-</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>&amp;CenterDot;</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>&amp;CenterDot;</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>-</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>&amp;CenterDot;</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>&amp;CenterDot;</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>-</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </mtd> </mtr> </mtable> </mfenced> </mrow>
Wherein, σ1(t,Si,Sj) represent block SiWith block SjBetween quadratic term matrix in block form on pixel t, T is transposition symbol；03×3 The null matrix of expression 3 × 3；Represent pixel t on block SiCenter of gravity coordinate；Represent that pixel t is closed In block SjCenter of gravity coordinate；
<mfenced open = "" close = ""> <mtable> <mtr> <mtd> <mrow> <msub> <mi>h</mi> <mrow> <mi>s</mi> <mi>r</mi> </mrow> </msub> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>,</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>=</mo> <msup> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <msubsup> <mn>0</mn> <mrow> <mn>3</mn> <mo>&amp;times;</mo> <mn>3</mn> </mrow> <mi>T</mi> </msubsup> </mtd> <mtd> <mn>...</mn> </mtd> <mtd> <mrow> <msub> <mi>h</mi> <mi>i</mi> </msub> <msup> <mrow> <mo>(</mo> <mi>t</mi> <mo>,</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mi>T</mi> </msup> </mrow> </mtd> <mtd> <mn>...</mn> </mtd> <mtd> <mrow> <msub> <mi>h</mi> <mi>j</mi> </msub> <msup> <mrow> <mo>(</mo> <mi>t</mi> <mo>,</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mi>T</mi> </msup> </mrow> </mtd> <mtd> <mn>...</mn> </mtd> <mtd> <msubsup> <mn>0</mn> <mrow> <mn>3</mn> <mo>&amp;times;</mo> <mn>3</mn> </mrow> <mi>T</mi> </msubsup> </mtd> </mtr> </mtable> </mfenced> <mi>T</mi> </msup> </mrow> </mtd> </mtr> </mtable> </mfenced>
<mrow> <msub> <mi>h</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>,</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <mrow> <mo>(</mo> <msub> <mi>d</mi> <msub> <mi>j</mi> <mn>0</mn> </msub> </msub> <mo>(</mo> <mi>t</mi> <mo>)</mo> <mo>-</mo> <msub> <mi>d</mi> <msub> <mi>i</mi> <mn>0</mn> </msub> </msub> <mo>(</mo> <mi>t</mi> <mo>)</mo> <mo>)</mo> </mrow> <msup> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> <mtd> <mn>1</mn> </mtd> </mtr> </mtable> </mfenced> <mi>T</mi> </msup> </mrow>
<mrow> <msub> <mi>h</mi> <mi>j</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>,</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <mrow> <mo>(</mo> <msub> <mi>d</mi> <msub> <mi>i</mi> <mn>0</mn> </msub> </msub> <mo>(</mo> <mi>t</mi> <mo>)</mo> <mo>-</mo> <msub> <mi>d</mi> <msub> <mi>j</mi> <mn>0</mn> </msub> </msub> <mo>(</mo> <mi>t</mi> <mo>)</mo> <mo>)</mo> </mrow> <msup> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> <mtd> <mover> <msub> <mi>t</mi> <mrow> <mi>j</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> </mtd> <mtd> <mn>1</mn> </mtd> </mtr> </mtable> </mfenced> <mi>T</mi> </msup> </mrow>
Wherein, hi(t,Si,Sj) represent hsr(Si,Sj, it is t) interior, on SiMatrix in block form；hj(t,Si,Sj) represent hsr(Si,Sj, T) in, on SjMatrix in block form；Represent according to block SiDisparity plane equation initial value, the pixel t's calculated regards Difference；Represent according to block SjDisparity plane equation initial value, the pixel t calculated parallax；Represent picture Plain t is on block SiCenter of gravity coordinate；Represent pixel t on block SjCenter of gravity coordinate；
Step 5：According to data item EdataWith smooth item Esmooth, build the complete of the stereopsis dense Stereo Matching based on the optimization of global block Office's energy function, wherein, above-mentioned global energy Function Extreme Value optimal solution, the as result of stereopsis dense Stereo Matching；
The disparity map that D represents stereopsis matching is defined, E (D) represents the stereopsis dense Stereo Matching optimized based on global block Global energy function, then, above-mentioned global energy function is defined as：
<mfenced open = "" close = ""> <mtable> <mtr> <mtd> <mrow> <mi>E</mi> <mrow> <mo>(</mo> <mi>D</mi> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mi>E</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> </msub> <mo>+</mo> <msub> <mi>E</mi> <mrow> <mi>s</mi> <mi>m</mi> <mi>o</mi> <mi>o</mi> <mi>t</mi> <mi>h</mi> </mrow> </msub> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>=</mo> <msup> <mover> <mi>x</mi> <mo>~</mo> </mover> <mi>T</mi> </msup> <mrow> <mo>(</mo> <msub> <mi>G</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> </msub> <mo>+</mo> <msub> <mi>G</mi> <mi>s</mi> </msub> <mo>)</mo> </mrow> <mover> <mi>x</mi> <mo>~</mo> </mover> <mo>-</mo> <mn>2</mn> <mrow> <mo>(</mo> <msubsup> <mi>H</mi> <mrow> <mi>d</mi> <mi>a</mi> <mi>t</mi> <mi>a</mi> </mrow> <mi>T</mi> </msubsup> <mo>+</mo> <msubsup> <mi>H</mi> <mi>s</mi> <mi>T</mi> </msubsup> <mo>)</mo> </mrow> <mover> <mi>x</mi> <mo>~</mo> </mover> </mrow> </mtd> </mtr> </mtable> </mfenced>
Wherein, EdataRepresent the data item of the global energy function of the stereopsis dense Stereo Matching optimized based on global block；Esmooth Represent the smooth item of the global energy function of the stereopsis dense Stereo Matching optimized based on global block；
The derivation of equationMost Small value, is equivalent to askCan directly it be calculated using least square methodObtain The stereopsis dense Stereo Matching disparity map of global optimum.
2. the stereopsis dense Stereo Matching method optimized according to claim 1 based on global block, it is characterised in that：The step In rapid 3, correlation coefficient r (a0,b0,c0|Si) expression formula be：
<mrow> <mi>r</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>c</mi> <mn>0</mn> </msub> <mo>|</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <msub> <mo>&amp;Sigma;</mo> <mrow> <mi>p</mi> <mo>&amp;Element;</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> </mrow> </msub> <mrow> <mo>(</mo> <msub> <mi>I</mi> <mi>b</mi> </msub> <mo>(</mo> <mi>p</mi> <mo>)</mo> <mo>-</mo> <mover> <mrow> <msub> <mi>I</mi> <mi>b</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mo>&amp;OverBar;</mo> </mover> <mo>)</mo> </mrow> <mrow> <mo>(</mo> <msub> <mi>I</mi> <mi>r</mi> </msub> <mo>(</mo> <mi>q</mi> <mo>)</mo> <mo>-</mo> <mover> <mrow> <msub> <mi>I</mi> <mi>r</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>|</mo> <mover> <msub> <mi>n</mi> <mi>i</mi> </msub> <mo>&amp;RightArrow;</mo> </mover> <mo>)</mo> </mrow> </mrow> <mo>&amp;OverBar;</mo> </mover> <mo>)</mo> </mrow> </mrow> <msqrt> <mrow> <msub> <mo>&amp;Sigma;</mo> <mrow> <mi>p</mi> <mo>&amp;Element;</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> </mrow> </msub> <msup> <mrow> <mo>(</mo> <msub> <mi>I</mi> <mi>b</mi> </msub> <mo>(</mo> <mi>p</mi> <mo>)</mo> <mo>-</mo> <mover> <mrow> <msub> <mi>I</mi> <mi>b</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mo>&amp;OverBar;</mo> </mover> <mo>)</mo> </mrow> <mn>2</mn> </msup> <msub> <mo>&amp;Sigma;</mo> <mrow> <mi>p</mi> <mo>&amp;Element;</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> </mrow> </msub> <msup> <mrow> <mo>(</mo> <msub> <mi>I</mi> <mi>r</mi> </msub> <mo>(</mo> <mi>q</mi> <mo>)</mo> <mo>-</mo> <mover> <mrow> <msub> <mi>I</mi> <mi>r</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>|</mo> <mover> <msub> <mi>n</mi> <mi>i</mi> </msub> <mo>&amp;RightArrow;</mo> </mover> <mo>)</mo> </mrow> </mrow> <mo>&amp;OverBar;</mo> </mover> <mo>)</mo> </mrow> <mn>2</mn> </msup> </mrow> </msqrt> </mfrac> </mrow>
In above formula, IbRepresent reference images, IrExpression refers to image, and p represents block SiAn interior pixel；Represent block SiIt is interior The average gray of all pixels,Represent block SiAverage ash of the correspondence with reference to all pixels in the same name block (NAM) on image Degree, Ib(p) gray scale of pixel p in reference images, I are representedr(q) gray scale with reference to pixel q on image, pixel p and pixel q are represented Belong to matching same place；
Single order partial differentialU=a, b or c are by correlation coefficient r (a0,b0,c0|Si) micro- to unknown number u single order Divide to calculate and obtain, second order partial differentialu1=a, b or c, u2=a, b or c are by correlation coefficient r (a0,b0, c0|Si) to unknown number u1、u2Second-order differential calculate obtain.
3. the stereopsis dense Stereo Matching method optimized according to claim 2 based on global block, it is characterised in that：The step In rapid 3, block SiThe first matching disparity map that correspondingly position of the same name block (NAM) on reference image can be in step 1 is obtained.
4. the stereopsis dense Stereo Matching method optimized according to claim 1 based on global block, it is characterised in that：The step In rapid 4, center of gravity coordinateCalculation be：
<mfenced open = "" close = ""> <mtable> <mtr> <mtd> <mrow> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>=</mo> <msub> <mi>t</mi> <mi>x</mi> </msub> <mo>-</mo> <msub> <mi>c</mi> <mrow> <mi>i</mi> <mi>x</mi> </mrow> </msub> </mrow> </mtd> <mtd> <mrow> <mover> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>=</mo> <msub> <mi>t</mi> <mi>y</mi> </msub> <mo>-</mo> <msub> <mi>c</mi> <mrow> <mi>i</mi> <mi>y</mi> </mrow> </msub> </mrow> </mtd> </mtr> </mtable> </mfenced>
In formula, (tx,ty)TRepresent coordinates of the pixel t in reference images；(cix,ciy)TRepresent block SiBarycentric coodinates.
5. the stereopsis dense Stereo Matching method optimized according to claim 1 based on global block, it is characterised in that：The step In rapid 4, connectivity penalty coefficient Pn1The calculation formula of (i, j) is：
<mrow> <msub> <mi>Pn</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> <mo>=</mo> <mi>P</mi> <mo>&amp;CenterDot;</mo> <mi>exp</mi> <mrow> <mo>(</mo> <mo>-</mo> <mo>|</mo> <mover> <mrow> <msub> <mi>I</mi> <mi>b</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <mover> <mrow> <msub> <mi>I</mi> <mi>b</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>S</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> </mrow> <mo>&amp;OverBar;</mo> </mover> <mo>|</mo> <mo>/</mo> <mi>&amp;sigma;</mi> <mo>)</mo> </mrow> </mrow>
In above formula, P represents the penalty coefficient that user specifies, and exp represents exponential function；σ represents pixel grey scale smoothing factor,Represent block SiThe average gray of interior all pixels,Represent block SjThe average gray of interior all pixels.
6. the stereopsis dense Stereo Matching method optimized according to claim 1 or 5 based on global block, it is characterised in that：Institute State in step 4, coplanarity penalty coefficient Pn2The calculation formula of (i, j) is：
Pn2(i, j)=Pn1(i,j)·(cos<Si,Sj>)m
In above formula, cos represents cosine function,<Si,Sj>Represent block SiWith block SjBetween normal vector angle, m represents power factor, Pn1(i, j) is represented according to block SiWith block SjThe connectivity penalty coefficient that calculates of syntople.
CN201710254284.0A 2017-04-18 2017-04-18 Stereopsis dense Stereo Matching method based on the optimization of global block Active CN107170000B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710254284.0A CN107170000B (en) 2017-04-18 2017-04-18 Stereopsis dense Stereo Matching method based on the optimization of global block

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710254284.0A CN107170000B (en) 2017-04-18 2017-04-18 Stereopsis dense Stereo Matching method based on the optimization of global block

Publications (2)

Publication Number Publication Date
CN107170000A true CN107170000A (en) 2017-09-15
CN107170000B CN107170000B (en) 2019-09-10

Family

ID=59812218

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710254284.0A Active CN107170000B (en) 2017-04-18 2017-04-18 Stereopsis dense Stereo Matching method based on the optimization of global block

Country Status (1)

Country Link
CN (1) CN107170000B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110060283A (en) * 2019-04-17 2019-07-26 武汉大学 It is a kind of to estimate half global dense Stereo Matching algorithm more
CN112233246A (en) * 2020-09-24 2021-01-15 中山大学 Satellite image dense matching method and system based on SRTM constraint
CN112767421A (en) * 2021-01-15 2021-05-07 重庆大学 Stereo image dense matching method and system combining semantic information
CN113395504A (en) * 2021-03-31 2021-09-14 北京迈格威科技有限公司 Disparity map optimization method and device, electronic equipment and computer-readable storage medium
CN113989250A (en) * 2021-11-02 2022-01-28 中国测绘科学研究院 Improved block dense matching method, system, terminal and medium based on depth map
CN114049562A (en) * 2021-11-30 2022-02-15 中国科学院地理科学与资源研究所 Method for fusing and correcting land cover data

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105160702A (en) * 2015-08-20 2015-12-16 武汉大学 Stereoscopic image dense matching method and system based on LiDAR point cloud assistance
CN106530337A (en) * 2016-10-31 2017-03-22 武汉市工程科学技术研究院 Non local stereopair dense matching method based on image gray scale guiding

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105160702A (en) * 2015-08-20 2015-12-16 武汉大学 Stereoscopic image dense matching method and system based on LiDAR point cloud assistance
CN106530337A (en) * 2016-10-31 2017-03-22 武汉市工程科学技术研究院 Non local stereopair dense matching method based on image gray scale guiding

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
XU HUANG: "IMAGE-GUIDED NON-LOCAL DENSE MATCHING WITH THREE-STEPS OPTIMIZATION", 《ISPRS ANNALS OF THE PHOTOGRAMMETRY, REMOTE SENSING AND SPATIAL INFORMATION SCIENCES》 *

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110060283A (en) * 2019-04-17 2019-07-26 武汉大学 It is a kind of to estimate half global dense Stereo Matching algorithm more
CN110060283B (en) * 2019-04-17 2020-10-30 武汉大学 Multi-measure semi-global dense matching method
CN112233246A (en) * 2020-09-24 2021-01-15 中山大学 Satellite image dense matching method and system based on SRTM constraint
CN112233246B (en) * 2020-09-24 2024-04-16 中山大学 Satellite image dense matching method and system based on SRTM constraint
CN112767421A (en) * 2021-01-15 2021-05-07 重庆大学 Stereo image dense matching method and system combining semantic information
CN112767421B (en) * 2021-01-15 2023-09-15 重庆大学 Stereoscopic image dense matching method and system combining semantic information
CN113395504A (en) * 2021-03-31 2021-09-14 北京迈格威科技有限公司 Disparity map optimization method and device, electronic equipment and computer-readable storage medium
CN113395504B (en) * 2021-03-31 2023-04-04 北京迈格威科技有限公司 Disparity map optimization method and device, electronic equipment and computer-readable storage medium
CN113989250A (en) * 2021-11-02 2022-01-28 中国测绘科学研究院 Improved block dense matching method, system, terminal and medium based on depth map
CN113989250B (en) * 2021-11-02 2022-07-05 中国测绘科学研究院 Improved block dense matching method, system, terminal and medium based on depth map
CN114049562A (en) * 2021-11-30 2022-02-15 中国科学院地理科学与资源研究所 Method for fusing and correcting land cover data

Also Published As

Publication number Publication date
CN107170000B (en) 2019-09-10

Similar Documents

Publication Publication Date Title
CN107170000A (en) The stereopsis dense Stereo Matching method optimized based on global block
CN111462329B (en) Three-dimensional reconstruction method of unmanned aerial vehicle aerial image based on deep learning
CN103106688B (en) Based on the indoor method for reconstructing three-dimensional scene of double-deck method for registering
CN101383054B (en) Hybrid three-dimensional reconstructing method based on image and scanning data
CN107767413A (en) A kind of image depth estimation method based on convolutional neural networks
CN100533487C (en) Smooth symmetrical surface 3-D solid model rebuilding method based on single image
CN102254348B (en) Virtual viewpoint mapping method based o adaptive disparity estimation
CN103606151B (en) Based on the virtual geographical scene method for auto constructing on a large scale of imaging point cloud
CN104539928B (en) A kind of grating stereo printing image combining method
CN104065947B (en) The depth map acquisition methods of a kind of integration imaging system
CN102074014A (en) Stereo matching method by utilizing graph theory-based image segmentation algorithm
CN109598754A (en) A kind of binocular depth estimation method based on depth convolutional network
CN103702103B (en) Based on the grating stereo printing images synthetic method of binocular camera
CN104954780A (en) DIBR (depth image-based rendering) virtual image restoration method applicable to high-definition 2D/3D (two-dimensional/three-dimensional) conversion
CN102665086A (en) Method for obtaining parallax by using region-based local stereo matching
CN104517317A (en) Three-dimensional reconstruction method of vehicle-borne infrared images
CN108648264A (en) Underwater scene method for reconstructing based on exercise recovery and storage medium
CN103020963B (en) A kind of multi-eye stereo matching process cut based on the figure of self-adaptation watershed divide
CN114359509A (en) Multi-view natural scene reconstruction method based on deep learning
CN114119884A (en) Building LOD1 model construction method based on high-score seven-satellite image
CN107689060A (en) Visual processing method, device and the equipment of view-based access control model processing of destination object
CN103049929A (en) Multi-camera dynamic scene 3D (three-dimensional) rebuilding method based on joint optimization
CN109218706B (en) Method for generating stereoscopic vision image from single image
CN114255279A (en) Binocular vision three-dimensional reconstruction method based on high-precision positioning and deep learning
CN114998532B (en) Three-dimensional image visual transmission optimization method based on digital image reconstruction

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