CN107248179A - Three-dimensional matching method for building up for disparity computation - Google Patents
Three-dimensional matching method for building up for disparity computation Download PDFInfo
- Publication number
- CN107248179A CN107248179A CN201710427806.2A CN201710427806A CN107248179A CN 107248179 A CN107248179 A CN 107248179A CN 201710427806 A CN201710427806 A CN 201710427806A CN 107248179 A CN107248179 A CN 107248179A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msup
- disparity
- subgraph
- gradient
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
- G06T7/85—Stereo camera calibration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/50—Depth or shape recovery
- G06T7/55—Depth or shape recovery from multiple images
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses the three-dimensional matching method for building up for disparity computation:(a) two video cameras are demarcated respectively;(b) image for setting two video cameras acquisitions is respectively supergraph picture, subgraph, sets up the fundamental matrix relation of two video cameras:F=Ar ‑T[t]×RAl ‑1;(c) epipolar-line constraint relation is set up:m’TFm=0;(d) neighborhood image in supergraph picture, subgraph near characteristic point is obtained, and calculates the gray average in supergraph picture in neighborhood imageGray average in the gray value I of subgraph, subgraph in neighborhood image(e) gray scale similarity S is calculated;(f) gradient of disparity d is calculatedgr;(g) according to the gray scale similarity, gradient of disparity calculated, false matches value is removed, three-dimensional matching is carried out.The problem of present invention is to solve gray scale in the prior art and gradient of disparity interference three-dimensional matching precision, realizes and quantum chemical method is carried out to gray scale and gradient of disparity, so as to eliminating false matches, put forward high-precision purpose.
Description
Technical field
The present invention relates to three-dimensional measurement field, and in particular to the three-dimensional matching method for building up for disparity computation.
Background technology
Parallax is exactly the direction difference produced by same target from two points for having certain distance.In terms of target
Angle between two points, is called the parallactic angle of the two points, and the distance between 2 points are referred to as baseline.Only it is to be understood that parallactic angle degree
And baseline length, it is possible to calculate the distance between target and observer.In three-dimensional measurement field, that is, refer to from two shootings
Difference on the image obtained in machine between same pixel point.Three-dimensional matching is, according to the calculating to selected feature, to set up feature
Between corresponding relation, photosites of the same spatial point in different images are mapped, and thus obtain corresponding parallax
The technology of image.Traditional three-dimensional matching is all based on camera calibration and carried out, due to the interference of gray scale and gradient of disparity,
Easily there are false matches, and then influence three-dimensional matching precision.
The content of the invention
It is grey in the prior art to solve it is an object of the invention to provide the three-dimensional matching method for building up for disparity computation
The problem of degree and gradient of disparity interference three-dimensional matching precision, realize and quantum chemical method is carried out to gray scale and gradient of disparity, so as to eliminating
False matches, put forward high-precision purpose.
The present invention is achieved through the following technical solutions:
For the three-dimensional matching method for building up of disparity computation, comprise the following steps:
(a) two video cameras are demarcated respectively, obtains outer parameter matrix R, translation vector t and two video cameras
Respective Intrinsic Matrix Al、Ar;
(b) image for setting two video cameras acquisitions is respectively supergraph picture, subgraph, sets up the basic square of two video cameras
Battle array relation:F=Ar -T[t]×RAl -1;Wherein T is female image intensity value;
(c) epipolar-line constraint relation is set up:m’TFm=0;Wherein m, m ' it is that a pair in supergraph picture, subgraph match
Characteristic point;
(d) neighborhood image in supergraph picture, subgraph near characteristic point is obtained, and is calculated in supergraph picture in neighborhood image
Gray averageGray average in the gray value I of subgraph, subgraph in neighborhood image
(e) gray scale similarity S is calculated:
Wherein, (x, y) is m point coordinates, and (x ', y ') is m ' coordinates;
(f) gradient of disparity d is calculatedgr:
dgr=| da-db|/|dcs(am,bm)|
Wherein, daFor the coordinate difference for the characteristic point that a pair match, dbThe coordinate difference of the characteristic point matched for another pair
Value, dcs(am, bm) for the vector at two pairs of match point line midpoints;
(g) according to the gray scale similarity, gradient of disparity calculated, false matches value is removed, remaining match point is brought into pole
In line restriction relation, three-dimensional matching is carried out.
The problem of for gray scale in the prior art and gradient of disparity interference three-dimensional matching precision, the present invention proposes that one kind is used for
Two video cameras are demarcated, obtain outer parameter matrix R, translation vector by the three-dimensional matching method for building up of disparity computation first
The respective Intrinsic Matrix A of t and two video cameral、Ar.Wherein it is using any existing method to the method for camera calibration
Can, it will not be described here.The image that two video cameras of setting are obtained is respectively supergraph picture, subgraph, sets up two video cameras
Fundamental matrix relation:F=Ar -T[t]×RAl -1;Afterwards in testee surface selected characteristic point, epipolar-line constraint relation is set up:m
’TFm=0.Neighborhood image in supergraph picture, subgraph near characteristic point is obtained by graphical analysis, and calculated in supergraph picture
Gray average in neighborhood imageGray average in the gray value I of subgraph, subgraph in neighborhood imageBring into m,
The two-dimensional coordinate that 2 points of m ', the calculation formula for calculating gray scale similarity S, S is as follows:
Gradient of disparity is calculated again:
dgr=| da-db|/|dcs(am,bm)|
After the gray scale similarity that calculates, gradient of disparity, need to exclude that gray scale similarity is too low, regard according to measurement accuracy
The excessive value of poor gradient, remaining match point is brought into epipolar-line constraint relation, can carry out three-dimensional matching.The present invention is compared to biography
The three-dimensional matching process of system, overcomes gray scale, the error that gradient of disparity is brought, and eliminates false matches, being capable of high degree
Upper raising measurement accuracy.
It is preferred that, match point of the gray scale similarity less than 0.8 is removed as false matches value.
It is preferred that, the match point that gradient of disparity is more than 0.2 is removed as false matches value.
It is preferred that, step (e) and (f) order are adjustable.
The present invention compared with prior art, has the following advantages and advantages:
Three-dimensional matching method for building up of the present invention for disparity computation, overcomes gray scale, the error that gradient of disparity is brought,
False matches are eliminated, measurement accuracy can be largely improved.
Brief description of the drawings
Accompanying drawing described herein is used for providing further understanding the embodiment of the present invention, constitutes one of the application
Point, do not constitute the restriction to the embodiment of the present invention.In the accompanying drawings:
Fig. 1 is the schematic flow sheet of the specific embodiment of the invention.
Embodiment
For the object, technical solutions and advantages of the present invention are more clearly understood, with reference to embodiment and accompanying drawing, to this
Invention is described in further detail, and exemplary embodiment and its explanation of the invention is only used for explaining the present invention, does not make
For limitation of the invention.
Embodiment 1:
As shown in Figure 1 is used for the three-dimensional matching method for building up of disparity computation, comprises the following steps:(a) respectively to two
Video camera is demarcated, and obtains outer parameter matrix R, translation vector t and the respective Intrinsic Matrix A of two video camerasl、Ar;
(b) image for setting two video cameras acquisitions is respectively supergraph picture, subgraph, sets up the fundamental matrix relation of two video cameras:F
=Ar -T[t]×RAl -1;Wherein T is female image intensity value;(c) epipolar-line constraint relation is set up:m’TFm=0;Wherein m, m ' for mother
The characteristic point matched for a pair in image, subgraph;(d) neighborhood image in supergraph picture, subgraph near characteristic point is obtained,
And calculate the gray average in supergraph picture in neighborhood imageAsh in the gray value I of subgraph, subgraph in neighborhood image
Spend average(e) gray scale similarity S is calculated:
Wherein, (x, y) is m point coordinates, and (x ', y ') is m ' coordinates;(f) gradient of disparity d is calculatedgr:
dgr=| da-db|/|dcs(am,bm)|
Wherein, daFor the coordinate difference for the characteristic point that a pair match, dbThe coordinate difference of the characteristic point matched for another pair
Value, dcs(am, bm) for the vector at two pairs of match point line midpoints;(g) according to the gray scale similarity, gradient of disparity calculated, go
Except false matches value, remaining match point is brought into epipolar-line constraint relation, three-dimensional matching is carried out.Wherein, gray scale similarity is less than
0.8 match point is removed as false matches value.The match point that gradient of disparity is more than 0.2 is removed as false matches value.
Above-described embodiment, has been carried out further to the purpose of the present invention, technical scheme and beneficial effect
Describe in detail, should be understood that the embodiment that the foregoing is only the present invention, be not intended to limit the present invention
Protection domain, within the spirit and principles of the invention, any modification, equivalent substitution and improvements done etc. all should be included
Within protection scope of the present invention.
Claims (4)
1. the three-dimensional matching method for building up for disparity computation, it is characterised in that comprise the following steps:
(a) two video cameras are demarcated respectively, obtains outer parameter matrix R, translation vector t and two video cameras each
Intrinsic Matrix Al、Ar;
(b) image for setting two video cameras acquisitions is respectively supergraph picture, subgraph, and the fundamental matrix for setting up two video cameras is closed
System:F=Ar -T[t]×RAl -1;Wherein T is female image intensity value;
(c) epipolar-line constraint relation is set up:m’TFm=0;Wherein m, m ' it is the feature matched for a pair in supergraph picture, subgraph
Point;
(d) neighborhood image in supergraph picture, subgraph near characteristic point is obtained, and calculates the ash in supergraph picture in neighborhood image
Spend averageGray average in the gray value I of subgraph, subgraph in neighborhood image
(e) gray scale similarity S is calculated:
<mrow>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<msup>
<mi>y</mi>
<mo>&prime;</mo>
</msup>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>n</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<msup>
<mi>x</mi>
<mo>&prime;</mo>
</msup>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<mrow>
<mo>(</mo>
<mi>T</mi>
<mo>(</mo>
<mrow>
<msup>
<mi>x</mi>
<mo>&prime;</mo>
</msup>
<mo>,</mo>
<msup>
<mi>y</mi>
<mo>&prime;</mo>
</msup>
</mrow>
<mo>)</mo>
<mo>-</mo>
<mover>
<mi>T</mi>
<mo>&OverBar;</mo>
</mover>
<mo>)</mo>
</mrow>
<mrow>
<mo>(</mo>
<mi>I</mi>
<mo>(</mo>
<mrow>
<mi>x</mi>
<mo>+</mo>
<msup>
<mi>x</mi>
<mo>&prime;</mo>
</msup>
<mo>,</mo>
<mi>y</mi>
<mo>+</mo>
<msup>
<mi>y</mi>
<mo>&prime;</mo>
</msup>
</mrow>
<mo>)</mo>
<mo>-</mo>
<mover>
<mi>I</mi>
<mo>&OverBar;</mo>
</mover>
<mo>)</mo>
</mrow>
</mrow>
Wherein, (x, y) is m point coordinates, and (x ', y ') is m ' coordinates;
(f) gradient of disparity d is calculatedgr:
dgr=| da-db|/|dcs(am,bm)|
Wherein, daFor the coordinate difference for the characteristic point that a pair match, dbThe coordinate difference of the characteristic point matched for another pair,
dcs(am, bm) for the vector at two pairs of match point line midpoints;
(g) according to the gray scale similarity, gradient of disparity calculated, false matches value is removed, remaining match point is brought into polar curve about
In beam relation, three-dimensional matching is carried out.
2. the three-dimensional matching method for building up according to claim 1 for disparity computation, it is characterised in that gray scale similarity
Match point less than 0.8 is removed as false matches value.
3. the three-dimensional matching method for building up according to claim 1 for disparity computation, it is characterised in that gradient of disparity is big
Match point in 0.2 is removed as false matches value.
4. the three-dimensional matching method for building up according to claim 1 for disparity computation, it is characterised in that step (e) and
(f) order is adjustable.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710427806.2A CN107248179A (en) | 2017-06-08 | 2017-06-08 | Three-dimensional matching method for building up for disparity computation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710427806.2A CN107248179A (en) | 2017-06-08 | 2017-06-08 | Three-dimensional matching method for building up for disparity computation |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107248179A true CN107248179A (en) | 2017-10-13 |
Family
ID=60017881
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710427806.2A Pending CN107248179A (en) | 2017-06-08 | 2017-06-08 | Three-dimensional matching method for building up for disparity computation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107248179A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116797463A (en) * | 2023-08-22 | 2023-09-22 | 佗道医疗科技有限公司 | Feature point pair extraction method and image stitching method |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101887589A (en) * | 2010-06-13 | 2010-11-17 | 东南大学 | Stereoscopic vision-based real low-texture image reconstruction method |
CN103310421A (en) * | 2013-06-27 | 2013-09-18 | 清华大学深圳研究生院 | Rapid stereo matching method and disparity map obtaining method both aiming at high-definition image pair |
CN104268880A (en) * | 2014-09-29 | 2015-01-07 | 沈阳工业大学 | Depth information obtaining method based on combination of features and region matching |
CN104680510A (en) * | 2013-12-18 | 2015-06-03 | 北京大学深圳研究生院 | RADAR parallax image optimization method and stereo matching parallax image optimization method and system |
CN106183995A (en) * | 2016-07-26 | 2016-12-07 | 武汉大学 | A kind of visual parking device method based on stereoscopic vision |
CN106228605A (en) * | 2016-07-29 | 2016-12-14 | 东南大学 | A kind of Stereo matching three-dimensional rebuilding method based on dynamic programming |
CN106600686A (en) * | 2016-12-06 | 2017-04-26 | 西安电子科技大学 | Three-dimensional point cloud reconstruction method based on multiple uncalibrated images |
CN106780442A (en) * | 2016-11-30 | 2017-05-31 | 成都通甲优博科技有限责任公司 | A kind of solid matching method and system |
-
2017
- 2017-06-08 CN CN201710427806.2A patent/CN107248179A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101887589A (en) * | 2010-06-13 | 2010-11-17 | 东南大学 | Stereoscopic vision-based real low-texture image reconstruction method |
CN103310421A (en) * | 2013-06-27 | 2013-09-18 | 清华大学深圳研究生院 | Rapid stereo matching method and disparity map obtaining method both aiming at high-definition image pair |
CN104680510A (en) * | 2013-12-18 | 2015-06-03 | 北京大学深圳研究生院 | RADAR parallax image optimization method and stereo matching parallax image optimization method and system |
CN104268880A (en) * | 2014-09-29 | 2015-01-07 | 沈阳工业大学 | Depth information obtaining method based on combination of features and region matching |
CN106183995A (en) * | 2016-07-26 | 2016-12-07 | 武汉大学 | A kind of visual parking device method based on stereoscopic vision |
CN106228605A (en) * | 2016-07-29 | 2016-12-14 | 东南大学 | A kind of Stereo matching three-dimensional rebuilding method based on dynamic programming |
CN106780442A (en) * | 2016-11-30 | 2017-05-31 | 成都通甲优博科技有限责任公司 | A kind of solid matching method and system |
CN106600686A (en) * | 2016-12-06 | 2017-04-26 | 西安电子科技大学 | Three-dimensional point cloud reconstruction method based on multiple uncalibrated images |
Non-Patent Citations (1)
Title |
---|
李健等: "《基于极线几何的改进多约束图像立体匹配》", 《微型机与应用》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116797463A (en) * | 2023-08-22 | 2023-09-22 | 佗道医疗科技有限公司 | Feature point pair extraction method and image stitching method |
CN116797463B (en) * | 2023-08-22 | 2023-11-21 | 佗道医疗科技有限公司 | Feature point pair extraction method and image stitching method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111260597B (en) | Parallax image fusion method of multiband stereo camera | |
CN107977997B (en) | Camera self-calibration method combined with laser radar three-dimensional point cloud data | |
CN102065313B (en) | Uncalibrated multi-viewpoint image correction method for parallel camera array | |
CN109919911B (en) | Mobile three-dimensional reconstruction method based on multi-view photometric stereo | |
CN113034568B (en) | Machine vision depth estimation method, device and system | |
CN102665086B (en) | Method for obtaining parallax by using region-based local stereo matching | |
CN111210481A (en) | Depth estimation acceleration method of multiband stereo camera | |
CN108020175B (en) | multi-grating projection binocular vision tongue surface three-dimensional integral imaging method | |
CN109840922B (en) | Depth acquisition method and system based on binocular light field camera | |
CN103945207B (en) | A kind of stereo-picture vertical parallax removing method based on View Synthesis | |
CN109712232B (en) | Object surface contour three-dimensional imaging method based on light field | |
CN106340045B (en) | Calibration optimization method in three-dimensional facial reconstruction based on binocular stereo vision | |
CN111091076B (en) | Tunnel limit data measuring method based on stereoscopic vision | |
CN104778716B (en) | Lorry compartment volume measuring method based on single image | |
CN104167001B (en) | Large-visual-field camera calibration method based on orthogonal compensation | |
CN106203429A (en) | Based on the shelter target detection method under binocular stereo vision complex background | |
CN102914295A (en) | Computer vision cube calibration based three-dimensional measurement method | |
CN109285189A (en) | A kind of straight path quick calculation method synchronous without binocular | |
CN107909543A (en) | A kind of flake binocular vision Stereo matching space-location method | |
CN115880344A (en) | Binocular stereo matching data set parallax truth value acquisition method | |
CN111382591A (en) | Binocular camera ranging correction method and vehicle-mounted equipment | |
CN113643427A (en) | Binocular ranging and three-dimensional reconstruction method | |
CN102842118A (en) | New robust stereopair correcting method | |
CN107248179A (en) | Three-dimensional matching method for building up for disparity computation | |
CN110487254B (en) | Rapid underwater target size measuring method for ROV |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171013 |