CN107680083A - Parallax determines method and parallax determining device - Google Patents
Parallax determines method and parallax determining device Download PDFInfo
- Publication number
- CN107680083A CN107680083A CN201710823614.3A CN201710823614A CN107680083A CN 107680083 A CN107680083 A CN 107680083A CN 201710823614 A CN201710823614 A CN 201710823614A CN 107680083 A CN107680083 A CN 107680083A
- Authority
- CN
- China
- Prior art keywords
- pixel
- parallax
- image
- edge
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/97—Determining parameters from multiple pictures
-
- 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
- G06T2207/10004—Still image; Photographic image
- G06T2207/10012—Stereo images
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Measurement Of Optical Distance (AREA)
- Length Measuring Devices By Optical Means (AREA)
- Image Processing (AREA)
Abstract
The present invention is to determine method and device on parallax, and this method includes:Determine the reference edge bianry image and edges matched bianry image of scene;Thick matching anaglyph is determined according to the pixel that effective parallax be present and corresponding effectively parallax;The effective parallax of non-edge pixels point and edge pixel point in slightly matching anaglyph in benchmark image, determine the disparity range of non-edge pixels point and edge pixel point;According to the value of the cost function of pixel in the grey decision-making calculating benchmark image of pixel in the grey decision-making of pixel in disparity range and benchmark image and matching image;Corresponding target parallax during the value minimum of cost function is determined, according to benchmark image and target parallax generation essence matching anaglyph.According to an embodiment of the invention, the disparity range for partial pixel point shifted matching window in image is matched is reduced, can quickly determine the parallax of different pixels point in whole benchmark image.
Description
Technical field
The present invention relates to Stereo Matching Technology field, more particularly to parallax to determine method and parallax determining device.
Background technology
During Stereo matching, in order to determine to match parallax of the image relative to benchmark image, in benchmark image
Each pixel benchmark window can be set, for matching image in each pixel match window, and root can be set
The cost function determined according to the grey decision-making of pixel in two windows, then according to being moved in image is matched in certain disparity range
Dynamic match window, using corresponding parallax during the value maximum of the cost function of determination as the parallax of pixel in benchmark window, enter
And the parallax of each pixel in benchmark image can be determined, so as to generate anaglyph.
But currently for each pixel in benchmark image, the regarding selected by shifted matching window in image is matched
Poor scope is all identical, but each pixel point feature obvious degree actually in benchmark image is different, in correlation technique
Do not treated with a certain discrimination for different pixels to choose suitable disparity range shifted matching window, which results in
Determine the less efficient of parallax.
The content of the invention
The present invention provides parallax and determines method and parallax determining device, to solve the deficiency in correlation technique.
First aspect according to embodiments of the present invention, there is provided a kind of parallax determines method, including:
Benchmark image and matching image based on two camera shooting Same Scenes, determine the reference edge two of the scene
It is worth image and edges matched bianry image;
According to the pixel that effective parallax in the reference edge bianry image and the edges matched bianry image be present
With corresponding effective parallax, the thick matching anaglyph of the benchmark image is determined;
Effective parallax of non-edge pixels point and edge pixel point in the thick matching anaglyph, it is determined that described
The non-edge pixels point of benchmark image and the disparity range of edge pixel point;
The grey decision-making of pixel in the disparity range, the benchmark image and the matching image, calculates institute
State the value of the cost function of pixel in benchmark image;
The value of the cost function is defined as to the parallax value of pixel in the benchmark image for the parallax value of minimum value.
Second aspect according to embodiments of the present invention, there is provided a kind of parallax determining device, including:
Bianry image determining unit, for shooting the benchmark image and matching image of Same Scene based on two cameras, really
The reference edge bianry image and edges matched bianry image of the fixed scene;
Thick matching unit, for there are according in the reference edge bianry image and the edges matched bianry image
The pixel of parallax and corresponding effective parallax are imitated, determines the thick matching anaglyph of the benchmark image;
Scope determining unit, for the non-edge pixels point in the thick matching anaglyph and edge pixel point
Effective parallax, determine the non-edge pixels point of the benchmark image and the disparity range of edge pixel point;
Cost calculating unit, for the pixel in the disparity range, the benchmark image and the matching image
The grey decision-making of point, calculate the value of the cost function of pixel in the benchmark image;
Smart matching unit, for the value of the cost function to be defined as in the benchmark image for the parallax value of minimum value
The parallax value of pixel.
The third aspect according to embodiments of the present invention, there is provided a kind of terminal, including memory, processor and it is stored in storage
On device and the computer program that can run on a processor, following walk is realized described in the computing device during computer program
Suddenly:
Benchmark image and matching image based on two camera shooting Same Scenes, determine the reference edge two of the scene
It is worth image and edges matched bianry image;
According to the pixel that effective parallax in the reference edge bianry image and the edges matched bianry image be present
With corresponding effective parallax, the thick matching anaglyph of the benchmark image is determined;
Effective parallax of non-edge pixels point and edge pixel point in the thick matching anaglyph, it is determined that described
The non-edge pixels point of benchmark image and the disparity range of edge pixel point;
The grey decision-making of pixel in the disparity range, the benchmark image and the matching image, calculates institute
State the value of the cost function of pixel in benchmark image;
The value of the cost function is defined as to the parallax value of pixel in the benchmark image for the parallax value of minimum value.
Fourth aspect according to embodiments of the present invention, there is provided a kind of computer-readable recording medium, be stored thereon with calculating
Machine program, the computer program realize following steps when being executed by processor:
Benchmark image and matching image based on two camera shooting Same Scenes, determine the reference edge two of the scene
It is worth image and edges matched bianry image;
According to the pixel that effective parallax in the reference edge bianry image and the edges matched bianry image be present
With corresponding effective parallax, the thick matching anaglyph of the benchmark image is determined;
Effective parallax of non-edge pixels point and edge pixel point in the thick matching anaglyph, it is determined that described
The non-edge pixels point of benchmark image and the disparity range of edge pixel point;
The grey decision-making of pixel in the disparity range, the benchmark image and the matching image, calculates institute
State the value of the cost function of pixel in benchmark image;
The value of the cost function is defined as to the parallax value of pixel in the benchmark image for the parallax value of minimum value.
From above-described embodiment, for non-edge pixels point, the first parallax of shifted matching window in match window
Scope, and for the pixel to match of effective parallax be present, the second parallax of shifted matching window in match window
Scope, respectively less than default disparity range, it is directed to partial pixel point shifted matching window in image is matched so as to reduce and regards
Poor scope, the parallax of different pixels point in whole benchmark image can be quickly determined, improve the efficiency for determining parallax, Jin Erti
The efficiency of height generation anaglyph, and the degree of accuracy of matching is improved to a certain extent.
It should be appreciated that the general description and following detailed description of the above are only exemplary and explanatory, not
Can the limitation present invention.
Brief description of the drawings
Accompanying drawing herein is merged in specification and forms the part of this specification, shows the implementation for meeting the present invention
Example, and for explaining principle of the invention together with specification.
Fig. 1 shows that a kind of parallax according to an embodiment of the invention determines the schematic flow diagram of method.
Fig. 2 is a kind of schematic diagram of Filtering Template according to one embodiment of the invention.
Fig. 3 is a kind of schematic diagram of source images according to one embodiment of the invention.
Fig. 4 is a kind of schematic diagram of edge binary images according to one embodiment of the invention.
Fig. 5 is a kind of schematic diagram of window according to one embodiment of the invention.
Fig. 6 shows a kind of signal of thick matching anaglyph for determining benchmark image according to an embodiment of the invention
Flow chart.
Fig. 7 shows a kind of schematic flow diagram for determining disparity range according to an embodiment of the invention.
Fig. 8 shows another schematic flow diagram for determining disparity range according to an embodiment of the invention.
Fig. 9 shows that another parallax according to an embodiment of the invention determines the schematic flow diagram of method.
Figure 10 shows that another parallax according to an embodiment of the invention determines the schematic flow diagram of method.
Figure 11 is the benchmark image according to one embodiment of the invention.
Figure 12 is the matching image according to one embodiment of the invention.
Figure 13 is that the Stereo matching disparity map based on side is shown according to one embodiment of the invention.
Figure 14 is a kind of schematic block diagram of parallax determining device according to one embodiment of the invention.
Figure 15 is the schematic block diagram of another parallax determining device according to one embodiment of the invention.
Figure 16 is the schematic block diagram of another parallax determining device according to one embodiment of the invention.
Figure 17 is the schematic block diagram of another parallax determining device according to one embodiment of the invention.
Figure 18 is the schematic block diagram of another parallax determining device according to one embodiment of the invention.
Figure 19 is the schematic block diagram of another parallax determining device according to one embodiment of the invention.
Embodiment
Here exemplary embodiment will be illustrated in detail, its example is illustrated in the accompanying drawings.Following description is related to
During accompanying drawing, unless otherwise indicated, the same numbers in different accompanying drawings represent same or analogous key element.Following exemplary embodiment
Described in embodiment do not represent and the consistent all embodiments of the present invention.On the contrary, they be only with it is such as appended
The example of the consistent apparatus and method of some aspects being described in detail in claims, of the invention.
Fig. 1 shows that a kind of parallax according to an embodiment of the invention determines the schematic flow diagram of method.Fig. 1 so,
This method includes:
In step s 11, benchmark image and matching image based on two camera shooting Same Scenes, determine the scene
Reference edge bianry image and edges matched bianry image.
In one embodiment, image can be gathered respectively according to two-sector model as collecting device, such as camera, will wherein
The image of one image capture device collection is as benchmark image, using the image of another image capture device collection as matching
Image.
In one embodiment, for benchmark image and matching image, marginalisation processing can be first carried out, is determined in image
Edge pixel point and non-edge pixels point, then binary conversion treatment is carried out for two class pixels, and then according to benchmark image
Reference edge bianry image is obtained, and edges matched bianry image is obtained according to matching image.
In one embodiment, before marginalisation and binary conversion treatment is carried out, benchmark image and matching can also be schemed
As being filtered smoothing processing, to remove the point of the noise pixel in image.Wherein it is possible to benchmark is regarded respectively using gaussian filtering
Figure and match views are filtered smoothing processing.
Fig. 2 is a kind of schematic diagram of Filtering Template according to one embodiment of the invention.
In one embodiment, 3 × 3 Filtering Template as shown in Figure 2 can be used, to benchmark image and movement images
In each pixel carry out gaussian filtering, for the sake of convenient, filtered benchmark image might as well be designated as imgR, after filtering
Movement images be designated as imgL.
In one embodiment, edge binary conversion treatment can specifically include two steps:
First, extracting reference edge pixel from benchmark image, edges matched pixel is extracted from matching image, its
In, the process of extraction edge pixel point can use canny operators, can also use Sobel operators, Isotropic Sobel
Operator, Roberts operators, Prewitt operators, Laplacian operators.
Second, carrying out binary conversion treatment to reference edge pixel, binary conversion treatment is carried out to edges matched pixel, its
In, the process of binary conversion treatment can use canny operators, can also use Kirsch operators.
Preferably, above-mentioned two step can be completed according to canny operators to operate.
Fig. 3 is a kind of schematic diagram of source images according to one embodiment of the invention.
As shown in figure 3, left hand view is the left image collected according to an image capture device, right part of flg is according to another
The right image that portion's image capture device collects, can will wherein left image as benchmark image, will wherein right image as
With image.
Fig. 4 is a kind of schematic diagram of edge binary images according to one embodiment of the invention.
As shown in figure 4, left hand view is that the reference edge two-value that edge binary conversion treatment obtains is carried out to benchmark image in Fig. 3
Image, right part of flg are that the edges matched bianry image that edge binary conversion treatment obtains is carried out to benchmark image in Fig. 3.
In step s 12, it is effective according to existing in the reference edge bianry image and the edges matched bianry image
The pixel of parallax and corresponding effective parallax, determine the thick matching anaglyph of the benchmark image.
In one embodiment, benchmark binarized pixel point can be determined in reference edge bianry image, with benchmark two-value
Benchmark two-value window is determined centered on pixel, matching binarized pixel point is determined in bianry image is matched, to match two-value picture
Matching two-value window is determined centered on vegetarian refreshments.
In one embodiment, benchmark window and window to be matched can be n × n window, namely include n × n picture
The window of vegetarian refreshments.
In one embodiment, binarized pixel point on the basis of the pixel in reference edge bianry image, benchmark two-value picture
Vegetarian refreshments includes reference edge binarized pixel point and benchmark non-edge binarized pixel point, and correspondingly, matching binarized pixel point includes
With edge binarized pixel point and matching non-edge binarized pixel point.Wherein, benchmark is determined only for reference edge binarized pixel point
Two-value window, and do not dealt with then for benchmark non-edge binarized pixel point.
Fig. 5 is a kind of schematic diagram of window according to one embodiment of the invention.
In one embodiment, as shown in figure 5, edge binary images, grey rectangle frame therein on the basis of left-side images
That is benchmark window, benchmark binarized pixel point are located at the center of benchmark window.Image right is edges matched bianry image, therein
Grey rectangle frame is window to be matched, and matching binarized pixel point is positioned at the center of matching two-value window, white rectangle frame therein
Illustrate the shifted matching two-value window in default disparity range.Certainly, in actual application, the size of window regards with default
Poor scope can be configured as needed, and the size of window can be with the difference shown in Fig. 5, and the scope of window movement
Can also be with the difference shown in Fig. 5.
In one embodiment, can in the default disparity range in edges matched bianry image shifted matching two-value window
Mouthful, it is determined that being the number N of 1 pixel with the values of corresponding pixel points in benchmark two-value window in matching two-value windowr, and
The number N of the different pixel of the values of corresponding pixel pointsw, calculateDetermine C maximum CmaxWhether it is more than pre-
If cost value, if being more than, there is effective parallax in pixel corresponding to record benchmark two-value window, and effectively parallax is CmaxIt is corresponding
Parallax, if being not more than, record benchmark two-value window corresponding to pixel effective parallax is not present.Then according to reference edge two
The pixel and corresponding effectively parallax that effective parallax in value image be present determine thick matching anaglyph.
Wherein, C is bigger, illustrate to match position in two-value window and benchmark two-value window (refer in the window position without
Must be the position in image) it is identical, and be edge binarized pixel point (namely a certain position pixel in two-value window is matched
Point is edges matched binarized pixel point, the binarized pixel point in edge on the basis of the pixel of same position in benchmark two-value window)
Pixel large percentage in the window, namely the matching degree of explanation matching two-value window and benchmark two-value window is higher, therefore
A default cost value can be set, if determining C maximum C in default disparity rangemaxIt is sufficiently large, namely more than default
Cost value, then can largely determine CmaxCorresponding parallax is the parallax of pixel corresponding to benchmark two-value window,
Effective parallax be present so as to record the pixel.Correspondingly, if CmaxIt is not sufficiently large, then to be then difficult to ensure that CmaxIt is corresponding
Parallax be pixel corresponding to benchmark two-value window parallax, then record the pixel and effective parallax be not present.
It should be noted that above-mentioned C denominator except can include match two-value window in the benchmark two-value window
The value of middle corresponding pixel points is the number of 1 pixel, and the number of the different pixel of value of corresponding pixel points, may be used also
It is the number of 1 pixel with the value including corresponding pixel points.
In step s 13, the non-edge pixels point in the thick matching anaglyph and edge pixel point is effective
Parallax, determine the non-edge pixels point of the benchmark image and the disparity range of edge pixel point.
In one embodiment, for non-edge pixels point and edge pixel point, the determination mode of disparity range is not
Together.
Wherein, for non-edge pixels point, benchmark non-edge two-value picture that can first in reference edge bianry image
Vegetarian refreshments determines non-edge pixels point in benchmark image, is then determined in slightly matching anaglyph on the left of non-edge pixels point most
Left effective parallax of the left hand edge pixel of the effective parallax of near presence, and the right hand edge of the nearest effective parallax of presence in right side
Right effective parallax of pixel, the first disparity range is determined according to left effectively parallax and right effectively parallax, wherein, described first regards
Poor scope is in above-mentioned default disparity range.
, can be according to the reference edge binarized pixel point in reference edge bianry image in reference map for edge pixel point
As in determine edge pixel point, in edge pixel point determine slightly matching anaglyph in exist effective parallax pixel and
Its effective parallax.
Further, can be according to effective parallax and the second default parallax for the edge pixel of effective parallax point be present
Threshold value determines the second disparity range, wherein, the second disparity range is in above-mentioned default disparity range.Effectively regarded except determining to exist
The edge pixel point of difference, the non-edges matched pixel in the absence of effective parallax can also be determined in edge pixel point, for
Non- edges matched pixel, it may be determined that its disparity range is above-mentioned default disparity range.
In step S14, the ash of the pixel in the disparity range, the benchmark image and the matching image
Rank is worth, and calculates the value of the cost function of pixel in the benchmark image.
In one embodiment, for non-edge pixels point, if can determine that a left side effectively regards in slightly matching anaglyph
Difference and right effective parallax, then the shifted matching window in the first disparity range in matching image, otherwise, in image is matched
Shifted matching window in the default disparity range.
For the edge pixel of effective parallax point be present, the shifted matching window in the second disparity range in matching image
Mouthful.
For non-edges matched pixel, shifted matching window in the default disparity range in the matching image
Mouthful.
Based on this, for non-edge pixels point, the edge pixel point of effective parallax and non-edges matched pixel be present, can
To determine that different disparity ranges carrys out shifted matching window respectively in image is matched, to be accurately determined in matching image and base
The pixel of pixel Point matching in quasi- image.
In one embodiment, cost function can be similitude cost function, such as zero-mean gray scale and coefficient correlation
Function (ZCC), Normalized Grey Level cross-correlation function (NCC), formula difference are as follows:
ZCC and NCC can be used for calculating GTG similitude.
Cost function can also be difference cost function, such as the absolute value and function (SAD) of pixel difference, and pixel is poor
Sum of squares function (SSD) etc., formula is as follows:
SAD and SSD can be used for calculating grayscale difference.
In above-mentioned formula, W is window, I1(x, y) represents in benchmark bianry image coordinate as the ash of the pixel of (x, y)
Rank, I2(x+d, y) represent matching bianry image in coordinate be the pixel of (x+d, y) GTG, d expression parallax, avgW1Represent
Pixel mean gray-scale, avgW in benchmark window2Represent pixel average average in window to be matched.
In the case where cost function is similitude cost function, corresponding parallax is picture during the value maximum of cost function
The parallax value of vegetarian refreshments;In the case where cost function is difference cost function, corresponding parallax during the value minimum of cost function
For the parallax value of pixel.
In step S15, the value of the cost function is defined as picture in the benchmark image for the parallax value of minimum value
The parallax value of vegetarian refreshments.
In one embodiment, each pixel in benchmark image can be directed to and determine parallax value respectively, then according to every
Parallax value corresponding to individual pixel, adjusts the display effect of pixel, for example, parallax value it is larger pixel it is shown in red, depending on
The less pixel of difference is shown as blueness, can both obtain anaglyph.
In one embodiment, for non-edge pixels point, the first parallax model of shifted matching window in match window
Enclose, and for the edge pixel of effective parallax point be present, the second disparity range of shifted matching window in match window,
Less than default disparity range, so as to reduce the disparity range for being directed to partial pixel point shifted matching window in image is matched,
The parallax of different pixels point in whole benchmark image can be quickly determined, improves the efficiency for determining parallax, and then improve generation
The efficiency of anaglyph, additionally it is possible to improve the degree of accuracy of matching to a certain extent.
Fig. 6 shows a kind of signal of thick matching anaglyph for determining benchmark image according to an embodiment of the invention
Flow chart.It is as shown in fig. 6, described effective according to existing in the reference edge bianry image and the edges matched bianry image
The pixel of parallax and corresponding effective parallax, the thick matching anaglyph of the benchmark image is determined, including:
In step S121, it is determined that the pixel and benchmark two-value of edges matched bianry image described in matching two-value window
The pixel value of the pixel of reference edge bianry image described in window is the quantity N of 1 pixelr;
In step S122, it is determined that the pixel and benchmark two-value of edges matched bianry image described in matching two-value window
The quantity N of the different pixel of the pixel values of the corresponding pixel points of reference edge bianry image described in windoww;
In step S123, the mobile matching two-value in the default disparity range in the edges matched bianry image
Window, according to formulaDetermine C maximum Cmax;
In step S124, if CmaxMore than default cost value, it is determined that pixel corresponding to the benchmark two-value window
Parallax value be effective parallax, and the effectively parallax is CmaxCorresponding parallax;
In step s 125, if CmaxNo more than default cost value, it is determined that pixel corresponding to the benchmark two-value window
The parallax value of point is invalid parallax;
It is effective according to existing in the reference edge bianry image and the edges matched bianry image in step S126
The pixel of parallax and corresponding effective parallax, determine the thick matching anaglyph of the benchmark image.
In one embodiment, C is bigger, illustrates that matching two-value window is identical with position in benchmark two-value window, and be
The pixel of edge binarized pixel point large percentage, namely of explanation matching two-value window and benchmark two-value window in the window
It is higher with spending.
Therefore a default cost value can be set, if determining C maximum C in default disparity rangemaxIt is sufficiently large,
Namely CmaxMore than default cost value, then can largely determine CmaxCorresponding parallax is that benchmark two-value window is corresponding
Pixel parallax, effective parallax be present so as to record the pixel.Correspondingly, if CmaxIt is not sufficiently large, namely
CmaxNo more than default cost value, then be just difficult to ensure that CmaxCorresponding parallax is regarding for pixel corresponding to benchmark two-value window
Difference, then record the pixel and effective parallax is not present.
Further, for the pixel of effective parallax be present, the information of effective parallax can be attached to corresponding pixel
On point, such as according to different effective parallax set pixel that there is different colors, so as to obtain reference edge two-value
Image has the disparity map that the pixel of effective parallax is formed, namely slightly matches disparity map.
Fig. 7 shows a kind of schematic flow diagram for determining disparity range according to an embodiment of the invention.Such as Fig. 7 institutes
Show, effective parallax of the non-edge pixels point and edge pixel point in the thick matching anaglyph, it is determined that described
The non-edge pixels point of benchmark image and the disparity range of edge pixel point include:
In step S131, if the non-edge pixels point in the benchmark image is in the thick matching anaglyph
Be present left effectively parallax and right effective parallax in neighbor pixel, then determined according to the left effectively parallax and the right effectively parallax
First disparity range of the non-edge pixels point, wherein, first disparity range is in default disparity range;
In step S132, if the non-edge pixels point in the benchmark image is in the thick matching anaglyph
Left effectively parallax and right effective parallax is not present in neighbor pixel, and the disparity range for determining the non-edge pixels point is described pre-
If disparity range.
In one embodiment, it is true in slightly matching anaglyph for the non-edge pixels point in the benchmark image
Left effective parallax of the left hand edge pixel of the nearest effective parallax of presence on the left of the fixed non-edge pixels point, and right side are nearest
The effective parallax of presence right hand edge pixel right effective parallax;
For the non-edge pixels pixel in some object in benchmark image, picture corresponding in image is being matched
Plain pixel, also one be positioned in the object, namely within the edge of the object, without positioned at the edge of the object it
Outside.Based on this, for the non-edge pixels pixel in benchmark image, it may be determined that the nearest edge pixel pixel in its left side
The nearest edge pixel pixel (namely right hand edge pixel pixel) of point (namely left hand edge pixel pixel) and right side, if depositing
In left hand edge pixel pixel and right hand edge pixel pixel, then it may further determine that left hand edge pixel pixel in true edge
Effective parallax (namely left effective parallax) corresponding to corresponding reference edge binarized pixel pixel, also has the right side in edge bianry image
Edge pixel pixel in benchmark bianry image effective parallax corresponding to corresponding reference edge binarized pixel pixel (namely
Right effective parallax).So in image is matched, also should for the pixel pixel to match determined by the pixel pixel
Between the disparity range that corresponding left effectively parallax and right effective parallax are defined in matching image.
Correspondingly, if the phase of non-edge pixels pixel in the benchmark image in the thick matching anaglyph
Left effectively parallax and right effective parallax is not present in adjacent pixel pixel, namely in the absence of left hand edge pixel pixel and right hand edge picture
Plain pixel, or even if side edge pixel pixel be present, but the parallax of the lateral edges pixel pixel regards to be invalid
Difference, namely the parallax of pixel pixel point to match in matching image can not be determined according to left effectively parallax and right effectively parallax
Scope, the then disparity range that can determine the non-edge pixels pixel point are the default disparity range.
And because left hand edge pixel and right hand edge pixel are respectively positioned in matching image, therefore according to left effectively parallax and
Right effectively border of the parallax as the first disparity range, it is determined that the first disparity range be less than entirely mobile in matching image
With the default disparity range corresponding to window, so as to which according to the first disparity range shifted matching window, matching speed can be improved.
It should be noted that above-mentioned steps S131, S132 can adjust execution sequence as needed, both can be according to Fig. 7
Shown order is performed, and can also perform in other orders, such as the two is performed side by side.
Alternatively, it is described that the of the non-edge pixels point is determined according to the left effectively parallax and the right effectively parallax
One disparity range includes:
According to the left effectively parallax dLWith the right effectively parallax dRAnd the first default parallax threshold value dT1Determine first
Disparity range,
In dL<dRIn the case of, the first disparity range is [dL-dT1, dR+dT1], wherein, it is [d to preset disparity rangemin,
dmax], dL-dT1More than or equal to dmin, dR+dT1Less than or equal to dmax;
In dL>dRIn the case of, the first disparity range is [dR-dT1, dL+dT1], wherein, it is [d to preset disparity rangemin,
dmax], dR-dT1More than or equal to dmin, dL+dT1Less than or equal to dmax。
Fig. 8 shows another schematic flow diagram for determining disparity range according to an embodiment of the invention.Such as Fig. 8 institutes
Show, the non-edge pixels point and edge pixel point in the benchmark image having in the thick matching anaglyph
Parallax is imitated, determining the disparity range of the non-edge pixels point and edge pixel point includes:
In step S133, if the edge pixel point in the benchmark image exists in the thick matching anaglyph
Edges matched pixel, then according to effective parallax of the edges matched pixel and the second default parallax threshold value, it is determined that described
Second disparity range of edges matched pixel, wherein, second disparity range is in default disparity range;
In step S134, if the edge pixel point in the benchmark image is not deposited in the thick matching anaglyph
In edges matched pixel, it is determined that the second disparity range of the edges matched pixel is default disparity range.
In one embodiment, for edge pixel point, the embodiment according to Fig. 6 can record it with the presence or absence of having
Imitate parallax, and specific effective parallax in the case where effective parallax be present.For the edge pixel of effective parallax point be present,
Also that is, effective parallax of pixel in benchmark image is roughly determined, wherein, effective parallax of pixel or it is
Effective parallax, or be parallax corresponding to the pixel to match in effective parallax smaller range.Therefore can be with
Second disparity range is determined according to effective parallax and the second default parallax threshold value.
It should be noted that above-mentioned steps S133 and S134 can adjust execution sequence as needed, both can be according to Fig. 8
Shown order performs, and can also perform in other orders, such as both perform side by side.
Alternatively, it is described that the edges matched pixel the is determined according to the effectively parallax and the second default parallax threshold value
Two disparity ranges include:
Parallax threshold value d is preset according to the effectively parallax d and secondT2Determine the second disparity range [d-dT2, d+dT2], its
In, it is [d to preset disparity rangemin, dmax], d-dT2More than or equal to dmin, d+dT2Less than or equal to dmax。
In one embodiment, for edge pixel point, it is not present and effectively regards if the embodiment according to Fig. 6 records it
Difference, then illustrate the embodiment according to Fig. 6 determine matching degree mode not in bianry image match determination and
The pixel that pixel matches in benchmark image, then when carrying out essence matching in matching image, then need entirely matching
Preset corresponding to image in disparity range, matched pixel is redefined, to determine its corresponding parallax.
Fig. 9 shows that another parallax according to an embodiment of the invention determines the schematic flow diagram of method.Such as Fig. 9 institutes
Show, the value for calculating the cost function of pixel in the benchmark image includes:
In step s 141, according to pixel in the grey decision-making of pixel in the benchmark window and the match window
Grey decision-making calculates the cost function value sum C of the non-edge pixels pointS(x,y), and the cost function of edge pixel point
It is worth sum CS’(x,y);
In step S142, calculation cost function C(x,y)=λ CS’(x,y)+(1-λ)CS(x,y)Value, wherein, 0.5<λ<1.
In one embodiment, due to 0.5<λ<1, therefore (1- λ)<The weight of edge pixel point in λ, namely match window
Larger, the weight of non-edge pixels point is smaller.And it is relative in general, to be primarily upon feature during calculation cost function
The cost function value of obvious pixel (such as edge pixel point), therefore higher weight is set for edge pixel point, improve
Weight of the more obvious pixel of feature (namely edge pixel point) in cost function in image, and then improve and calculate generation
The accuracy of valency function.
Figure 10 shows that another parallax according to an embodiment of the invention determines the schematic flow diagram of method.Such as Figure 10
Shown, after the anaglyph is generated, methods described also includes:
In step s 16, filtering processing is carried out to the edge pixel point in the anaglyph;
In step S17, handled according to the target parallax of edge pixel point in the benchmark bianry image and filtering
Edge pixel point afterwards, generates the Stereo matching anaglyph based on side.
In one embodiment, filtering process, and only extraction filter can be only carried out to the edge pixel point in anaglyph
Edge pixel point after ripple smoothing processing determines target parallax, so as to obtain only existing the anaglyph of edge pixel point,
Namely the Stereo matching anaglyph based on side.
Figure 11 is the benchmark image according to one embodiment of the invention.Figure 12 is shown according to one embodiment of the invention
The matching image gone out.Figure 13 is that the Stereo matching disparity map based on side is shown according to one embodiment of the invention.
In one embodiment, can be to benchmark image (such as left image) as shown in figure 11 according to above-described embodiment
Handled with matching image (such as right image) as shown in figure 12, so as to accurately determine the edge picture for wherein belonging to edge
Vegetarian refreshments, and then the Stereo matching disparity map based on side as shown in fig. 13 that is accurately generated according to edge pixel point.
The embodiment of method is determined with above-mentioned parallax accordingly, and the invention also provides a kind of reality of parallax determining device
Apply example.
Figure 14 is a kind of schematic block diagram of parallax determining device according to one embodiment of the invention.Such as Figure 14 institutes
Show, parallax determining device includes:
Bianry image determining unit 1, for shooting the benchmark image and matching image of Same Scene based on two cameras, really
The reference edge bianry image and edges matched bianry image of the fixed scene;
Thick matching unit 2, for existing according in the reference edge bianry image and the edges matched bianry image
The pixel of effective parallax and corresponding effective parallax, determine the thick matching anaglyph of the benchmark image;
Scope determining unit 3, for the non-edge pixels point and edge pixel point in the thick matching anaglyph
Effective parallax, determine the non-edge pixels point of the benchmark image and the disparity range of edge pixel point;
Cost calculating unit 4, for the pixel in the disparity range, the benchmark image and the matching image
The grey decision-making of point, calculate the value of the cost function of pixel in the benchmark image;
Smart matching unit 5, for the value of the cost function to be defined as into the benchmark image for the parallax value of minimum value
The parallax value of middle pixel.
Figure 15 is the schematic block diagram of another parallax determining device according to one embodiment of the invention.Such as Figure 15 institutes
Show, the thick matching unit 2 includes:
Quantity determination subelement 21, for determine matching two-value window described in edges matched bianry image pixel with
The pixel value of the pixel of reference edge bianry image described in benchmark two-value window is the quantity N of 1 pixelr;And
It is determined that the pixel of edges matched bianry image described in matching two-value window and reference edge two described in benchmark two-value window
It is worth the quantity N of the different pixel of pixel value of the corresponding pixel points of imagew;
Cost value subelement 22, for mobile described in the default disparity range in the edges matched bianry image
With two-value window, according to formulaDetermine C maximum Cmax;
Parallax subelement 23, for existing according in the reference edge bianry image and the edges matched bianry image
The pixel of effective parallax and corresponding effective parallax, determine the thick matching anaglyph of the benchmark image.
Figure 16 is the schematic block diagram of another parallax determining device according to one embodiment of the invention.Such as Figure 16 institutes
Show, the scope determining unit 3 includes:
First scope determination subelement 31, if the non-edge pixels point in the benchmark image is in the thick matching parallax
There is left effective parallax and right effective parallax in the neighbor pixel in image, then according to the left effectively parallax and the right side effectively
Parallax determines the first disparity range of the non-edge pixels point, wherein, first disparity range is in default disparity range;
If neighbor pixel of the non-edge pixels point in the thick matching anaglyph in the benchmark image has in the absence of a left side
Parallax and right effective parallax are imitated, the disparity range for determining the non-edge pixels point is the default disparity range.
Alternatively, the scope determination subelement is used for according to the left effectively parallax dLWith the right effectively parallax dRWith
And the first default parallax threshold value dT1The first disparity range is determined,
In dL<dRIn the case of, the first disparity range is [dL-dT1, dR+dT1], wherein, it is [d to preset disparity rangemin,
dmax], dL-dT1More than or equal to dmin, dR+dT1Less than or equal to dmax;
In dL>dRIn the case of, the first disparity range is [dR-dT1, dL+dT1], wherein, it is [d to preset disparity rangemin,
dmax], dR-dT1More than or equal to dmin, dL+dT1Less than or equal to dmax。
Figure 17 is the schematic block diagram of another parallax determining device according to one embodiment of the invention.Such as Figure 17 institutes
Show, the scope determining unit 3 includes:
Second scope determination subelement 32, if the edge pixel point in the benchmark image is in the thick matching disparity map
Edges matched pixel be present as in, then parallax threshold value preset according to effective parallax of the edges matched pixel and second,
The second disparity range of the edges matched pixel is determined, wherein, second disparity range is in default disparity range;Such as
Edges matched pixel is not present in the thick matching anaglyph in edge pixel point in benchmark image described in fruit, it is determined that
Second disparity range of the edges matched pixel is default disparity range.
Alternatively, the second scope determination subelement is used to preset parallax threshold value according to the effectively parallax d and second
dT2Determine the second disparity range [d-dT2, d+dT2], wherein, it is [d to preset disparity rangemin, dmax], d-dT2It is more than or equal to
dmin, d+dT2Less than or equal to dmax。
Figure 18 is the schematic block diagram of another parallax determining device according to one embodiment of the invention.Such as Figure 18 institutes
Show, the cost calculating unit 4 includes:
Cost and subelement 41, for according to picture in the grey decision-making of pixel in the benchmark window and the match window
The grey decision-making of vegetarian refreshments calculates the cost function value sum C of the non-edge pixels pointS(x,y), and the generation of edge pixel point
Valency functional value sum CS’(x,y);
Subelement 42 is weighted, for calculation cost function C(x,y)=λ CS’(x,y)+(1-λ)CS(x,y)Value, wherein, 0.5<λ<
1。
Figure 19 is the schematic block diagram of another parallax determining device according to one embodiment of the invention.Such as Figure 19 institutes
Show, in addition to:
Smoothing processing unit 7, for carrying out filtering processing to the edge pixel point in the anaglyph;
Image generation unit 8, for flat according to the target parallax of edge pixel point in the benchmark bianry image and filtering
Edge pixel point after sliding processing, generates the Stereo matching anaglyph based on side.
On the device in above-described embodiment, wherein unit performs the concrete mode of operation in the method for correlation
Embodiment in be described in detail, explanation will be not set forth in detail herein.
For device embodiment, because it corresponds essentially to embodiment of the method, so related part is real referring to method
Apply the part explanation of example.Device embodiment described above is only schematical, wherein described be used as separating component
The unit of explanation can be or may not be physically separate, can be as the part that unit is shown or can also
It is not physical location, you can with positioned at a place, or can also be distributed on multiple NEs.Can be according to reality
Need to select some or all of module therein to realize the purpose of the present invention program.Those of ordinary skill in the art are not paying
In the case of going out creative work, you can to understand and implement.
Embodiments of the invention are related to a kind of terminal, including memory, processor and storage on a memory and can located
The computer program run on reason device, following steps are realized described in the computing device during computer program:
Benchmark image and matching image based on two camera shooting Same Scenes, determine the reference edge two of the scene
It is worth image and edges matched bianry image;
According to the pixel that effective parallax in the reference edge bianry image and the edges matched bianry image be present
With corresponding effective parallax, the thick matching anaglyph of the benchmark image is determined;
Effective parallax of non-edge pixels point and edge pixel point in the thick matching anaglyph, it is determined that described
The non-edge pixels point of benchmark image and the disparity range of edge pixel point;
The grey decision-making of pixel in the disparity range, the benchmark image and the matching image, calculates institute
State the value of the cost function of pixel in benchmark image;
The value of the cost function is defined as to the parallax value of pixel in the benchmark image for the parallax value of minimum value.
Embodiments of the invention are related to a kind of computer-readable recording medium, are stored thereon with computer program, the meter
Calculation machine program realizes following steps when being executed by processor:
Benchmark image and matching image based on two camera shooting Same Scenes, determine the reference edge two of the scene
It is worth image and edges matched bianry image;
According to the pixel that effective parallax in the reference edge bianry image and the edges matched bianry image be present
With corresponding effective parallax, the thick matching anaglyph of the benchmark image is determined;
Effective parallax of non-edge pixels point and edge pixel point in the thick matching anaglyph, it is determined that described
The non-edge pixels point of benchmark image and the disparity range of edge pixel point;
The grey decision-making of pixel in the disparity range, the benchmark image and the matching image, calculates institute
State the value of the cost function of pixel in benchmark image;
The value of the cost function is defined as to the parallax value of pixel in the benchmark image for the parallax value of minimum value.
Those skilled in the art will readily occur to the present invention its after considering specification and putting into practice disclosure disclosed herein
Its embodiment.It is contemplated that cover the present invention any modification, purposes or adaptations, these modifications, purposes or
Person's adaptations follow the general principle of the present invention and including undocumented common knowledges in the art of the invention
Or conventional techniques.Description and embodiments are considered only as exemplary, and true scope and spirit of the invention are by following
Claim is pointed out.
It should be appreciated that the invention is not limited in the precision architecture for being described above and being shown in the drawings, and
And various modifications and changes can be being carried out without departing from the scope.The scope of the present invention is only limited by appended claim.
Claims (10)
1. a kind of parallax determines method, it is characterised in that including:
Benchmark image and matching image based on two camera shooting Same Scenes, determine the reference edge binary map of the scene
Picture and edges matched bianry image;
According to the pixel and phase that effective parallax in the reference edge bianry image and the edges matched bianry image be present
The effective parallax answered, determine the thick matching anaglyph of the benchmark image;
Effective parallax of non-edge pixels point and edge pixel point in the thick matching anaglyph, determines the benchmark
The non-edge pixels point of image and the disparity range of edge pixel point;
The grey decision-making of pixel in the disparity range, the benchmark image and the matching image, calculates the base
The value of the cost function of pixel in quasi- image;
The value of the cost function is defined as to the parallax value of pixel in the benchmark image for the parallax value of minimum value.
2. according to the method for claim 1, it is characterised in that described according to the reference edge bianry image and described
With the pixel that effective parallax in edge binary images be present and corresponding effective parallax, the thick matching of the benchmark image is determined
Anaglyph, including:
It is determined that the pixel of edges matched bianry image described in matching two-value window and true edge described in benchmark two-value window
The pixel value of the pixel of edge bianry image is the quantity N of 1 pixelr;
It is determined that the pixel of edges matched bianry image described in matching two-value window and true edge described in benchmark two-value window
The quantity N of the different pixel of the pixel values of the corresponding pixel points of edge bianry imagew;
The mobile matching two-value window in default disparity range in the edges matched bianry image, according to formulaDetermine C maximum Cmax;
If CmaxMore than default cost value, it is determined that the parallax value of pixel corresponding to the benchmark two-value window is effectively to regard
Difference, and the effectively parallax is CmaxCorresponding parallax;
If CmaxNo more than default cost value, it is determined that the parallax value of pixel corresponding to the benchmark two-value window is invalid
Parallax;
According to the pixel and phase that effective parallax in the reference edge bianry image and the edges matched bianry image be present
The effective parallax answered, determine the thick matching anaglyph of the benchmark image.
3. method according to claim 1 or 2, it is characterised in that described non-in the thick matching anaglyph
Effective parallax of edge pixel point and edge pixel point, determine the non-edge pixels point and edge pixel point of the benchmark image
Disparity range includes:
If neighbor pixel of the non-edge pixels point in the thick matching anaglyph in the benchmark image exists left
Effective parallax and right effective parallax, then the non-edge pixels point is determined according to the left effectively parallax and the right effectively parallax
The first disparity range, wherein, first disparity range is in default disparity range;
If neighbor pixel of the non-edge pixels point in the thick matching anaglyph in the benchmark image is not present
Left effectively parallax and right effective parallax, the disparity range for determining the non-edge pixels point is the default disparity range.
4. method according to claim 1 or 2, it is characterised in that described non-in the thick matching anaglyph
Effective parallax of edge pixel point and edge pixel point, determine the non-edge pixels point and edge pixel point of the benchmark image
Disparity range includes:
If edges matched pixel be present in the thick matching anaglyph in the edge pixel point in the benchmark image,
According to effective parallax of the edges matched pixel and the second default parallax threshold value, the of the edges matched pixel is determined
Two disparity ranges, wherein, second disparity range is in default disparity range;
If edges matched pixel is not present in the thick matching anaglyph in the edge pixel point in the benchmark image,
The second disparity range for then determining the edges matched pixel is default disparity range.
5. method according to any one of claim 1 to 4, it is characterised in that it is described according to the disparity range, it is described
The grey decision-making of benchmark image and the pixel in the matching image, calculates the cost function of pixel in the benchmark image
Value includes:
Calculated according to the grey decision-making of pixel in the grey decision-making of pixel in the benchmark window and the match window described non-
The cost function value sum C of edge pixel pointS(x,y), and the cost function value sum C of edge pixel pointS’(x,y);
Calculation cost function C(x,y)=λ CS’(x,y)+(1-λ)CS(x,y)Value, wherein, 0.5<λ<1.
A kind of 6. parallax determining device, it is characterised in that including:
Bianry image determining unit, for benchmark image and matching image based on two camera shooting Same Scenes, determine institute
State the reference edge bianry image and edges matched bianry image of scene;
Thick matching unit, for effectively regarding according to existing in the reference edge bianry image and the edges matched bianry image
The pixel of difference and corresponding effective parallax, determine the thick matching anaglyph of the benchmark image;
Scope determining unit, for the effective of the non-edge pixels point in the thick matching anaglyph and edge pixel point
Parallax, determine the non-edge pixels point of the benchmark image and the disparity range of edge pixel point;
Cost calculating unit, for the pixel in the disparity range, the benchmark image and the matching image
Grey decision-making, calculate the value of the cost function of pixel in the benchmark image;
Smart matching unit, for the value of the cost function to be defined as into pixel in the benchmark image for the parallax value of minimum value
The parallax value of point.
7. device according to claim 6, it is characterised in that the thick matching unit includes:
Quantity determination subelement, for determining the pixel and benchmark two of edges matched bianry image described in matching two-value window
The pixel value of the pixel of reference edge bianry image described in value window is the quantity N of 1 pixelr;And determination
With reference edge bianry image described in the pixel of edges matched bianry image described in two-value window and benchmark two-value window
Corresponding pixel points the different pixel of pixel value quantity Nw;
Cost value subelement, for the mobile matching two-value in the default disparity range in the edges matched bianry image
Window, according to formulaDetermine C maximum Cmax;
Parallax subelement, for effectively regarding according to existing in the reference edge bianry image and the edges matched bianry image
The pixel of difference and corresponding effective parallax, determine the thick matching anaglyph of the benchmark image.
8. the device according to claim 6 or 7, it is characterised in that the scope determining unit includes:
First scope determination subelement, if the non-edge pixels point in the benchmark image is in the thick matching anaglyph
Neighbor pixel left effectively parallax and right effective parallax be present, then it is true according to the left effectively parallax and the right effectively parallax
First disparity range of the fixed non-edge pixels point, wherein, first disparity range is in default disparity range;If institute
Left effective parallax is not present in neighbor pixel of the non-edge pixels point in the thick matching anaglyph stated in benchmark image
With right effective parallax, the disparity range for determining the non-edge pixels point is the default disparity range.
9. a kind of terminal, including memory, processor and storage are on a memory and the computer journey that can run on a processor
Sequence, it is characterised in that realize following steps during computer program described in the computing device:
Benchmark image and matching image based on two camera shooting Same Scenes, determine the reference edge binary map of the scene
Picture and edges matched bianry image;
According to the pixel and phase that effective parallax in the reference edge bianry image and the edges matched bianry image be present
The effective parallax answered, determine the thick matching anaglyph of the benchmark image;
Effective parallax of non-edge pixels point and edge pixel point in the thick matching anaglyph, determines the benchmark
The non-edge pixels point of image and the disparity range of edge pixel point;
The grey decision-making of pixel in the disparity range, the benchmark image and the matching image, calculates the base
The value of the cost function of pixel in quasi- image;
The value of the cost function is defined as to the parallax value of pixel in the benchmark image for the parallax value of minimum value.
10. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that the computer program
Following steps are realized when being executed by processor:
Benchmark image and matching image based on two camera shooting Same Scenes, determine the reference edge binary map of the scene
Picture and edges matched bianry image;
According to the pixel and phase that effective parallax in the reference edge bianry image and the edges matched bianry image be present
The effective parallax answered, determine the thick matching anaglyph of the benchmark image;
Effective parallax of non-edge pixels point and edge pixel point in the thick matching anaglyph, determines the benchmark
The non-edge pixels point of image and the disparity range of edge pixel point;
The grey decision-making of pixel in the disparity range, the benchmark image and the matching image, calculates the base
The value of the cost function of pixel in quasi- image;
The value of the cost function is defined as to the parallax value of pixel in the benchmark image for the parallax value of minimum value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710823614.3A CN107680083B (en) | 2017-09-13 | 2017-09-13 | Parallax determination method and parallax determination device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710823614.3A CN107680083B (en) | 2017-09-13 | 2017-09-13 | Parallax determination method and parallax determination device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107680083A true CN107680083A (en) | 2018-02-09 |
CN107680083B CN107680083B (en) | 2020-05-29 |
Family
ID=61136428
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710823614.3A Active CN107680083B (en) | 2017-09-13 | 2017-09-13 | Parallax determination method and parallax determination device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107680083B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108921902A (en) * | 2018-06-06 | 2018-11-30 | 链家网(北京)科技有限公司 | A kind of method and device of correcting principle light camera deviation |
CN109993781A (en) * | 2019-03-28 | 2019-07-09 | 北京清微智能科技有限公司 | Based on the matched anaglyph generation method of binocular stereo vision and system |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101625761A (en) * | 2009-08-06 | 2010-01-13 | 浙江工业大学 | Computer binocular vision matching method based on global and local algorithms |
CN102665086A (en) * | 2012-04-26 | 2012-09-12 | 清华大学深圳研究生院 | Method for obtaining parallax by using region-based local stereo matching |
CN102999913A (en) * | 2012-11-29 | 2013-03-27 | 清华大学深圳研究生院 | Local three-dimensional matching method based on credible point spreading |
CN103440653A (en) * | 2013-08-27 | 2013-12-11 | 北京航空航天大学 | Binocular vision stereo matching method |
JP2015019346A (en) * | 2013-06-12 | 2015-01-29 | 日本放送協会 | Parallax image generator |
US20160173852A1 (en) * | 2014-12-16 | 2016-06-16 | Kyungpook National University Industry-Academic Cooperation Foundation | Disparity computation method through stereo matching based on census transform with adaptive support weight and system thereof |
-
2017
- 2017-09-13 CN CN201710823614.3A patent/CN107680083B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101625761A (en) * | 2009-08-06 | 2010-01-13 | 浙江工业大学 | Computer binocular vision matching method based on global and local algorithms |
CN102665086A (en) * | 2012-04-26 | 2012-09-12 | 清华大学深圳研究生院 | Method for obtaining parallax by using region-based local stereo matching |
CN102999913A (en) * | 2012-11-29 | 2013-03-27 | 清华大学深圳研究生院 | Local three-dimensional matching method based on credible point spreading |
JP2015019346A (en) * | 2013-06-12 | 2015-01-29 | 日本放送協会 | Parallax image generator |
CN103440653A (en) * | 2013-08-27 | 2013-12-11 | 北京航空航天大学 | Binocular vision stereo matching method |
US20160173852A1 (en) * | 2014-12-16 | 2016-06-16 | Kyungpook National University Industry-Academic Cooperation Foundation | Disparity computation method through stereo matching based on census transform with adaptive support weight and system thereof |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108921902A (en) * | 2018-06-06 | 2018-11-30 | 链家网(北京)科技有限公司 | A kind of method and device of correcting principle light camera deviation |
CN109993781A (en) * | 2019-03-28 | 2019-07-09 | 北京清微智能科技有限公司 | Based on the matched anaglyph generation method of binocular stereo vision and system |
Also Published As
Publication number | Publication date |
---|---|
CN107680083B (en) | 2020-05-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11360571B2 (en) | Information processing device and method, program and recording medium for identifying a gesture of a person from captured image data | |
Saha et al. | Mutual spectral residual approach for multifocus image fusion | |
CN108377374B (en) | Method and system for generating depth information related to an image | |
JP4597391B2 (en) | Facial region detection apparatus and method, and computer-readable recording medium | |
Ansar et al. | Enhanced real-time stereo using bilateral filtering | |
CN101542529A (en) | Generation of depth map for an image | |
CN110287791B (en) | Screening method and system for face pictures | |
KR20150079576A (en) | Depth map generation from a monoscopic image based on combined depth cues | |
CN107864337A (en) | Sketch image processing method, device and equipment | |
CN108257165A (en) | Image solid matching method, binocular vision equipment | |
Anantrasirichai et al. | Terrain classification from body-mounted cameras during human locomotion | |
CN104751406A (en) | Method and device used for blurring image | |
Chibber et al. | Smart traffic light controller using edge detection in digital signal processing | |
CN110705634A (en) | Heel model identification method and device and storage medium | |
CN107680083A (en) | Parallax determines method and parallax determining device | |
CN113628202B (en) | Determination method, cleaning robot and computer storage medium | |
CN107403448A (en) | Cost function generation method and cost function generating means | |
WO2024016632A1 (en) | Bright spot location method, bright spot location apparatus, electronic device and storage medium | |
CN106611417B (en) | Method and device for classifying visual elements into foreground or background | |
RU2580466C1 (en) | Device for recovery of depth map of scene | |
Veksler | Semi-dense stereo correspondence with dense features | |
CN110349110B (en) | Blurred image enhancement method based on accumulative frame over-fusion and application | |
CN116579968A (en) | Identification method and device for food material image, steaming and baking equipment and storage medium | |
Tran et al. | Variational disparity estimation framework for plenoptic images | |
Sandhu et al. | Comparative Analysis of Edge Detection Techniques for extracting Refined Boundaries |
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 |