CN104408772A - Grid projection-based three-dimensional reconstructing method for free-form surface - Google Patents
Grid projection-based three-dimensional reconstructing method for free-form surface Download PDFInfo
- Publication number
- CN104408772A CN104408772A CN201410647810.6A CN201410647810A CN104408772A CN 104408772 A CN104408772 A CN 104408772A CN 201410647810 A CN201410647810 A CN 201410647810A CN 104408772 A CN104408772 A CN 104408772A
- Authority
- CN
- China
- Prior art keywords
- point
- grid
- intersection point
- form surface
- dimensional
- 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
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/30—Polynomial surface description
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Physics (AREA)
- Mathematical Analysis (AREA)
- Algebra (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a grid projection-based three-dimensional reconstructing method for a free-form surface. There is no regular texture and certain feature point at the surface of the free-form surface, and the positions and number of feature points extracted through an existing three-dimensional reconstructing algorithm for the free-form surface are uncertain due to the surface reflectivity difference, local luminance difference of an image, image noise and the like in the running process. In order to solve the problems, grids are projected to the surface of a curved surface to be reconstructed to enable the free-form surface to have certain identifiable texture and features; on this basis, surface features of the free-form surface are extracted to realize precise feature point matching and reconstructing and convert the three-dimensional reconstruction of the free-form surface into matching and reconstruction through identifying the grid projection features of the surface of the free-form surface. The three-dimensional reconstructing for an entity model verifies that the method is feasible and effective.
Description
Technical field
The present invention relates to a kind of free form surface three-dimensional rebuilding method, particularly relate to the free form surface three-dimensional rebuilding method based on Grid Projection.
Background technology
Stereoscopic vision is computing machine always
figurethe important research problem of shape and computer vision carries out perception or identification according to the principle of mankind's binocular vision to scene or target, namely utilizes mathematical method from two width or several two dimension solids
figurethe three-dimensional information of picture centering restoration scenario or target.Current stereoscopic vision is at industrial circle---and product quality detects and external packing detects, the industrial flaw detection of the metal product that dispatches from the factory; At medical domain---Modling model before medical operating, carries out diagnosing and providing therapeutic scheme for assist physician; In archaeology field---digitized processing is carried out to historical relic, realizes multiple fields such as cultural heritage protection and be widely used.Reaching its maturity along with 3D printing technique simultaneously, meeting future by the prerequisite of 3D printing technique quick copy object is also ripe convenient three-dimensional reconstruction technique.
Stereoscopic vision, according to the difference of measuring method, can be divided into Active measuring and passive measurement.Active measuring, as laser scanning, structure light scan, the same time takes from different perspectives to target, can rebuild high-precision model.Passive method is from several based on multiple view geometry principle
figuretwo-dimensional signal reconstruction model is extracted in picture.Current method for three-dimensional measurement mainly comprises phase measurement consistency profiles, network optical mode etc., uses these methods can recover the three-dimensional information of regular shape, the obvious scene of feature or target.But phase measurement consistency profiles exists to be forbidden due to phase shift and the non-sinusoidal of light field can both produce the problem of measuring error, when therefore measuring, just need the sinusoidal grating introducing the high phase changer of precision comparison and standard; Network optical mode is also called trellis coding scheme, the impact of unique point in its Existential Space neighborhood, so coding
figurethe identification of picture and decoding more difficult, and the measuring error of system is increased, also has resolution low and be subject to the shortcomings such as the not identical impact with color of scenery surface reflectivity.
Therefore, in order to solve this kind of problem, herein in three-dimensional reconstruction process, passive measurement method is combined with Active measuring method, adding artificial restraint condition, is the surface be made up of regular grid by random or without texture surface information unification, to binocular in spatial domain
figurerecover three-dimensional information after picture process, reduce the uncertainty in three-dimensional reconstruction process, improve reconstructed results precision.Context of methods is lower compared to the method for reconstructing such as laser scanning, phase method cost, and system architecture is simple, does not need to do any pre-service to experimental subjects, can produce a desired effect.
Summary of the invention
(1) technical matters that will solve
This
inventionobject be to overcome the deficiencies in the prior art, provide a kind of free form surface three-dimensional rebuilding method based on Grid Projection, the method reduces the uncertainty in three-dimensional reconstruction process, improves reconstructed results precision.
(2) technical scheme
For above problem, this
inventionthe free form surface three-dimensional rebuilding method based on Grid Projection proposed, comprises the following steps:
1) design alternative system hardware, builds binocular vision experiment table, gathers experimental subjects
figuresheet;
2) by step 1) in obtain two width
figurepicture carries out distortion correction and various pretreatment operation, and suitable preprocess method is selected in comparative analysis by experiment, is the later stage
figurethe process of picture is ready;
3) by step 2) gained two width
figurepicture utilizes morphological method to carry out mesh lines thinning;
4) by step 3) gained two width
figurepicture extracts the unique point between grid intersection point and grid intersection point.According to pixel distribution feature in the distribution characteristics of grid intersection point and four neighborhoods and eight neighborhood thereof, at each intersection point position extract minutiae, final retain be positioned at position, bosom unique point as grid intersection point.According to single pixel characteristic of grid lines after grid intersection point and refinement, extract all unique points between grid intersection point.
5) by step 4) gained two width
figurepicture unique point, grid intersection point carries out secondary coding according to the position of intersection point to it, and row-coordinate point is within the specific limits classified as same a line, and row coordinate point is within the specific limits classified as same row, by two width after coding
figureit is right that the intersection point having a same coding in picture is classified as coupling.Unique point between grid intersection point is mated according to epipolar geom etry constraint, similarity constraint, continuity constraint, Ordinal Consistency constraint;
6) by step 5) gained two width
figurepicture Feature Points Matching pair, according to the unique point of matching double points separately
figuretwo-dimensional coordinate in picture, the inner parameter of binocular vision camera and external parameter and stereo vision three-dimensional rebuilding formula, calculate the three-dimensional coordinate of unique point, and complete reconstruction.
(3) beneficial effect
This
inventionthe free form surface three-dimensional rebuilding method based on Grid Projection can be used on curved surface well, test and rebuild for mouse surface, achieve reasonable experiment effect, there is versatility.
Accompanying drawing explanation
fig. 1for this
invention is realexecute the left and right camera of example 1
figurepicture gray processing treatment effect
figure, wherein, (a) for left camera original
figurepicture (after distortion correction), (b) for right camera original
figurepicture (after distortion correction);
fig. 2for this
invention is realexecute the left and right camera of example 1
figureeffect is completed as pre-service
figure, wherein, (a) completes for left camera pre-service
figure, (b) completes for right camera pre-service
figure;
fig. 3for this
invention is realexecute the grid intersection point extraction effect of example 1
figure, wherein, (a) is left camera extraction intersection point, and (b) is that right camera extracts intersection point;
fig. 4for this
invention is realfeature point extraction effect between the grid intersection point executing example 1
figure, wherein, (a) is left
figurevertical direction unique point, (b) is right
figurevertical direction unique point, (c) is left
figurehorizontal direction unique point, (d) is right
figurehorizontal direction unique point;
fig. 5for this
invention is realexecute the Feature Points Matching effect of example 1
figure, wherein, (a) is vertical direction Feature Points Matching result, and (b) is horizontal direction Feature Points Matching result;
fig. 6for this
invention is realexecute the Feature Points Matching flow process of example 1
figure;
fig. 7for this
invention is realexecute example 1 three-dimensional reconstruction effect
figure;
Embodiment
This
inventionthe free form surface three-dimensional rebuilding method based on Grid Projection proposed combines
accompanying drawingbe described as follows with embodiment, following embodiment only for illustration of this development, and is not to this
inventionrestriction, this
inventionscope of patent protection should be limited by each claim.
This
inventin three-dimensional reconstruction process, passive measurement method being combined with Active measuring method, add artificial restraint condition, is the surface be made up of regular grid by random or without texture surface information unification, to binocular in spatial domain
figurerecover three-dimensional information after picture process, reduce the uncertainty in three-dimensional reconstruction process, improve reconstructed results precision.
This
inventionbe research object with mouse, study and which kind of algorithm more accurately can carry out three-dimensional reconstruction to curved surface with.
Embodiment
The three-dimensional reconstruction system built herein is made up of CCD camera, Grid Projection, binocular camera support, computing machine and system software etc.Left and right two video cameras adopt any attitude of non-parallel optical axis to place, and place Grid Projection in two video camera line midpoint.During measurement, Grid Projection to measured surface, and forms latticed light bar on measured surface, CCD camera collection measure two original
figurepicture,
as Fig. 1.
To video camera obtain original
figureafter picture carries out the pre-service such as distortion correction, denoising, binaryzation, tessellated mesh lines, only can there be is gridding information
figureas right,
as Fig. 2shown in.After Grid Projection, extraction grid lines, by the target of random feature, random texture
figurepicture is converted to the unified grid lines of texture, by processing grid lines, can obtain the textural characteristics of rule to any Free-form Surface Reconstruction target.Contrast original
figurepicture and pre-service
figurepicture can be seen, by original
figureafter in picture in algorithm with less than information reject, at utmost improve algorithm travelling speed.
For
fig. 2carry out feature extraction, follow the principle from point to line, namely first extract angle point, the intersection point of high, the easy identification of accuracy rate, then according to these angle points, intersection point extraction unique point therebetween.First is extracted target is mesh lines intersection point.In extraction mesh lines intersection point process, owing to deforming after Grid Projection to target surface, the mesh lines intersection region after refinement has trickle difference.We need to take into account different grid intersection features, extract the intersection point of all mesh lines.To multiple grids of different experiments object shooting
figurecarry out observing, analyze and adding up as intersection point and neighborhood territory pixel feature thereof.Pixel quantity in four neighborhoods of Research statistics intersection point place pixel and eight neighborhood, can be summarized as 5 types.The unique point of described 5 types can cover most grid intersection points.If there is extremely indivedual grid intersection not meet any one type above, then this kind of point of crossing is designated as vacancy point.Because pixel distribution situation near grid intersection is complicated, the unique point of multiple the above-mentioned type can be extracted near same point of crossing.For this reason, arrange prioritization to the intersection point of every type, according to said method can ensure the unique point after extracting and point of crossing one_to_one corresponding, the final unique point quantity extracted is equal to or less than actual grid point of crossing quantity,
as Fig. 3shown in.
According to grid intersection point, extract the unique point between grid intersection point further.This part unique point is divided into two classes, and a class is the unique point near vertical directivity curve section, and a class is the unique point in level of approximation directivity curve section.?
figurein picture, because mesh lines has been refined as single pixel lines, therefore according to the grid intersection point extracted, the pixel coordinate between all grid intersection points can be obtained.Feature point extraction result between grid
as Fig. 4shown in.
According to feature point extraction process, unique point is divided into grid intersection point, the unique point of near vertical directivity curve section and the unique point of level of approximation directivity curve section, Feature Points Matching process also will be carried out according to these three kinds of classification, flow process
as Fig. 6shown in.
For grid intersection point, after extracting grid intersection point, intersection point can be obtained and exist
figuretwo-dimensional coordinate (x, y) in picture.Intersection point can be obtained according to the two-dimensional coordinate of intersection point to exist
figurepositional information in picture.Secondary coding is carried out to it in position according to intersection point by us, and row-coordinate point is within the specific limits classified as same a line, and row coordinate point is within the specific limits classified as same row.With
fig. 2in grid intersection point be example, will
in figure25 points are encoded to (1 successively according to a position, 1), (1,2) ... (5,4), (5,5), the intersection point marked price arranged as the 1st row the 1st is for (1,1), the intersection point that 1st row the 2nd arranges is labeled as (1,2), by that analogy.
In the matching process, grid intersection point may be had and fail and extract the problem of grid intersection point vacancy of causing, if consistency constraint proceeds coupling in order, intersection point couplings all after this intersection point can be caused
mistake.Therefore, using the mean value of the spacing according to two intersection points as threshold value, if row-coordinate or row coordinate are greater than 1.5 times of this threshold value between two intersection points, then think and between 2, to have point of crossing to fail extraction, coding between 2 o'clock is done vacancy process, avoids occurring error in ensuing matching process.Will
figureafter the grid intersection difference secondary coding of sheet centering, exact matching can be carried out.The unique point with corresponding secondary coding is mated, such as, by a left side
figurethe unique point of (1,1) is encoded to right in picture
figure(1,1) point in picture mates, by a left side
figurethe unique point of (1,2) is encoded to right in picture
figure(1,2) point in picture mates, and by that analogy, mates all grid intersection points.In the matching process, grid intersection point may be had and fail and extract the problem of grid intersection point vacancy of causing, if consistency constraint proceeds coupling in order, intersection point couplings all after this intersection point can be caused
mistake.Therefore, using the mean value of the spacing according to two intersection points as threshold value, if row-coordinate or row coordinate are greater than 1.5 times of this threshold value between two intersection points, then think and between 2, to have point of crossing to fail extraction, coding between 2 o'clock is done vacancy process, avoids occurring error in ensuing matching process.Will
figureafter the grid intersection difference secondary coding of sheet centering, exact matching can be carried out.The unique point with corresponding secondary coding is mated, such as, by a left side
figurethe unique point of (1,1) is encoded to right in picture
figure(1,1) point in picture mates, by a left side
figurethe unique point of (1,2) is encoded to right in picture
figure(1,2) point in picture mates, and by that analogy, mates all grid intersection points.
For near vertical direction and level of approximation direction grid intersection point between unique point, carry out constrained matching according to constraint conditions such as epipolar geom etry constraint, similarity constraint, continuity constraint and Ordinal Consistency constraints.First right
figurepicture is to carrying out polar curve correction, and eliminate the distance in y direction between corresponding match point point, namely corresponding match point is positioned on same level line.So, then can linear search on this horizontal line when searching for corresponding match point, and do not need whole
figuretwo-dimensional search is carried out in picture.In linear search process, corresponding match point necessarily appears in corresponding grid lines, as long as therefore search for the point be positioned on these grid lines on this horizontal line can find corresponding match point.Such as, search is left
figureon picture the 1st row mesh lines, certain point to be matched is on the right side
figurecorresponding point in picture, find the horizontal line at this place, determine the y coordinate of corresponding point; Then find the right side
figurepicture the 1st row mesh lines, searches for the horizontal intersection point of this mesh lines and this place and can be defined as corresponding point.After the same method, by the right side
figurepicture in point as point to be matched, on a left side
figurecorresponding point are searched in picture.The Corresponding matching point that twice search obtains to
figurenamely distance in picture thinks correct corresponding point in certain threshold value, thus sets up matching relationship.In like manner, for the unique point in level of approximation direction, similarity constraint, continuity constraint and Ordinal Consistency is utilized to find corresponding point.Due to projection relation, the unique point on level of approximation segment of curve can not set up relation one to one, namely left
figureeach point in picture not necessarily can on the right side
figurecorresponding point is found to match in picture.Left and right
figurepicture Feature Points Matching result
as Fig. 5shown in.
According to the unique point of matching double points separately
figuretwo-dimensional coordinate in picture, the inner parameter of binocular vision camera and external parameter and stereo vision three-dimensional rebuilding formula, obtain the three-dimensional coordinate of matching double points.According to three-dimensional coordinate, fitting surface obtains three-dimensional model,
as Fig. 7.
in figureblue round dot is the unique point after rebuilding, the change of the change reflection curved surface Z axis coordinate of curved surface color.
Claims (5)
1., based on a free form surface three-dimensional rebuilding method for Grid Projection, comprise the Preprocessing method to Grid Projection rear surface image, Feature Points Extraction, characteristic point matching method, three-dimensional rebuilding method.
2. the free form surface three-dimensional rebuilding method based on Grid Projection according to claim 1, is characterized in that the smoothing method in real domain is carried out detailed analysis and experiment by pretreatment stage, and improves medium filtering, adopts adaptive median filter.To filtered image, adopt morphological method to extract grid lines after binarization, and grid lines are carried out single pixel refinement.
3. the free form surface three-dimensional rebuilding method based on Grid Projection according to claim 1, is characterized in that comprising two parts in feature extraction phases, feature point extraction between grid intersection point feature point extraction and grid intersection point.It is characterized in that:
1) feature extracting grid intersection point unique point has added up the pixel quantity in four neighborhoods of intersection point place pixel and eight neighborhood, and can be summarized as 5 types.Extract intersection point respectively according to this 5 type, retain a most suitable unique point at each position of intersecting point, ensure that each position of intersecting point only has a unique point.
2) between grid intersection point, the feature of feature point extraction is according to grid Intersection Point and single pixel lines two constraint conditions, extracts all unique points be positioned between grid Intersection Point.
4., according to claim 1 based on the free form surface three-dimensional rebuilding method of Grid Projection, the matching process of unique point comprises two parts, and Feature Points Matching between grid intersection point Feature Points Matching and grid intersection point, is characterized in that:
1) grid intersection point coupling, is carry out secondary coding according to the position of intersection point to it, row-coordinate point is within the specific limits classified as same a line, row coordinate point is within the specific limits classified as same row.
2) Feature Points Matching between grid, being according to epipolar geom etry constraint, Ordinal Consistency constraint, take piece image as benchmark, search for corresponding match point in same a line of another piece image after, with searched image for benchmark, in former benchmark image, search for corresponding match point.According to match check consistance, determine final matching double points.
5. the free form surface three-dimensional rebuilding method based on Grid Projection according to claim 1, it is characterized in that: according to the two-dimensional coordinate of unique point in respective image of matching double points, the inner parameter of binocular vision camera and external parameter and stereo vision three-dimensional rebuilding formula, calculate the three-dimensional coordinate of unique point, and complete reconstruction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410647810.6A CN104408772A (en) | 2014-11-14 | 2014-11-14 | Grid projection-based three-dimensional reconstructing method for free-form surface |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410647810.6A CN104408772A (en) | 2014-11-14 | 2014-11-14 | Grid projection-based three-dimensional reconstructing method for free-form surface |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104408772A true CN104408772A (en) | 2015-03-11 |
Family
ID=52646400
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410647810.6A Pending CN104408772A (en) | 2014-11-14 | 2014-11-14 | Grid projection-based three-dimensional reconstructing method for free-form surface |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104408772A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105928484A (en) * | 2016-03-28 | 2016-09-07 | 江南大学 | Elevator guide rail automatic measurement system based on binocular vision |
| CN107330906A (en) * | 2017-06-28 | 2017-11-07 | 江南大学 | Improvement thinning algorithm based on curve matching |
| CN107992868A (en) * | 2017-11-15 | 2018-05-04 | 辽宁警察学院 | A kind of High Precision Stereo footprint Quick Acquisition method |
| CN109725640A (en) * | 2018-12-18 | 2019-05-07 | 天津理工大学 | Three-dimensional road method for reconstructing in a kind of automatic Pilot based on stereoscopic vision and laser grating |
| CN109785431A (en) * | 2018-12-18 | 2019-05-21 | 天津理工大学 | A kind of road ground three-dimensional feature acquisition method and device based on laser network |
| CN110930344A (en) * | 2018-08-29 | 2020-03-27 | 杭州海康威视数字技术股份有限公司 | Target quality determination method, device and system and electronic equipment |
| CN111213066A (en) * | 2017-08-23 | 2020-05-29 | 洛桑联邦理工学院 | Image reconstruction method based on model |
| CN111652901A (en) * | 2020-06-02 | 2020-09-11 | 山东大学 | Texture-free three-dimensional object tracking method based on confidence coefficient and feature fusion |
| CN112837380A (en) * | 2021-02-05 | 2021-05-25 | 深圳市万物云科技有限公司 | Obstacle avoidance method and device based on laser network, computer equipment and storage medium |
| CN113655056A (en) * | 2021-08-20 | 2021-11-16 | 重庆交通大学 | Investigation and statistics method for river benthonic animals |
| CN117253012A (en) * | 2023-09-18 | 2023-12-19 | 东南大学 | Method for restoring plane building free-form surface grid structure to three-dimensional space |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101866497A (en) * | 2010-06-18 | 2010-10-20 | 北京交通大学 | Binocular stereo vision based intelligent three-dimensional human face rebuilding method and system |
| CN102855665A (en) * | 2012-08-01 | 2013-01-02 | 北京天人同达软件科技有限公司 | Method for reconstructing three-dimensional building model from single image |
| CN103500468A (en) * | 2013-09-09 | 2014-01-08 | 电子科技大学 | Space curved surface reconstruction method based on boundary constraint |
-
2014
- 2014-11-14 CN CN201410647810.6A patent/CN104408772A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101866497A (en) * | 2010-06-18 | 2010-10-20 | 北京交通大学 | Binocular stereo vision based intelligent three-dimensional human face rebuilding method and system |
| CN102855665A (en) * | 2012-08-01 | 2013-01-02 | 北京天人同达软件科技有限公司 | Method for reconstructing three-dimensional building model from single image |
| CN103500468A (en) * | 2013-09-09 | 2014-01-08 | 电子科技大学 | Space curved surface reconstruction method based on boundary constraint |
Non-Patent Citations (3)
| Title |
|---|
| 吴剑波等: "《光栅立体视觉三维表面重建***及其图像处理方法（一）》", 《计算机工程与应用》 * |
| 吴剑波等: "《光栅立体视觉三维表面重建***及其图像处理方法（二）》", 《计算机工程与应用》 * |
| 吴剑波等: "《网格投影式立体视觉三维表面重建***》", 《光电工程》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105928484B (en) * | 2016-03-28 | 2019-05-17 | 江南大学 | Cage guide automatic measurement system based on binocular vision |
| CN105928484A (en) * | 2016-03-28 | 2016-09-07 | 江南大学 | Elevator guide rail automatic measurement system based on binocular vision |
| CN107330906A (en) * | 2017-06-28 | 2017-11-07 | 江南大学 | Improvement thinning algorithm based on curve matching |
| US11175392B2 (en) | 2017-08-23 | 2021-11-16 | Ecole Polytechnique Federale De Lausanne (Epfl) | Model-based image reconstruction method |
| CN111213066A (en) * | 2017-08-23 | 2020-05-29 | 洛桑联邦理工学院 | Image reconstruction method based on model |
| CN111213066B (en) * | 2017-08-23 | 2021-11-12 | 洛桑联邦理工学院 | Image reconstruction method based on model |
| CN107992868A (en) * | 2017-11-15 | 2018-05-04 | 辽宁警察学院 | A kind of High Precision Stereo footprint Quick Acquisition method |
| CN110930344A (en) * | 2018-08-29 | 2020-03-27 | 杭州海康威视数字技术股份有限公司 | Target quality determination method, device and system and electronic equipment |
| CN110930344B (en) * | 2018-08-29 | 2023-05-05 | 杭州海康威视数字技术股份有限公司 | Target quality determination method, device and system and electronic equipment |
| CN109725640A (en) * | 2018-12-18 | 2019-05-07 | 天津理工大学 | Three-dimensional road method for reconstructing in a kind of automatic Pilot based on stereoscopic vision and laser grating |
| CN109785431A (en) * | 2018-12-18 | 2019-05-21 | 天津理工大学 | A kind of road ground three-dimensional feature acquisition method and device based on laser network |
| CN111652901A (en) * | 2020-06-02 | 2020-09-11 | 山东大学 | Texture-free three-dimensional object tracking method based on confidence coefficient and feature fusion |
| CN112837380A (en) * | 2021-02-05 | 2021-05-25 | 深圳市万物云科技有限公司 | Obstacle avoidance method and device based on laser network, computer equipment and storage medium |
| CN113655056A (en) * | 2021-08-20 | 2021-11-16 | 重庆交通大学 | Investigation and statistics method for river benthonic animals |
| CN113655056B (en) * | 2021-08-20 | 2024-05-03 | 重庆交通大学 | River benthonic animal investigation and statistics method |
| CN117253012A (en) * | 2023-09-18 | 2023-12-19 | 东南大学 | Method for restoring plane building free-form surface grid structure to three-dimensional space |
| CN117253012B (en) * | 2023-09-18 | 2024-03-19 | 东南大学 | Method for restoring plane building free-form surface grid structure to three-dimensional space |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104408772A (en) | Grid projection-based three-dimensional reconstructing method for free-form surface | |
| CN110264567B (en) | Real-time three-dimensional modeling method based on mark points | |
| CN106355570B (en) | A kind of binocular stereo vision matching method of combination depth characteristic | |
| CN108053476B (en) | Human body parameter measuring system and method based on segmented three-dimensional reconstruction | |
| Kamencay et al. | Improved Depth Map Estimation from Stereo Images Based on Hybrid Method. | |
| Pantilie et al. | SORT-SGM: Subpixel optimized real-time semiglobal matching for intelligent vehicles | |
| CN110487216A (en) | A kind of fringe projection 3-D scanning method based on convolutional neural networks | |
| US20060210146A1 (en) | Creating 3D images of objects by illuminating with infrared patterns | |
| CN104346608A (en) | Sparse depth map densing method and device | |
| CN104077804A (en) | Method for constructing three-dimensional human face model based on multi-frame video image | |
| CN105279789A (en) | A three-dimensional reconstruction method based on image sequences | |
| CN106155299B (en) | A kind of pair of smart machine carries out the method and device of gesture control | |
| CN105960570A (en) | Structured light matching of a set of curves from two cameras | |
| Zhao et al. | Double propagation stereo matching for urban 3-d reconstruction from satellite imagery | |
| Sanz et al. | Depth estimation-an introduction | |
| Liu et al. | The applications and summary of three dimensional reconstruction based on stereo vision | |
| CN111023994B (en) | Grating three-dimensional scanning method and system based on multiple measurement | |
| CN109816738A (en) | A kind of striped boundary extraction algorithm based on coded structured light | |
| Gonzalez-Aguilera et al. | From point cloud to CAD models: Laser and optics geotechnology for the design of electrical substations | |
| CN113313740A (en) | Disparity map and surface normal vector joint learning method based on plane continuity | |
| Ran et al. | High-precision human body acquisition via multi-view binocular stereopsis | |
| CN109712230B (en) | Three-dimensional model supplementing method and device, storage medium and processor | |
| Damjanović et al. | Local Stereo Matching Using Adaptive Local Segmentation: ISRN Machine Vision | |
| CN116843829A (en) | Concrete structure crack three-dimensional reconstruction and length quantization method based on binocular video | |
| CN114998532B (en) | Three-dimensional image visual transmission optimization method based on digital image reconstruction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150311 |
|
| WD01 | Invention patent application deemed withdrawn after publication |