CN107392950A - A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection - Google Patents

A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection Download PDF

Info

Publication number
CN107392950A
CN107392950A CN201710631310.7A CN201710631310A CN107392950A CN 107392950 A CN107392950 A CN 107392950A CN 201710631310 A CN201710631310 A CN 201710631310A CN 107392950 A CN107392950 A CN 107392950A
Authority
CN
China
Prior art keywords
mrow
msup
mtd
msub
munder
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201710631310.7A
Other languages
Chinese (zh)
Inventor
卢迪
张美玲
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.)
Harbin University of Science and Technology
Original Assignee
Harbin University of Science and Technology
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 Harbin University of Science and Technology filed Critical Harbin University of Science and Technology
Priority to CN201710631310.7A priority Critical patent/CN107392950A/en
Publication of CN107392950A publication Critical patent/CN107392950A/en
Pending legal-status Critical Current

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
    • G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
    • G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/90—Determination of colour characteristics
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10024—Color image

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Magnetic Resonance Imaging Apparatus (AREA)

Abstract

A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection of the present invention belongs to computer vision field, more particularly to the solid matching method to weak texture image, comprises the following steps：Two width coloured images are inputted, two described width coloured images are respectively left image and right image, and weak skin texture detection and segmentation are carried out to picture using the gradient information of left image；Matching power flow is calculated according to the colouring information and gradient information of left image and right image；On the basis of weak skin texture detection and segmentation result in above-mentioned, the interior yardstick based on gaussian filtering and the polymerization of across yardstick cost are carried out；Take policy calculation parallax entirely using the person of winning；Parallax is refined using left and right consistency detection and the method based on adaptive weighting, exports anaglyph.The present invention is realized on the premise of texture region matching accuracy is ensured, improves weak texture region matching accuracy, obtains the technical purpose of more preferable disparity map.

Description

A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection
Technical field
A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection of the present invention belongs to computer vision neck Domain, more particularly to the solid matching method to weak texture image.
Background technology
Binocular stereo vision (BinocularStereoVision) is a kind of important form of computer vision, and it is base In principle of parallax and two images of the imaging device from different position acquisition testees are utilized, by calculating image corresponding points Between position deviation, to obtain the method for object dimensional geological information.And the quality that three-dimensional information obtains depends primarily on solid The height of matching gained disparity map accuracy.The problem of Stereo matching is present at present mainly has uneven illumination, over-exposed etc. outer Boundary's factor, and picture can exist and block, weak texture, repeat picture itself feature that the computer such as texture is difficult to differentiate between in itself. Although a large amount of scholars to three-dimensional matching correct for many years, the matching for weak texture region is still the one of image processing field Individual difficult point.On the premise of how ensureing texture region matching accuracy, weak texture region matching accuracy is improved, is obtained more preferable Disparity map is a significant problem.
The content of the invention
The invention provides a kind of across yardstick cost based on weak skin texture detection to polymerize solid matching method, can ensure On the premise of texture region matching accuracy, weak texture region matching accuracy is improved, obtains more preferable disparity map.
The object of the present invention is achieved like this：
A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection, comprises the following steps：
Step a, two width coloured images are inputted, two described width coloured images are respectively left image and right image, utilize a left side The gradient information of image carries out weak skin texture detection and segmentation to picture；
Step b, Matching power flow is calculated according to the colouring information and gradient information of left image and right image；
Step c, on the basis of weak skin texture detection and segmentation result in step a, the interior yardstick based on gaussian filtering is carried out It polymerize with across yardstick cost；
Step d, policy calculation parallax is taken entirely using the person of winning；
Step e, the method using left and right consistency detection and based on adaptive weighting is refined to parallax, exports disparity map Picture.
Described a kind of across yardstick cost polymerization solid matching method based on weak skin texture detection, to figure in the step a Piece carries out weak skin texture detection and segmentation is specially：
It is that the Grad of (x, y) place pixel is g (x, y) to calculate left image coordinate, and with Grads threshold gTCompare, judge Whether it is weak texture region, its calculation formula is：
G (x, y) ＜ gT
In formula:N (x, y) represents the window centered on pixel (x, y), and M represents the number of pixel in window, I (x, y) table Show the gray value of pixel.
A kind of described across yardstick cost based on weak skin texture detection polymerize solid matching method, is calculated in the step b Matching power flow is specially：
Three-dimensional color image is calculated to left image ILWith right image IRMatching power flow C (p, d), its calculation formula is：
C (p, d)=(1- α) CAD(p,d)+α·(Cgrad_x(p,d)+Cgrad_y(p,d))
In formula：P is a bit in left image, and i=R in formula, G, B represent three passages of coloured image, T respectivelyADAnd Tgrad The interceptive value of color and gradient is represented respectively；Gradient operator of the picture in x, y direction is represented respectively；α is colour-difference Balance factor between gradient difference.
A kind of described across yardstick cost polymerization solid matching method based on weak skin texture detection, cost in the step c Polymerization is specially：
Wherein,The Matching power flow after polymerization is represented, z is desired optimization target values, and W is gaussian filtering core, and N is pixel P neighborhood window, q are p neighborhood territory pixel points；S ∈ { 0,1 ..., S } are scale parameter, during s=0, C0The original chi of representative image Spend Matching power flow；The polymerization cost of S+1 yardstick of representative image；
In formula, λ is regularization factors,WithRepresent The optimization object function of formula (11), orderHave：
Wherein, ThighAnd TlowThe texture region hereinbefore detected and weak texture region are represented respectively；C1And C1/2Respectively The Matching power flow of original image yardstick and half yardstick is represented, gaussian filtering is carried out with different size of window, after fusion To final Matching power flow.
A kind of described across yardstick cost polymerization solid matching method based on weak skin texture detection, parallax in the step e Refine specially：
|D'L(P)-D'R(P-D'R(P)) | ＜ δ
DLRC(P)=min (D'(PL), D'(PR))
Wherein, the left figure parallax value D' of the point p in disparity mapLAnd right figure parallax value D' (p)R(p-D'L(p)), δ LRC Threshold value；D ' (PL) is the parallax value of first unshielding point in left side, and D (PR) is the parallax of first unshielding point on right side Value；WBpq(IL) be left image function, Δ cpqWith Δ spqRespectively left image midpoint p and q heterochromia and space are European Distance,WithThe respectively adjustment parameter of heterochromia and distance difference；Dw(p) filtered image.
Beneficial effect：
A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection of the present invention, the present invention proposes a kind of new Stereo Matching Algorithm, whether belong to weak texture for picture region, choose appropriate matching process, so as to improve Stereo matching Accuracy, obtain more preferable disparity map.
Using the Stereo matching image pair of present embodiment algorithm process, in the texture region of picture and weak texture region all Preferable effect can be obtained, erroneous matching rate decreases (than without low 5%) of weak texture region partitioning algorithm.Illustrate this Embodiment algorithm can improve weak texture region matching accuracy, obtain on the premise of texture region matching accuracy is ensured Obtain more preferable disparity map.
Brief description of the drawings
Fig. 1 is a kind of across yardstick cost polymerization solid matching method flow chart based on weak skin texture detection.
Fig. 2 is Bowling1 disparity maps.
Fig. 3 is Lampshade1 disparity maps.
Fig. 4 is Monopoly disparity maps.
Fig. 5 is Plastic disparity maps.
Embodiment
The specific embodiment of the invention is described in further detail below in conjunction with the accompanying drawings.
Specific embodiment one
A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection, as shown in figure 1, including following step Suddenly：
Step a, two width coloured images are inputted, two described width coloured images are respectively left image and right image, utilize a left side The gradient information of image carries out weak skin texture detection and segmentation to picture；
Step b, Matching power flow is calculated according to the colouring information and gradient information of left image and right image；
Step c, on the basis of weak skin texture detection and segmentation result in step a, the interior yardstick based on gaussian filtering is carried out It polymerize with across yardstick cost；
Step d, policy calculation parallax is taken entirely using the person of winning；
Step e, the method using left and right consistency detection and based on adaptive weighting is refined to parallax, exports disparity map Picture.
According to above step, four pictures are selected to be contrasted, as shown in Fig. 2, Fig. 3, Fig. 4 and Fig. 5,
In Fig. 2, Fig. 2 (a) is Bowling1 original image left figures；Fig. 2 (b) is the true disparity maps of Bowling1；Fig. 2 (c) is The weak skin texture detection results of Bowling1；Fig. 2 (d) is Bowling1 final parallax；Fig. 2 (e) is Bowling1 without weak line Manage the disparity map of detection.
In Fig. 3, Fig. 3 (a) is Lampshade1 original image left figures；Fig. 3 (b) is the true disparity maps of Lampshade1；Fig. 3 (c) it is the weak skin texture detection results of Lampshade1；Fig. 3 (d) is Lampshade1 final parallax；Fig. 3 (e) is Lampshade1 Without the disparity map of weak skin texture detection.
In Fig. 4, Fig. 4 (a) is Monopoly original image left figures；Fig. 4 (b) is the true disparity maps of Monopoly；Fig. 4 (c) is The weak skin texture detection results of Monopoly；Fig. 4 (d) is Monopoly final parallax；Fig. 4 (e) is Monopoly without weak line Manage the disparity map of detection.
In Fig. 5, Fig. 5 (a) is Plastic original image left figures；Fig. 5 (b) is the true disparity maps of Plastic；Fig. 5 (c) is The weak skin texture detection results of Plastic；Fig. 5 (d) is Plastic final parallax；Fig. 5 (e) is that Plastic examines without weak texture The disparity map of survey.
To Fig. 2 (a)~Fig. 2 (e), Fig. 3 (a)~Fig. 3 (e), Fig. 4 (a)~Fig. 4 (e), Fig. 5 (a)~figure from visual effect Disparity map in 5 (e) carries out subjective assessment.Black portions represent the weak texture region detected, white in Fig. 2~Fig. 5 (c) Part represents texture region.Compare disparity map, it can be seen that in weak texture region, the parallax obtained using present embodiment algorithm Figure result is more far better than without the algorithm disparity map effect of weak skin texture detection.
The inventive method is evaluated from objective evaluation index.
Table 1 is given using the weak obvious image pair of texture region of two kinds of algorithm process middlebury image sets 4 Erroneous matching rate.
Table 1
As can be seen from Table 1, in the test result of two kinds of algorithm process Stereo matching images pair, using present embodiment The image comparison of algorithm process reduces 5% without weak skin texture detection and the algorithmic error matching rate of segmentation.Illustrate this implementation Mode algorithm can improve weak texture region matching accuracy, obtain more on the premise of texture region matching accuracy is ensured Good disparity map.
Specific embodiment two
Described a kind of across yardstick cost polymerization solid matching method based on weak skin texture detection, to figure in the step a Piece carries out weak skin texture detection and segmentation is specially：
It is that the Grad of (x, y) place pixel is g (x, y) to calculate left image coordinate, and with Grads threshold gTCompare, judge Whether it is weak texture region, its calculation formula is：
G (x, y) ＜ gT
In formula:N (x, y) represents the window centered on pixel (x, y), and M represents the number of pixel in window, I (x, y) table Show the gray value of pixel.
A kind of described across yardstick cost based on weak skin texture detection polymerize solid matching method, is calculated in the step b Matching power flow is specially：
Three-dimensional color image is calculated to left image ILWith right image IRMatching power flow C (p, d), its calculation formula is：
C (p, d)=(1- α) CAD(p,d)+α·(Cgrad_x(p,d)+Cgrad_y(p,d))
In formula：P is a bit in left image, and i=R in formula, G, B represent three passages of coloured image, T respectivelyADAnd Tgrad The interceptive value of color and gradient is represented respectively；Gradient operator of the picture in x, y direction is represented respectively；α is colour-difference Balance factor between gradient difference.
A kind of described across yardstick cost polymerization solid matching method based on weak skin texture detection, cost in the step c Polymerization is specially：
Wherein,The Matching power flow after polymerization is represented, z is desired optimization target values, and W is gaussian filtering core, and N is pixel P neighborhood window, q are p neighborhood territory pixel points；S ∈ { 0,1 ..., S } are scale parameter, during s=0, C0The original chi of representative image Spend Matching power flow；The polymerization cost of S+1 yardstick of representative image；
In formula, λ is regularization factors,WithRepresent public The optimization object function of formula (11), orderHave：
Wherein, ThighAnd TlowThe texture region hereinbefore detected and weak texture region are represented respectively；C1And C1/2Respectively The Matching power flow of original image yardstick and half yardstick is represented, gaussian filtering is carried out with different size of window, after fusion To final Matching power flow.
A kind of described across yardstick cost polymerization solid matching method based on weak skin texture detection, parallax in the step e Refine specially：
|D'L(P)-D'R(P-D'R(P)) | ＜ δ
DLRC(P)=min (D'(PL), D'(PR))
Wherein, the left figure parallax value D' of the point p in disparity mapLAnd right figure parallax value D' (p)R(p-D'L(p)), δ LRC Threshold value；D ' (PL) is the parallax value of first unshielding point in left side, and D ' (PR) is regarding for first unshielding point on right side Difference；WBpq(IL) be left image function, Δ cpqWith Δ spqRespectively left image midpoint p and q heterochromia and space Europe Formula distance,WithThe respectively adjustment parameter of heterochromia and distance difference；Dw(p) filtered image.

Claims (5)

1. a kind of across yardstick cost polymerization solid matching method based on weak skin texture detection, its sign is, comprises the following steps：
Step a, two width coloured images are inputted, two described width coloured images are respectively left image and right image, utilize left image Gradient information weak skin texture detection and segmentation are carried out to picture；
Step b, Matching power flow is calculated according to the colouring information and gradient information of left image and right image；
Step c, on the basis of weak skin texture detection and segmentation result in step a, carry out interior yardstick based on gaussian filtering and across Yardstick cost polymerize；
Step d, policy calculation parallax is taken entirely using the person of winning；
Step e, the method using left and right consistency detection and based on adaptive weighting is refined to parallax, exports anaglyph.
2. a kind of across yardstick cost polymerization solid matching method based on weak skin texture detection according to claim 1, it is special Sign is, carries out weak skin texture detection to picture in the step a and segmentation is specially：
It is that the Grad of (x, y) place pixel is g (x, y) to calculate left image coordinate, and with Grads threshold gTCompare, judge whether For weak texture region, its calculation formula is：
G (x, y) ＜ gT
<mrow> <mi>g</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mn>1</mn> <mi>M</mi> </mfrac> <mo>&amp;CenterDot;</mo> <munder> <mo>&amp;Sigma;</mo> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> <mo>&amp;Element;</mo> <mi>N</mi> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> </munder> <mrow> <mo>(</mo> <mo>|</mo> <mi>I</mi> <mo>(</mo> <mrow> <mi>u</mi> <mo>,</mo> <mi>v</mi> </mrow> <mo>)</mo> <mo>-</mo> <mi>I</mi> <mo>(</mo> <mrow> <mi>u</mi> <mo>+</mo> <mn>1</mn> <mo>,</mo> <mi>v</mi> </mrow> <mo>)</mo> <mo>+</mo> <mo>|</mo> <mi>I</mi> <mo>(</mo> <mrow> <mi>u</mi> <mo>,</mo> <mi>v</mi> </mrow> <mo>)</mo> <mo>-</mo> <mi>I</mi> <mo>(</mo> <mrow> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>+</mo> <mn>1</mn> </mrow> <mo>)</mo> <mo>|</mo> <mo>)</mo> </mrow> </mrow>
In formula:N (x, y) represents the window centered on pixel (x, y), and M represents the number of pixel in window, and I (x, y) represents picture The gray value of element.
3. a kind of across yardstick cost polymerization solid matching method based on weak skin texture detection according to claim 1, it is special Sign is that Matching power flow is calculated in the step b is specially：
Three-dimensional color image is calculated to left image ILWith right image IRMatching power flow C (p, d), its calculation formula is：
C (p, d)=(1- α) CAD(p,d)+α·(Cgrad_x(p,d)+Cgrad_y(p,d))
<mrow> <msub> <mi>C</mi> <mrow> <mi>A</mi> <mi>D</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>d</mi> <mo>)</mo> </mrow> <mo>=</mo> <mi>m</mi> <mi>i</mi> <mi>n</mi> <mrow> <mo>(</mo> <mfrac> <mn>1</mn> <mn>3</mn> </mfrac> <munder> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mi>R</mi> <mo>,</mo> <mi>G</mi> <mo>,</mo> <mi>B</mi> </mrow> </munder> <mo>|</mo> <msubsup> <mi>I</mi> <mi>L</mi> <mi>i</mi> </msubsup> <mo>(</mo> <mi>p</mi> <mo>)</mo> <mo>-</mo> <msubsup> <mi>I</mi> <mi>R</mi> <mi>i</mi> </msubsup> <mo>(</mo> <mrow> <mi>p</mi> <mo>,</mo> <mi>d</mi> </mrow> <mo>)</mo> <mo>|</mo> <mo>,</mo> <msub> <mi>T</mi> <mrow> <mi>A</mi> <mi>D</mi> </mrow> </msub> <mo>)</mo> </mrow> </mrow>
Cgrad_x(p, d)=min (| ▽xIL(p)-▽xIR(p,d)|,Tgrad)
Cgrad_y(p, d)=min (| ▽yIL(p)-▽yIR(p,d)|,Tgrad)
In formula：P is a bit in left image, and i=R in formula, G, B represent three passages of coloured image, T respectivelyADAnd TgradRespectively Represent the interceptive value of color and gradient；▽x、▽yGradient operator of the picture in x, y direction is represented respectively；α is colour-difference and ladder Balance factor between degree difference.
4. a kind of across yardstick cost polymerization solid matching method based on weak skin texture detection according to claim 1, it is special Sign is that cost, which polymerize, in the step c is specially：
<mrow> <mover> <mi>C</mi> <mo>~</mo> </mover> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>d</mi> <mo>)</mo> </mrow> <mo>=</mo> <munder> <mi>argmin</mi> <mi>z</mi> </munder> <munder> <mo>&amp;Sigma;</mo> <mrow> <mi>q</mi> <mo>&amp;Element;</mo> <mi>N</mi> </mrow> </munder> <mi>W</mi> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <mo>|</mo> <mo>|</mo> <mi>z</mi> <mo>-</mo> <mi>C</mi> <mrow> <mo>(</mo> <mi>q</mi> <mo>,</mo> <mi>d</mi> <mo>)</mo> </mrow> <mo>|</mo> <msup> <mo>|</mo> <mn>2</mn> </msup> </mrow>
<mrow> <mover> <mi>C</mi> <mo>~</mo> </mover> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>d</mi> <mo>)</mo> </mrow> <mo>=</mo> <munder> <mo>&amp;Sigma;</mo> <mrow> <mi>q</mi> <mo>&amp;Element;</mo> <mi>N</mi> </mrow> </munder> <mi>W</mi> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <mo>&amp;CenterDot;</mo> <mi>C</mi> <mrow> <mo>(</mo> <mi>q</mi> <mo>,</mo> <mi>d</mi> <mo>)</mo> </mrow> </mrow>
<mrow> <mover> <mi>v</mi> <mo>~</mo> </mover> <mo>=</mo> <munder> <mrow> <mi>a</mi> <mi>r</mi> <mi>g</mi> <mi>m</mi> <mi>i</mi> <mi>n</mi> </mrow> <msubsup> <mrow> <mo>{</mo> <msup> <mi>z</mi> <mi>s</mi> </msup> <mo>}</mo> </mrow> <mrow> <mi>s</mi> <mo>=</mo> <mn>0</mn> </mrow> <mi>S</mi> </msubsup> </munder> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>s</mi> <mo>=</mo> <mn>0</mn> </mrow> <mi>S</mi> </munderover> <munder> <mo>&amp;Sigma;</mo> <mrow> <msup> <mi>q</mi> <mi>s</mi> </msup> <mo>&amp;Element;</mo> <msub> <mi>N</mi> <mi>s</mi> </msub> </mrow> </munder> <mi>W</mi> <mrow> <mo>(</mo> <msup> <mi>p</mi> <mi>s</mi> </msup> <mo>,</mo> <msup> <mi>q</mi> <mi>s</mi> </msup> <mo>)</mo> </mrow> <mo>|</mo> <mo>|</mo> <msup> <mi>z</mi> <mi>s</mi> </msup> <mo>-</mo> <msup> <mi>C</mi> <mi>s</mi> </msup> <mrow> <mo>(</mo> <msup> <mi>q</mi> <mi>s</mi> </msup> <mo>,</mo> <msup> <mi>d</mi> <mi>s</mi> </msup> <mo>)</mo> </mrow> <mo>|</mo> <msup> <mo>|</mo> <mn>2</mn> </msup> </mrow>
Wherein,The Matching power flow after polymerization is represented, z is desired optimization target values, and W is gaussian filtering core, and N is the neighbour of pixel p Domain window, q are p neighborhood territory pixel points；S ∈ { 0,1 ..., S } are scale parameter, during s=0, C0Representative image original scale With cost；The polymerization cost of S+1 yardstick of representative image；
<mrow> <msup> <mover> <mi>C</mi> <mo>~</mo> </mover> <mi>s</mi> </msup> <mrow> <mo>(</mo> <msup> <mi>p</mi> <mi>s</mi> </msup> <mo>,</mo> <msup> <mi>d</mi> <mi>s</mi> </msup> <mo>)</mo> </mrow> <mo>=</mo> <munder> <mo>&amp;Sigma;</mo> <mrow> <msup> <mi>q</mi> <mi>s</mi> </msup> <mo>&amp;Element;</mo> <msub> <mi>N</mi> <mi>s</mi> </msub> </mrow> </munder> <mi>W</mi> <mrow> <mo>(</mo> <msup> <mi>p</mi> <mi>s</mi> </msup> <mo>,</mo> <msup> <mi>q</mi> <mi>s</mi> </msup> <mo>)</mo> </mrow> <mo>&amp;CenterDot;</mo> <msup> <mi>C</mi> <mi>s</mi> </msup> <mrow> <mo>(</mo> <msup> <mi>q</mi> <mi>s</mi> </msup> <mo>,</mo> <msup> <mi>d</mi> <mi>s</mi> </msup> <mo>)</mo> </mrow> </mrow>
<mrow> <mover> <mi>v</mi> <mo>~</mo> </mover> <mo>=</mo> <munder> <mrow> <mi>a</mi> <mi>r</mi> <mi>g</mi> <mi>m</mi> <mi>i</mi> <mi>n</mi> </mrow> <msubsup> <mrow> <mo>{</mo> <msup> <mi>z</mi> <mi>s</mi> </msup> <mo>}</mo> </mrow> <mrow> <mi>s</mi> <mo>=</mo> <mn>0</mn> </mrow> <mi>S</mi> </msubsup> </munder> <mrow> <mo>(</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>s</mi> <mo>=</mo> <mn>0</mn> </mrow> <mi>S</mi> </munderover> <munder> <mo>&amp;Sigma;</mo> <mrow> <msup> <mi>q</mi> <mi>s</mi> </msup> <mo>&amp;Element;</mo> <msub> <mi>N</mi> <mi>s</mi> </msub> </mrow> </munder> <mi>W</mi> <mo>(</mo> <mrow> <msup> <mi>p</mi> <mi>s</mi> </msup> <mo>,</mo> <msup> <mi>q</mi> <mi>s</mi> </msup> </mrow> <mo>)</mo> <mo>|</mo> <mo>|</mo> <msup> <mi>z</mi> <mi>s</mi> </msup> <mo>-</mo> <msup> <mi>C</mi> <mi>s</mi> </msup> <mo>(</mo> <mrow> <msup> <mi>q</mi> <mi>s</mi> </msup> <mo>,</mo> <msup> <mi>d</mi> <mi>s</mi> </msup> </mrow> <mo>)</mo> <mo>|</mo> <msup> <mo>|</mo> <mn>2</mn> </msup> <mo>+</mo> <mi>&amp;lambda;</mi> <mo>&amp;CenterDot;</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>s</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>S</mi> </munderover> <mo>|</mo> <mo>|</mo> <msup> <mi>z</mi> <mi>s</mi> </msup> <mo>-</mo> <msup> <mi>z</mi> <mrow> <mi>s</mi> <mo>-</mo> <mn>1</mn> </mrow> </msup> <mo>|</mo> <msup> <mo>|</mo> <mn>2</mn> </msup> <mo>)</mo> </mrow> </mrow>
In formula, λ is regularization factors,WithRepresentation formula (11) optimization object function, orderHave：
<mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <mo>(</mo> <mn>1</mn> <mo>+</mo> <mi>&amp;lambda;</mi> <mo>)</mo> <msup> <mi>z</mi> <mi>s</mi> </msup> <mo>-</mo> <msup> <mi>&amp;lambda;z</mi> <mrow> <mi>s</mi> <mo>+</mo> <mn>1</mn> </mrow> </msup> <mo>=</mo> <msup> <mover> <mi>C</mi> <mo>~</mo> </mover> <mi>s</mi> </msup> <mo>(</mo> <msup> <mi>p</mi> <mi>s</mi> </msup> <mo>,</mo> <msup> <mi>d</mi> <mi>s</mi> </msup> <mo>)</mo> <mo>,</mo> <mi>s</mi> <mo>=</mo> <mn>0</mn> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>-</mo> <msup> <mi>&amp;lambda;z</mi> <mrow> <mi>s</mi> <mo>-</mo> <mn>1</mn> </mrow> </msup> <mo>+</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>+</mo> <mn>2</mn> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <msup> <mi>z</mi> <mi>s</mi> </msup> <mo>-</mo> <msup> <mi>&amp;lambda;z</mi> <mrow> <mi>s</mi> <mo>+</mo> <mn>1</mn> </mrow> </msup> <mo>=</mo> <msup> <mover> <mi>C</mi> <mo>~</mo> </mover> <mi>s</mi> </msup> <mrow> <mo>(</mo> <msup> <mi>p</mi> <mi>s</mi> </msup> <mo>,</mo> <msup> <mi>d</mi> <mi>s</mi> </msup> <mo>)</mo> </mrow> <mo>,</mo> <mi>s</mi> <mo>=</mo> <mn>1</mn> <mo>,</mo> <mn>2</mn> <mo>,</mo> <mo>...</mo> <mo>,</mo> <mi>S</mi> <mo>-</mo> <mn>1</mn> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>-</mo> <msup> <mi>&amp;lambda;z</mi> <mrow> <mi>s</mi> <mo>-</mo> <mn>1</mn> </mrow> </msup> <mo>+</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>+</mo> <mn>2</mn> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <msup> <mi>z</mi> <mi>s</mi> </msup> <mo>=</mo> <msup> <mover> <mi>C</mi> <mo>~</mo> </mover> <mi>s</mi> </msup> <mrow> <mo>(</mo> <msup> <mi>p</mi> <mi>s</mi> </msup> <mo>,</mo> <msup> <mi>d</mi> <mi>s</mi> </msup> <mo>)</mo> </mrow> <mo>,</mo> <mi>s</mi> <mo>=</mo> <mi>S</mi> </mrow> </mtd> </mtr> </mtable> </mfenced>
<mrow> <mi>A</mi> <mover> <mi>v</mi> <mo>^</mo> </mover> <mo>=</mo> <mover> <mi>v</mi> <mo>~</mo> </mover> </mrow>
<mrow> <mi>A</mi> <mo>=</mo> <mfenced open = "[" close = "]"> <mtable> <mtr> <mtd> <mrow> <mn>1</mn> <mo>+</mo> <mi>&amp;lambda;</mi> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mi>&amp;lambda;</mi> </mrow> </mtd> <mtd> <mrow></mrow> </mtd> <mtd> <mrow></mrow> </mtd> <mtd> <mrow></mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>-</mo> <mi>&amp;lambda;</mi> </mrow> </mtd> <mtd> <mrow> <mn>1</mn> <mo>+</mo> <mn>2</mn> <mi>&amp;lambda;</mi> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mi>&amp;lambda;</mi> </mrow> </mtd> <mtd> <mrow></mrow> </mtd> <mtd> <mrow></mrow> </mtd> </mtr> <mtr> <mtd> <mrow></mrow> </mtd> <mtd> <mo>...</mo> </mtd> <mtd> <mo>...</mo> </mtd> <mtd> <mrow></mrow> </mtd> <mtd> <mrow></mrow> </mtd> </mtr> <mtr> <mtd> <mrow></mrow> </mtd> <mtd> <mrow></mrow> </mtd> <mtd> <mo>...</mo> </mtd> <mtd> <mo>...</mo> </mtd> <mtd> <mrow></mrow> </mtd> </mtr> <mtr> <mtd> <mrow></mrow> </mtd> <mtd> <mrow></mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mi>&amp;lambda;</mi> </mrow> </mtd> <mtd> <mrow> <mn>1</mn> <mo>+</mo> <mn>2</mn> <mi>&amp;lambda;</mi> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mi>&amp;lambda;</mi> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow></mrow> </mtd> <mtd> <mrow></mrow> </mtd> <mtd> <mrow></mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mi>&amp;lambda;</mi> </mrow> </mtd> <mtd> <mrow> <mn>1</mn> <mo>+</mo> <mi>&amp;lambda;</mi> </mrow> </mtd> </mtr> </mtable> </mfenced> </mrow>
<mrow> <mover> <mi>v</mi> <mo>^</mo> </mover> <mo>=</mo> <msup> <mi>A</mi> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msup> <mover> <mi>v</mi> <mo>~</mo> </mover> </mrow>
<mrow> <msub> <mover> <mi>C</mi> <mo>~</mo> </mover> <mrow> <mi>f</mi> <mi>i</mi> <mi>n</mi> <mi>a</mi> <mi>l</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>d</mi> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mi>T</mi> <mrow> <mi>h</mi> <mi>i</mi> <mi>g</mi> <mi>h</mi> </mrow> </msub> <mo>&amp;CenterDot;</mo> <munder> <mi>&amp;Sigma;</mi> <mrow> <mi>q</mi> <mo>&amp;Element;</mo> <mi>N</mi> </mrow> </munder> <msub> <mi>W</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <msub> <mi>C</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>q</mi> <mo>,</mo> <mi>d</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>T</mi> <mrow> <mi>l</mi> <mi>o</mi> <mi>w</mi> </mrow> </msub> <mo>&amp;CenterDot;</mo> <munder> <mi>&amp;Sigma;</mi> <mrow> <mi>q</mi> <mo>&amp;Element;</mo> <mi>N</mi> </mrow> </munder> <msub> <mi>W</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <msub> <mi>C</mi> <mrow> <mn>1</mn> <mo>/</mo> <mn>2</mn> </mrow> </msub> <mrow> <mo>(</mo> <mi>q</mi> <mo>,</mo> <mi>d</mi> <mo>)</mo> </mrow> </mrow>
Wherein, ThighAnd TlowThe texture region hereinbefore detected and weak texture region are represented respectively；C1And C1/2Represent respectively The Matching power flow of original image yardstick and half yardstick, gaussian filtering is carried out with different size of window, is obtained most after fusion Whole Matching power flow.
5. a kind of across yardstick cost polymerization solid matching method based on weak skin texture detection according to claim 1, it is special Sign is that parallax is refined specially in the step e：
|D'L(P)-D'R(P-D'R(P)) | ＜ δ
DLRC(P)=min (D'(PL), D'(PR))
<mrow> <msub> <mi>D</mi> <mi>w</mi> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>)</mo> </mrow> <mo>=</mo> <munder> <mo>&amp;Sigma;</mo> <mi>q</mi> </munder> <msub> <mi>WB</mi> <mrow> <mi>p</mi> <mi>q</mi> </mrow> </msub> <mrow> <mo>(</mo> <msub> <mi>I</mi> <mi>L</mi> </msub> <mo>)</mo> </mrow> <msub> <mi>D</mi> <mrow> <mi>L</mi> <mi>R</mi> <mi>C</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>q</mi> <mo>)</mo> </mrow> </mrow>
<mrow> <msub> <mi>WB</mi> <mrow> <mi>p</mi> <mi>q</mi> </mrow> </msub> <mo>=</mo> <mi>exp</mi> <mrow> <mo>(</mo> <mo>-</mo> <mfrac> <mrow> <msub> <mi>&amp;Delta;c</mi> <mrow> <mi>p</mi> <mi>q</mi> </mrow> </msub> </mrow> <msubsup> <mi>&amp;sigma;</mi> <mi>c</mi> <mn>2</mn> </msubsup> </mfrac> <mo>+</mo> <mfrac> <mrow> <msub> <mi>&amp;Delta;s</mi> <mrow> <mi>p</mi> <mi>q</mi> </mrow> </msub> </mrow> <msubsup> <mi>&amp;sigma;</mi> <mi>s</mi> <mn>2</mn> </msubsup> </mfrac> <mo>)</mo> </mrow> </mrow>
Wherein, the left figure parallax value D' of the point p in disparity mapLAnd right figure parallax value D' (p)R(p-D'L(p)), δ is LRC threshold Value；D ' (PL) is the parallax value of first unshielding point in left side, and D ' (PR) is the parallax value of first unshielding point on right side； WBpq(IL) be left image function, Δ cpqWith Δ spqRespectively left image midpoint p and q heterochromia and space it is European away from From,WithThe respectively adjustment parameter of heterochromia and distance difference；Dw(p) filtered image.
CN201710631310.7A 2017-07-28 2017-07-28 A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection Pending CN107392950A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710631310.7A CN107392950A (en) 2017-07-28 2017-07-28 A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710631310.7A CN107392950A (en) 2017-07-28 2017-07-28 A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection

Publications (1)

Publication Number Publication Date
CN107392950A true CN107392950A (en) 2017-11-24

Family

ID=60342086

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710631310.7A Pending CN107392950A (en) 2017-07-28 2017-07-28 A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection

Country Status (1)

Country Link
CN (1) CN107392950A (en)

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107945222A (en) * 2017-12-15 2018-04-20 东南大学 A kind of new Stereo matching cost calculates and parallax post-processing approach
CN108181319A (en) * 2017-12-12 2018-06-19 陕西三星洁净工程有限公司 A kind of laying dust detecting device and method based on stereoscopic vision
CN108510529A (en) * 2018-03-14 2018-09-07 昆明理工大学 A kind of figure based on adaptive weight cuts solid matching method
CN108596975A (en) * 2018-04-25 2018-09-28 华南理工大学 A kind of Stereo Matching Algorithm for weak texture region
CN108682026A (en) * 2018-03-22 2018-10-19 辽宁工业大学 A kind of binocular vision solid matching method based on the fusion of more Matching units
CN108765486A (en) * 2018-05-17 2018-11-06 长春理工大学 Based on sparse piece of aggregation strategy method of relevant Stereo matching in color
CN109816782A (en) * 2019-02-03 2019-05-28 哈尔滨理工大学 A kind of indoor scene three-dimensional rebuilding method based on binocular vision
CN109887021A (en) * 2019-01-19 2019-06-14 天津大学 Based on the random walk solid matching method across scale
CN109961417A (en) * 2017-12-26 2019-07-02 广州极飞科技有限公司 Image processing method, device and mobile device control method
CN111191694A (en) * 2019-12-19 2020-05-22 浙江科技学院 Image stereo matching method
CN111508013A (en) * 2020-04-21 2020-08-07 中国科学技术大学 Stereo matching method
CN112070694A (en) * 2020-09-03 2020-12-11 深兰人工智能芯片研究院（江苏）有限公司 Binocular stereo vision disparity map post-processing method and device
WO2021018093A1 (en) * 2019-07-31 2021-02-04 深圳市道通智能航空技术有限公司 Stereo matching method, image processing chip, and moving carrier

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105551035A (en) * 2015-12-09 2016-05-04 深圳市华和瑞智科技有限公司 Stereoscopic vision matching method based on weak edge and texture classification
CN106340036A (en) * 2016-08-08 2017-01-18 东南大学 Binocular stereoscopic vision-based stereo matching method
CN106530336A (en) * 2016-11-07 2017-03-22 湖南源信光电科技有限公司 Stereo matching algorithm based on color information and graph-cut theory

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105551035A (en) * 2015-12-09 2016-05-04 深圳市华和瑞智科技有限公司 Stereoscopic vision matching method based on weak edge and texture classification
CN106340036A (en) * 2016-08-08 2017-01-18 东南大学 Binocular stereoscopic vision-based stereo matching method
CN106530336A (en) * 2016-11-07 2017-03-22 湖南源信光电科技有限公司 Stereo matching algorithm based on color information and graph-cut theory

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
张华 等: ""基于跨尺度变窗口代价聚合的快速立体匹配"", 《计算机工程与应用》 *
曹晓倩 等: "基于弱纹理检测及视差图融合的立体匹配", 《仪器仪表学报》 *
林雪: "双目立体视觉中立体匹配技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *

Cited By (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108181319A (en) * 2017-12-12 2018-06-19 陕西三星洁净工程有限公司 A kind of laying dust detecting device and method based on stereoscopic vision
CN108181319B (en) * 2017-12-12 2020-09-11 陕西三星洁净工程有限公司 Accumulated dust detection device and method based on stereoscopic vision
CN107945222A (en) * 2017-12-15 2018-04-20 东南大学 A kind of new Stereo matching cost calculates and parallax post-processing approach
CN109961417A (en) * 2017-12-26 2019-07-02 广州极飞科技有限公司 Image processing method, device and mobile device control method
CN108510529A (en) * 2018-03-14 2018-09-07 昆明理工大学 A kind of figure based on adaptive weight cuts solid matching method
CN108682026A (en) * 2018-03-22 2018-10-19 辽宁工业大学 A kind of binocular vision solid matching method based on the fusion of more Matching units
CN108682026B (en) * 2018-03-22 2021-08-06 江大白 Binocular vision stereo matching method based on multi-matching element fusion
CN108596975A (en) * 2018-04-25 2018-09-28 华南理工大学 A kind of Stereo Matching Algorithm for weak texture region
CN108596975B (en) * 2018-04-25 2022-03-29 华南理工大学 Stereo matching algorithm for weak texture region
CN108765486A (en) * 2018-05-17 2018-11-06 长春理工大学 Based on sparse piece of aggregation strategy method of relevant Stereo matching in color
CN109887021A (en) * 2019-01-19 2019-06-14 天津大学 Based on the random walk solid matching method across scale
CN109887021B (en) * 2019-01-19 2023-06-06 天津大学 Cross-scale-based random walk stereo matching method
CN109816782A (en) * 2019-02-03 2019-05-28 哈尔滨理工大学 A kind of indoor scene three-dimensional rebuilding method based on binocular vision
WO2021018093A1 (en) * 2019-07-31 2021-02-04 深圳市道通智能航空技术有限公司 Stereo matching method, image processing chip, and moving carrier
CN111191694A (en) * 2019-12-19 2020-05-22 浙江科技学院 Image stereo matching method
CN111508013A (en) * 2020-04-21 2020-08-07 中国科学技术大学 Stereo matching method
CN111508013B (en) * 2020-04-21 2022-09-06 中国科学技术大学 Stereo matching method
CN112070694A (en) * 2020-09-03 2020-12-11 深兰人工智能芯片研究院（江苏）有限公司 Binocular stereo vision disparity map post-processing method and device

Similar Documents

Publication Publication Date Title
CN107392950A (en) A kind of across yardstick cost polymerization solid matching method based on weak skin texture detection
CN108596975A (en) A kind of Stereo Matching Algorithm for weak texture region
CN104680496B (en) A kind of Kinect depth map restorative procedures based on color images
CN110473217A (en) A kind of binocular solid matching process based on Census transformation
CN105744256B (en) Based on the significant objective evaluation method for quality of stereo images of collection of illustrative plates vision
CN107578418A (en) A kind of indoor scene profile testing method of confluent colours and depth information
CN110148181A (en) A kind of general binocular solid matching process
CN107578404A (en) The complete of view-based access control model notable feature extraction refers to objective evaluation method for quality of stereo images
CN107301642B (en) A kind of full-automatic prospect background segregation method based on binocular vision
CN106504276A (en) The combinations matches cost algorithms of non local Stereo Matching Algorithm and parallax joint filling algorithm
CN107301664A (en) Improvement sectional perspective matching process based on similarity measure function
CN104240204B (en) Solar silicon wafer and battery piece counting method based on image processing
CN107578430A (en) A kind of solid matching method based on adaptive weight and local entropy
CN104376535A (en) Rapid image repairing method based on sample
CN101866497A (en) Binocular stereo vision based intelligent three-dimensional human face rebuilding method and system
CN106408513A (en) Super-resolution reconstruction method of depth map
CN103384343B (en) A kind of method and device thereof filling up image cavity
CN102800111B (en) Color harmony based color fusion image color quality evaluation method
CN103325120A (en) Rapid self-adaption binocular vision stereo matching method capable of supporting weight
CN106600632A (en) Improved matching cost aggregation stereo matching algorithm
CN106530334B (en) A kind of airborne Interference synthetic aperture radar complex image corregistration method and complex image corregistration system
CN104102928B (en) A kind of Classifying Method in Remote Sensing Image based on texture primitive
CN103177451A (en) Three-dimensional matching algorithm between adaptive window and weight based on picture edge
CN103955945A (en) Self-adaption color image segmentation method based on binocular parallax and movable outline
CN107578399A (en) Full reference image quality appraisement method based on boundary characteristic segmentation

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
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20171124