CN108305286A - Multi-view stereo vision foot type method for three-dimensional measurement, system and medium based on color coding - Google Patents
Multi-view stereo vision foot type method for three-dimensional measurement, system and medium based on color coding Download PDFInfo
- Publication number
- CN108305286A CN108305286A CN201810074452.2A CN201810074452A CN108305286A CN 108305286 A CN108305286 A CN 108305286A CN 201810074452 A CN201810074452 A CN 201810074452A CN 108305286 A CN108305286 A CN 108305286A
- Authority
- CN
- China
- Prior art keywords
- foot type
- point
- characteristic point
- foot
- parameter
- 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
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/168—Segmentation; Edge detection involving transform domain methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/08—Indexing scheme for image data processing or generation, in general involving all processing steps from image acquisition to 3D model generation
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20048—Transform domain processing
- G06T2207/20061—Hough transform
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Computer Graphics (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a kind of multi-view stereo vision foot type method for three-dimensional measurement encoded based on color, system and medium, methods to include：By the acquisition platform of more mesh images, it is installed with the foot type of coloud coding pattern socks using polyphaser sync pulse jamming, carries out Image Acquisition；Polyphaser calibration algorithm according to the image collected, and based on sphere obtains camera inside and outside parameter；The matching of foot type dense feature point is carried out according to camera inside and outside parameter；It is matched according to camera inside and outside parameter and foot type dense feature point as a result, carrying out model construction and foot type parameter measurement.The present invention has many advantages, such as that at low cost, characteristic point is readily identified and matches, the point cloud density of generation is higher, calculating speed is fast, obtains accurately and completely foot model and relative dimensions parameter.
Description
Technical field
The present invention relates to foot shape measurement technical field more particularly to a kind of multi-view stereo vision foot type based on color coding
Method for three-dimensional measurement, system and medium.
Background technology
Currently, the tether-free technologies for being usually used in three-dimensional foot type measuring have：Based on laser scanning rebuild method, based on knot
The method that structure light is rebuild and the method rebuild based on stereoscopic vision.Foot type three-dimension measuring system in the market is based on laser scanning more
The method of reconstruction is formed closure plane using line laser and is scanned to foot, while being imaged using video camera, obtains foot type in light
It is movable in a vertical direction with the surface information on section, then by motor, obtains the threedimensional model of foot type.Based on structured light reconstruction
Measurement method by contain object dimensional information structure light obtain testee shape.Three based on stereoscopic vision
Measurement method is tieed up since it is low relative to laser scanning reconstructed cost, the reconstruction of opposed configuration light is not easy to be influenced by ambient light；And
It is fast to acquire image, advantageously reduces the error brought by shake.
But there is also defects for existing three-dimensional foot type measuring method：
Based on the method that laser scanning is rebuild, forms closure plane using laser and foot is scanned, obtain the three-dimensional mould of foot type
Type.This method sample frequency is high, and reconstruction model is fine and smooth, but its acquisition time is longer, and people's foot is easy to happen shake；And its price
Costliness, there are cost bottlenecks, it is difficult to promote.Such as Shoemaster, the foot type scanning system price such as Canfit-Plus CAD/CAM
It is sufficiently expensive.
Measurement method cost based on structured light reconstruction is relatively low, and technical maturity is medium, however this method be easy by
The influence of ambient light need to be placed under the soft environment of light and use, limit its scope of application.
Measurement method based on stereoscopic vision reconstruction is because its is at low cost, and acquisition image is quick, by more and more research people
Member's concern.But the reliability of the parameter measured is relatively low compared with the method for directly reconstructing to obtain model.Based on spot at random
The method of reconstruction is because its characteristic point is sparse, therefore it is also sparse to rebuild obtained model.Meanwhile the survey rebuild based on stereoscopic vision
Amount method is required to accurate camera parameter, and when gridiron pattern standardization is applied to polyphaser distributed more widely, camera parameter holds
Larger accumulated error is easily introduced during relation transmission.
Invention content
It is unrestricted that the present invention provides a kind of readily identified accurate, at low cost, characteristic point and matching, application range of measuring
Multi-view stereo vision foot type method for three-dimensional measurement, system and the medium based on color coding based on color coding.
To achieve the above object, the present invention provides a kind of multi-view stereo vision foot type three-dimensional measurement side encoded based on color
Method includes the following steps：
By the acquisition platform of more mesh images, the foot type of coloud coding pattern socks is installed with using polyphaser sync pulse jamming,
Carry out Image Acquisition；
Polyphaser calibration algorithm according to the image collected, and based on sphere obtains camera inside and outside parameter；
The matching of foot type dense feature point is carried out according to the camera inside and outside parameter；
It is matched according to the camera inside and outside parameter and foot type dense feature point to join with foot type as a result, carrying out model construction
Number measures.
Wherein, the acquisition platform of more mesh images has 360 ° of visual angles, has uniform illumination environment, is based on using two
The script of socket realizes the polyphaser synchronous acquisition of image.
Wherein, described according to the image collected, and the polyphaser calibration algorithm based on sphere obtains camera inside and outside parameter
The step of include：
According to the image collected, the loop truss algorithm based on Hough transformation estimates projection coordinate and figure of the centre of sphere in image
The size of diameter as in；
Spherical surface Corner Detection is carried out, 2D coordinates are obtained, the 3D coordinates of mark point and the centre of sphere is calculated, estimates the initial outer ginseng of camera
Number；
Nonlinear optimization is carried out using SBA, obtains accurate camera inside and outside parameter.
Wherein, described the step of carrying out the matching of foot type dense feature point according to the camera inside and outside parameter, includes：
According to the camera inside and outside parameter, contain the unique code value of characteristic point using the window features generation of pseudorandom arrays
Colored argyle design, extraction diamond shape angle point is as characteristic point；
The detection of multicolour pattern diamond shape array characteristic point is carried out, the image coordinate of dense feature point is obtained, matches intensive spy
Sign point.
Wherein, the progress multicolour pattern diamond shape array characteristic point detection, obtains the image coordinate of dense feature point, matches
The step of intensive characteristic point includes：
It uses preset Filtering Template using each characteristic point as anchor point, makees convolution operation with image, according to response size
Characteristic point is classified, the type code value of characteristic point is obtained；
Location feature point color code regions；
Diamond shape array color is decoded based on HCL color spaces, obtains the color code value of characteristic point；
The complete code value of characteristic point is worth in conjunction with the type code value and color code；
Code value constraint, limit restraint and the disparity range constraint of binding characteristic point, match intensive characteristic point.
Wherein, progress multicolour pattern diamond shape array characteristic point detection, before the step of matching intensive characteristic point also
Including：
Three-dimensional correction is carried out to input picture according to the camera inside and outside parameter, gray processing, gaussian filtering, contrast are drawn
It stretches, histogram equalization and expansion process.
Wherein, described matched according to the camera inside and outside parameter and foot type dense feature point as a result, carry out model structure
The step of building include：
After carrying out three-dimensional point cloud splicing using camera inside and outside parameter and Feature Points Matching relationship tectonic transition matrix, obtain dilute
Dredge threedimensional model；
Using Poisson surface algorithm for reconstructing, the dense model of foot type is obtained.
Wherein, described matched according to the camera inside and outside parameter and foot type dense feature point as a result, carry out foot type ginseng
Counting the step of measuring includes：
Generate footprint figure；
Foot type characteristic point is positioned on the footprint figure, back projection determines its three-dimensional position to foot model；
According to feature girth constructing definitions auxiliary plane, feature girth is calculated.
The present invention also proposes a kind of multi-view stereo vision foot type three-dimension measuring system encoded based on color, including storage
Device, processor and the computer program being stored on the memory, when the computer program is run by the processor
The step of realizing method as described above.
The present invention also proposes a kind of computer readable storage medium, and calculating is stored on the computer readable storage medium
The step of machine program, the computer program realizes method as described above when being run by processor.
The beneficial effects of the invention are as follows：The present invention is based on color coding multi-view stereo vision foot type method for three-dimensional measurement and
System, this method are installed with the foot type of coloud coding pattern socks by multiple camera sync photographies, by pattern Corner Detection,
Corners Matching, three-dimensional point set generation, Surface Reconstruction, foot type parameter based on color coding pattern calculate step, rebuild simultaneously
Multiple dimensional parameters of foot type are calculated, the polyphaser punctuate algorithm based on sphere is realized, obtains accurate camera parameter；Pass through
The colored argyle design that characteristic point code value is contained in generation prints on socks, and feature point coordinates is obtained using detection algorithm, carries out
Characteristic point decodes, and intensive characteristic point is matched under multiple constraint, realizes that foot type is rebuild and foot type parameter measurement.With existing method
It compares, the present invention has at low cost, characteristic point readily identified and matching, that the point cloud density that generates is higher, calculating speed is fast etc. is excellent
Point obtains accurately and completely foot model and relative dimensions parameter.
In general, have the characteristics that as follows：
1, image synchronization and quickly is acquired --- it is laid out, is built with conducive to figure by designing camera bracket and multi-cam
As the uniform light environment of acquisition, and polyphaser synchronous acquisition foot type figure is realized by the script based on socket interfaces
Case, synchronization time difference are less than 10ms.
2, camera calibration parameter is accurate --- the polyphaser calibration algorithm based on sphere is realized, in conjunction with Filtering Template and ladder
The detection algorithm for spending intensity obtains mark point 2D coordinates, and mark point 3D coordinates are calculated in conjunction with imaging model and sphere geometrical constraint,
It realizes reference points matching, obtains and join outside initial camera, and carry out nonlinear optimization, obtain camera interior and exterior parameter, realize
The coordinate system of polyphaser is unified.
3, reconstruction model is dense, measurement parameter is accurate --- propose a kind of foot type method for reconstructing based on color coding, it is raw
At the colored diamond shape array pattern for containing uniform and dense feature point, the precision of Corner Detection and the stabilization of characteristic matching are improved
Property, more dense foot type three dimensional point cloud can be quickly generated, the measurement based on footprint figure and auxiliary plane is finally utilized
Method realizes that 8 dimensional parameters of foot type measure.
Description of the drawings
Fig. 1 is the flow diagram of the multi-view stereo vision foot type method for three-dimensional measurement encoded the present invention is based on color；
Fig. 2 is color coding foot type three-dimension measuring system principle schematic of the present invention；
Fig. 3 is the polyphaser scaling method flow diagram the present invention is based on sphere；
Fig. 4 is multicolour pattern diamond shape array characteristic point detects schematic diagram of the present invention；
Fig. 5 is that foot surface of the present invention dense feature point detects and matches schematic diagram；
Fig. 6 is the spherical caliberating device schematic diagram for posting mark point；
Fig. 7 is mark point Filtering Template schematic diagram；
Fig. 8 is characteristic point classification Filtering Template schematic diagram；
Fig. 9 is characteristic point classification Filtering Template schematic diagram；
Figure 10 is foot type point cloud vertical view and local coordinate system direction schematic diagram；
Figure 11 is the foot type point cloud plane projection schematic diagram for indicating cut-off rule；
Figure 12 is the footprint figure for indicating foot type characteristic point；
Figure 13 is foot type girth auxiliary plane schematic diagram；
Figure 14 is point cloud slicing and auxiliary plane；
Figure 15 is the slice of data schematic diagram under polar coordinate system；
Figure 16 is that slice of data divides schematic diagram；
Figure 17 is that the preceding four groups of foot type of experiment rebuild input and reconstructed results schematic diagram.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific implementation mode
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
Fig. 1 is please referred to, Fig. 1 is the flow of the multi-view stereo vision foot type method for three-dimensional measurement encoded the present invention is based on color
Schematic diagram.
As shown in Figure 1, a kind of multi-view stereo vision foot type method for three-dimensional measurement based on color coding proposed by the present invention,
Include the following steps：
Step S1 is installed with coloud coding pattern socks by the acquisition platform of more mesh images using polyphaser sync pulse jamming
Foot type, carry out Image Acquisition；
Step S2, the polyphaser calibration algorithm according to the image collected, and based on sphere obtain camera inside and outside parameter；
Step S3 carries out the matching of foot type dense feature point according to the camera inside and outside parameter；
Step S4, it is matched according to the camera inside and outside parameter and foot type dense feature point as a result, carrying out model construction
With foot type parameter measurement.
Wherein, wherein the acquisition platform of more mesh images has 360 ° of visual angles, has uniform illumination environment, utilizes two
A script based on socket realizes the polyphaser synchronous acquisition of image.
It is described according to the image collected, and the step of polyphaser calibration algorithm based on sphere obtains camera inside and outside parameter
Including：
According to the image collected, the loop truss algorithm based on Hough transformation estimates projection coordinate and figure of the centre of sphere in image
The size of diameter as in；
Spherical surface Corner Detection is carried out, 2D coordinates are obtained, the 3D coordinates of mark point and the centre of sphere is calculated, estimates the initial outer ginseng of camera
Number；
Nonlinear optimization is carried out using SBA, obtains accurate camera inside and outside parameter.
It is described according to the camera inside and outside parameter carry out foot type dense feature point matching the step of include：
According to the camera inside and outside parameter, contain the unique code value of characteristic point using the window features generation of pseudorandom arrays
Colored argyle design, extraction diamond shape angle point is as characteristic point；
The detection of multicolour pattern diamond shape array characteristic point is carried out, the image coordinate of dense feature point is obtained, matches intensive spy
Sign point.
The progress multicolour pattern diamond shape array characteristic point detection, obtains the image coordinate of dense feature point, and matching is intensive
Characteristic point the step of include：
It uses preset Filtering Template using each characteristic point as anchor point, makees convolution operation with image, according to response size
Characteristic point is classified, the type code value of characteristic point is obtained；
Location feature point color code regions；
Diamond shape array color is decoded based on HCL color spaces, obtains the color code value of characteristic point；
The complete code value of characteristic point is worth in conjunction with the type code value and color code；
Code value constraint, limit restraint and the disparity range constraint of binding characteristic point, match intensive characteristic point.
It is described matched according to the camera inside and outside parameter and foot type dense feature point as a result, carry out model construction step
Suddenly include：
After carrying out three-dimensional point cloud splicing using camera inside and outside parameter and Feature Points Matching relationship tectonic transition matrix, obtain dilute
Dredge threedimensional model；
Using Poisson surface algorithm for reconstructing, the dense model of foot type is obtained.
It is described matched according to the camera inside and outside parameter and foot type dense feature point as a result, carry out foot type parameter measurement
The step of include：
Generate footprint figure；
Foot type characteristic point is positioned on the footprint figure, back projection determines its three-dimensional position to foot model；
According to feature girth constructing definitions auxiliary plane, feature girth is calculated.
The present invention realizes the polyphaser punctuate algorithm based on sphere, obtains accurate camera parameter；Contained by generating
The colored argyle design of characteristic point code value prints on socks, obtains feature point coordinates using detection algorithm, carries out characteristic point solution
Code matches intensive characteristic point under multiple constraint, realize that foot type is rebuild and foot type parameter measurement.Compared with the existing methods, originally
Invention has many advantages, such as that at low cost, characteristic point is readily identified and matches, the point cloud density of generation is higher, calculating speed is fast, accurately
And completely obtain foot model and relative dimensions parameter.
The present invention program is described in detail below：
The present invention relates to a kind of multi-view stereo vision foot type method for three-dimensional measurement and system based on color coding, this method
It is installed with the foot type of coloud coding pattern socks by multiple camera sync photographies, is compiled by pattern Corner Detection, based on color
Corners Matching, three-dimensional point set generation, Surface Reconstruction, the foot type parameter of code pattern calculate step, rebuild and calculate foot type
8 dimensional parameters.Present invention can apply to the occasions such as foot type reconstruction and dimensional measurement, such as foot shape measurement, footwear to customize
With the fields such as foot medical treatment.
The flow chart of foot shape measurement system of the present invention is as shown in Fig. 2, Fig. 3 is the polyphaser calibration based on sphere
Flow chart, Fig. 4 are the flow chart of colored diamond shape array angle point grid, and Fig. 5 is the intensive diamond shape array angle point encoded based on color
Match flow chart.
1, the acquisition platform of more mesh images
The present invention is based on color coding multi-view stereo vision three-dimensional foot type measuring system using computer, light source, camera,
2. the hardware such as holder, Switching Power Supply, router, interchanger construct with 360 ° of visual angles and have uniform illumination environment 3.
1., using two scripts based on socket 4. polyphaser acquisition platform realizes the synchronous acquisition of image.
2, the polyphaser calibration based on sphere
System uses a kind of calibration algorithm based on sphere 5. to obtain accurate camera inside and outside parameter.A diameter of D's
Multiple tag blocks being made of two roundlets are placed on the spherical surface of spherical displacer, as shown in Figure 6.Two small centers of circle are considered as mark point,
The spacing in two centers of circle is d1.Each mark point surrounding border circular areas is chequered with black and white.This method only needs each video camera to carry
For few image, usual one or two.Specifically the flow chart of the polyphaser calibration based on sphere is as shown in Figure 3.
Wherein 8. step is to estimate the size of centre of sphere diameter in the projection coordinate of image and image, is become by being based on Hough
The loop truss algorithm changed is realized.9. step is the subpixel coordinates of mark point on detection spherical surface, realized by following steps：
(1) circle that Hough transformation detects is set as to the region of interest ROI of image；
(2) make convolution with image using the mark point Filtering Template such as Fig. 7, realize the detection of candidate point；
(3) it is all Filtering Template response combinations to define a certain pixel and whether be characterized the likelihood value C of a candidate point
Maximum value, can be indicated by (2-1) to (2-4).
In formula：C --- pixel is the likelihood value of mark point；Si 1,2--- pixel is the mark point of the i-th class template detection
Likelihood value；fi x--- filter response of the pixel for the form x of the i-th class template；U --- pixel is for the i-th class template
Average filter responds；A, b, c, d --- four kinds of filters in Fig. 7.
(4) product that candidate point score Score is candidate point gradient intensity G and above-mentioned candidate point likelihood value C, profit are defined
It is indicated with formula (2-5).
Score=GC (2-5)
In formula：G --- candidate point gradient intensity value；Score --- candidate point score.
When Score is less than the threshold value threshold of setting, it is mark point to exclude the candidate point.Wherein gradient intensity G can
It is indicated with formula (2-6).
In formula：Gx--- the gradient intensity value of candidate point x-axis direction；Gy--- the gradient intensity value in candidate point y-axis direction.
X, Y direction gradient Gx, GyIt Sobel operators template can be used to do convolutional calculation with input picture to obtain, public affairs can be used
Formula (2-7) and (2-8) are indicated.
In formula：Img --- input picture；* --- convolution operation.
(5) angular coordinate of sub-pixel precision is extracted
10. step calculates the 3D coordinates of spherical surface mark point and the centre of sphere, wherein sphere centre coordinate (Xc0,Yc0,Zc0) can be according to initial
The geometrical constraint of internal reference A combination spheres is calculated, and can be indicated by formula (2-9) to (2-12).
In formula：The true diameter of D --- spherical displacer；Diameter of d --- the spherical displacer in image；A --- video camera internal reference
Matrix；u0--- principal point abscissa；v0--- principal point ordinate；fx--- the normalization focal length in x-axis；fy--- returning in y-axis
One changes focal length.
Spherical surface mark point 3D coordinates (Xcc,Ycc,Zcc) in combination with spherical surface mark point 2D coordinates, camera imaging model, internal reference
Initial value A and sphere geometrical constraint are calculated, and can be indicated by (2-13) to (2-15).
(Xcc-Xc0)2+(Ycc-Yc0)2+(Zcc-Zc0)2=D2 (2-15)
StepFeature Points Matching is realized using the thought of RANSAC, is obtained and is joined T and R outside initial camera.StepMake
Nonlinear optimization is carried out with SBA, to obtain the camera interior and exterior parameter of optimization.
3, instep dense feature point is generated designs with code value
For people's foot surface relative smooth, the less situation of high-frequency characteristic, to obtain more intensive three-dimension foot model, this
Invention generates a kind of colored argyle design containing the unique code value of characteristic point using the good window features of pseudorandom arrays, carries
It takes diamond shape angle point as characteristic point, matches intensive characteristic point, realize that foot type is rebuild and measured.Q member m rank pseudorandom arrays are good
Window features refer to k1×k2The window of size slides in the pseudorandom arrays of n1 × n2, the sequence arbitrarily arrived by window frame
It is unique for being listed in pseudorandom arrays all.Wherein, the period n of q members m rank pseudo-random sequences meets formula (3-1).k1*k2
The size of window, n are indicated respectively1,n2The size of pseudorandom arrays is indicated respectively, they meet formula (3-2) and arrive (3-4) respectively：
N=qm-1(3-1)
M=k1×k2(3-2)
Before the argyle design of code value is contained in generation, need the surface area for first estimating foot and the quasi- diamond block used big
It is small, to confirm the array for generating a few several ranks of member, that is, determine the value of q and m.By calculating, the present invention selects to generate 8 yuan of 6 rank puppet
Random array.I.e. given non-full zero original state a0a1a2a3a4a5=000001, table look-up the transmission function for obtaining pseudorandom arrays
For ai+6=ai+1+Aai, according to 8 operation rule of mould, the generation period is n=86The pseudo-random sequence of -1=262656
a0a1a2a3a4a5...=000001000120....Rule is converted into according to diagonal line, pseudo-random sequence is folded into 511 ×
513(n1=83- 1=262656, n2=n/n1=513) pseudorandom arrays of size.K at this time1=3, k2=2, i.e. three rows two row
Wicket frame in array be unique in 511 × 513 big array.If the array in the wicket frame that three rows two are arranged
It is given to diamond shape angle point respectively as code value, then the matching of dense feature point can be achieved.
As shown in table 1, the present invention is based on 9 kinds of differentiable colors of HCL color-space choosings respectively represent 8 kinds of primitive colors and
Background colour.As shown in figure 8, representing the numerical value of pseudorandom arrays with colored diamond block, the coding pattern based on color is generated.With water chestnut
Shape angle point defines one 7 code values for each characteristic point as characteristic point, and first place indicates the type of characteristic point, and latter six
Indicate color code value.Diamond shape array characteristic point is divided into I, II liang of class, I class is red point (I, J, K etc.), is indicated with number 8, it
Be two colored diamond shapes of left and right intersection point.II class is blue dot (A, B, C etc.), is indicated with number 9, they are upper and lower two colored water chestnuts
The intersection point of shape.The color code regions for defining each I category feature point are the rectangles arranged as three rows two of geometric center using it.With spy
For levying point J, the red block in Fig. 8 is its color code regions.And the color code regions of 1 category feature point be defined as with
The color code regions of its adjacent 1 category feature point of lower right are identical, i.e. the color code regions phase of characteristic point C and characteristic point J
Together, their code values are the difference is that the first feature vertex type.The color coding in conjunction with shown in table 1 in 8, it is known that characteristic point J
Code value be 8106470, the code value of characteristic point C is 9106470.
1 color coding schedule of table
4, instep dense feature point matches
6. step is foot type dense feature point matching operation, the flow chart of specific implementation is as shown in Figure 4.
(1) three-dimensional correction is carried out to input picture using camera parameterSo that corresponding points row coordinate is equal, reduces and correspond to
Point search range；
(2) stepRealize that the detection of foot type dense feature point, the flow chart of specific implementation are as shown in Figure 5.By three-dimensional school
Input picture after just uses after the operations such as gray processing, gaussian filtering, contrast stretching, histogram equalization and expansion
Step 9. in reference points detection algorithm, obtain dense feature point image coordinate；
(3) stepUsing Filtering Template as shown in Figure 9 using each characteristic point as anchor point, making convolution behaviour with image
Make, size classifies characteristic point according to response, obtains the type code value of characteristic point；
(4) stepLocation feature point color code regions need to navigate to red boxes in Fig. 8 by taking characteristic point J as an example
6 diamond shape color blocks.First by the ratio of distance and the shortest distance between solution characteristic point, four differences are found for point J
Adjacent feature point C, D, E and F of type can be indicated by formula (4-1) and (4-2).As ratio ＞=0.75, it is believed that be
Different types of adjacent feature point.
In formula：xk,yk--- the transverse and longitudinal coordinate value of II category feature point k；xj,yj--- the transverse and longitudinal coordinate value of I category feature point j；
dmin--- the shortest distances of all II category feature point k to I category feature point j；Ratio --- each II category feature point k is special to I class
Levy the distance and shortest distance d of point jminRatio.
After finding characteristic point C, D, E point, characteristic point A (x can be estimated by formula (4-3) to (4-5)A,yA), I (xI,
yI), L (xL,yL) coordinate.
(x in formulaC,yC), (xD,yD), (xE,yE), (xJ,yJ) be respectively C, D, E and J point in Fig. 8 coordinate.So that it is determined that
It is remaining similarly can to estimate that the coordinate of B, O, M, P, N, K, Q, G and H determine for the diamond block region of characteristic point J the first row first rows
Five pieces of diamond-shaped areas.
(5) stepDiamond shape array color is decoded based on HCL color spaces, obtains the color code value of characteristic point.In conjunction with class
Type code value and color code are worth to the complete code value of characteristic point.
(6) the code value constraint of binding characteristic point, limit restraint and disparity range constraint, match intensive characteristic point.
5, model construction and foot type parameter measurement
7. step realizes structure and the foot type parameter measurement of three-dimension foot model.After completing the matching between multiple view picture, at this time
The parallax of each point can be calculated by (5-1) on image.The D coordinates value of characteristic point is represented by (5-2).
D=xl-xr (5-1)
In formula：xl--- point is row coordinate value at left image midpoint；xr--- the row of this Corresponding matching point in right figure are sat
Scale value；Parallax value of d --- this o'clock between two views.
In formula：W --- arbitrary constant；The three-dimensional world coordinate value of X, Y, Z --- point.
The three-dimensional coordinate of this time point is (X, Y, Z), using camera parameter and Feature Points Matching relationship tectonic transition matrix into
After the splicing of row three-dimensional point cloud, sparse three-dimensional model is obtained.After using Poisson surface algorithm for reconstructing, the dense mould of foot type is obtained
Type.The present invention is using the foot shape measurement method based on footprint figure and auxiliary plane to the foot type dimensional parameters of influence shoes comfort level
It measures, the definition of foot type dimensional parameters is as shown in table 2.
2 foot type dimensional parameters of table define
| Foot type characteristic parameter | Parameter definition |
| The long L of foot | Distance farthest on foot length direction in foot type point cloud |
| The wide W of foot | Distance farthest in foot wide direction in foot type point cloud |
| Plantar toe girth C1 | The girth obtained around articulationes metatarsophalangeae point measurement |
| Shank girth C2 | Cross the girth of foot den |
| Pocket encloses C3 | On boat curved point and followed by point girth |
| Metatarsophalangeal joint height H1 | Apogee distance measures the height of foot plate surface in plantar toe girth |
| Preceding shank height H2 | The height of apogee distance foot plate surface in shank girth |
| Curved point height H3 on boat | Pocket places the height of apogee distance foot plate surface |
Foot type parameter measurement includes three steps：Footprint figure is firstly generated, second step positions foot type feature on footprint figure
Point, back projection to foot model determine that its three-dimensional position, third step calculate feature and enclose according to feature girth constructing definitions auxiliary plane
It is long.
5.1 footprint figures generate
The local coordinate system relative to foot type is defined, as shown in Figure 10, the direction of foot type length is defined as X-direction, foot
The direction of molded breadth degree is defined as Y direction, and the direction of foot type height is defined as Z-direction, and foot type coordinate system meets right-handed system rule
Then.Foot type bottom surface is XOY plane, is YOZ planes along foot wide direction along the position XOZ planes of foot length direction in foot type side.
On spot projection to XOY plane by foot type point cloud level degree less than 40mm.Using toe endpoint as starting point, edge is parallel to X
Axis and Y direction are found successively, will entirely be thrown with this four for scanning boundary point near four points on boundary on projection plane
Shadow borderline region is divided into five pieces, as shown in figure 11.
The vertical sweep line that foot wide direction is defined with a fixed step size is moved from point A to point B, between line segment AD and BC
It is moved in section, the point surveyed in foot type is found between adjacent two scan lines, AB sections of corresponding numbers are arrived in storage successively
In group.After obtaining AB sections of corresponding arrays, BC sections, CD sections, AD sections of corresponding arrays can be obtained in the same way.Draw number
Point in group, is linked to be line segment successively, has obtained the footprint figure of foot type, as shown in figure 12.
5.2 foot type positioning feature points
It is located at the local bump location of footprint figure, the actual bit of these characteristic points with the relevant characteristic point of characteristic parameter extraction
Set be respectively positioned on foot lean to one side on surface rather than foot plate surface.Each characteristic point is positioned on two-dimentional footprint Figure 12, and it is instead thrown
In shadow to three-dimensional point cloud, the three-dimensional position with the relevant each characteristic point of foot type characteristic parameter is found.Wherein feature 1 is toe end
Point, characteristic point 4 are heel point, are located on the position for two points that distance is farthest on foot length direction, this 2 points rectangular along foot
To distance be exactly foot type length.Characteristic point 3 and characteristic point 6 are the position for being located at two points that distance is farthest in foot wide direction respectively
It sets, their length in foot wide direction are exactly foot type width.Characteristic point 7 is foot den.Its condition met is foot den
To heel bump along 0.41 times that the distance of foot length direction is foot type length.
5.3 foot type girths calculate
In foot type characteristic parameter, Metatarsophalangeal joint height, preceding shank height, on boat curved point height be respectively plantar enclose, instep
It encloses, pocket encloses the highest point of middle height.Therefore, three foot type girths are the cores of foot type parameter measurement.According to foot type characteristic parameter
Definition obtain geometric position of each girth relative to foot type, for each girth, we construct an auxiliary plane respectively,
Each girth is sought by seeking the intersection on foot type surface and auxiliary plane, the auxiliary signal of construction is as shown in figure 13.
The length of 2 points of a length of foot thumb endpoint of foot and heel bump along foot length direction.The straight width of plantar toe is 3,6 points along foot
The length of wide direction.The corresponding auxiliary plane of plantar toe girth was Metatarsophalangeal joint point and the 5th articulationes metatarsophalangeae point, and and XOY
The vertical plane of plane, the intersection length with three-dimensional foot type are plantar toe girth length.Construct the auxiliary plane of shank girth
When, the parallel plane in the foot den faces Zuo YuYOZ is crossed first.When measuring left foot, using Z axis as rotary shaft, by plane to Y-axis positive direction
15 degree of rotation.When measuring right crus of diaphragm, plane is rotated 15 degree to Y-axis negative direction.Then with the phase of auxiliary plane at this time and XOY plane
Friendship straight line is axis, and plane is rotated 21 degree to Z axis negative direction, the corresponding auxiliary plane of shank girth is obtained, with three-dimensional foot type
Intersection length be shank girth length.Pocket, which encloses corresponding auxiliary plane, is put down by curved point on boat and heel point, and with XOZ
The vertical plane in face, the intersection length with three-dimensional foot type are pocket girth degree.
After obtaining auxiliary plane, certain thickness slice is done to foot type point cloud, i.e., it will be certain in auxiliary plane
The point of distance extracts.This distance selection is related and small as possible with the density of point cloud, just not will produce larger mistake in this way
Difference.Our selected distance auxiliary plane 1mm's, that is, do the slice of 2mm thickness.As shown in figure 14.
By on the spot projection to auxiliary plane in slice, if T is slice of data point set, the number of data point in the set
For n.The centre of form O of subpoint is determined according to formula (5-3)o。
Using the centre of form as polar origin, make a ray for being parallel to X-axis, as polar axis, establishes slice of data pole seat
Mark system, the coordinate (r, θ) of slice of data, as shown in figure 15.
Girth profile wire shaped is relatively simple, can to its by polar angle division handle, as shown in figure 16, select 1 ° as walk
It is long, the slice point data after projection is divided into multiple regions, the centre of form in each region is sought, by these centres of form according to polar angle sequence
It is sequentially connected to obtain multistage short-term section, sums to line segment, just obtain the length of girth.
Example is embodied
The present invention has built the polyphaser image capturing system at 360 ° of visual angles, is accurately taken the photograph by realizing that sphere calibration obtains
Camera inside and outside parameter prints special colored diamond shape array pattern on socks, and the foot type image of socks is worn in shooting, is realized close
The feature point extraction of collection and matching, accumulated error is larger when solving gridiron pattern standardization applied to polyphaser and instep stablizes spy
The problem of sign point is difficult to set up intensive model less.The effect of the present invention is illustrated in conjunction with following specific reconstructed results.Figure 17 is shown
Foot type three-dimensional reconstruction of the present invention as a result, first is classified as an input picture, secondary series is the point cloud rebuild, and third is classified as weight
The Visualization Model built.
38 groups of experiment foot type parameter measurements of table
The evaluation common index of algorithm for reconstructing, which has, rebuilds integrality and reconstruction accuracy.Foot type characteristic parameter measurement result is such as
Shown in table 3, statistical analysis can obtain, and within 1mm, the worst error of girth and height exists the error of foot type length and width
Within 2mm, the requirement that shoes are customized according to foot type disclosure satisfy that.
Compared with prior art, the present invention has the characteristics that as follows：
1, image synchronization and quickly is acquired --- it is laid out, is built with conducive to figure by designing camera bracket and multi-cam
As the uniform light environment of acquisition, and polyphaser synchronous acquisition foot type figure is realized by the script based on socket interfaces
Case, synchronization time difference are less than 10ms.
2, camera calibration parameter is accurate --- the polyphaser calibration algorithm based on sphere is realized, in conjunction with Filtering Template and ladder
The detection algorithm for spending intensity obtains mark point 2D coordinates, and mark point 3D coordinates are calculated in conjunction with imaging model and sphere geometrical constraint,
It realizes reference points matching, obtains and join outside initial camera, and carry out nonlinear optimization, obtain camera interior and exterior parameter, realize
The coordinate system of polyphaser is unified.
3, reconstruction model is dense, measurement parameter is accurate --- propose a kind of foot type method for reconstructing based on color coding, it is raw
At the colored diamond shape array pattern for containing uniform and dense feature point, the precision of Corner Detection and the stabilization of characteristic matching are improved
Property, more dense foot type three dimensional point cloud can be quickly generated, the measurement based on footprint figure and auxiliary plane is finally utilized
Method realizes that 8 dimensional parameters of foot type measure.
In addition, the present invention also proposes a kind of multi-view stereo vision foot type three-dimension measuring system encoded based on color, including
Memory, processor and the computer program being stored on the memory, the computer program are transported by the processor
The step of method as described above is realized when row, principle please refers to above method embodiment, and details are not described herein.
In addition, the present invention also proposes a kind of computer readable storage medium, stored on the computer readable storage medium
There are the step of computer program, the computer program realizes method as described above when being run by processor, principle that please join
According to above method embodiment, details are not described herein.
The foregoing is merely the preferred embodiment of the present invention, are not intended to limit the scope of the invention, every utilization
Equivalent structure made by description of the invention and accompanying drawing content or flow transformation, are applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of multi-view stereo vision foot type method for three-dimensional measurement based on color coding, which is characterized in that include the following steps：
By the acquisition platform of more mesh images, it is installed with the foot type of coloud coding pattern socks using polyphaser sync pulse jamming, carries out
Image Acquisition；
Polyphaser calibration algorithm according to the image collected, and based on sphere obtains camera inside and outside parameter；
The matching of foot type dense feature point is carried out according to the camera inside and outside parameter；
It is matched according to the camera inside and outside parameter and foot type dense feature point to be surveyed with foot type parameter as a result, carrying out model construction
Amount.
2. the multi-view stereo vision foot type method for three-dimensional measurement according to claim 1 based on color coding, feature exist
In the acquisition platform of more mesh images has 360 ° of visual angles, has uniform illumination environment, utilizes two feet based on socket
The polyphaser synchronous acquisition of this realization image.
3. the multi-view stereo vision foot type method for three-dimensional measurement according to claim 1 based on color coding, feature exist
In, it is described according to the image collected, and the step of acquisition camera inside and outside parameter of the polyphaser calibration algorithm based on sphere includes：
According to the image collected, the loop truss algorithm based on Hough transformation estimates the centre of sphere in the projection coordinate of image and image
The size of diameter；
Spherical surface Corner Detection is carried out, 2D coordinates are obtained, the 3D coordinates of mark point and the centre of sphere is calculated, estimates the initial outer parameter of camera；
Nonlinear optimization is carried out using SBA, obtains accurate camera inside and outside parameter.
4. the multi-view stereo vision foot type method for three-dimensional measurement according to claim 3 based on color coding, feature exist
In, it is described according to the camera inside and outside parameter carry out foot type dense feature point matching the step of include：
According to the camera inside and outside parameter, the colour for containing the unique code value of characteristic point is generated using the window features of pseudorandom arrays
Argyle design, extraction diamond shape angle point is as characteristic point；
The detection of multicolour pattern diamond shape array characteristic point is carried out, the image coordinate of dense feature point is obtained, matches intensive characteristic point.
5. the multi-view stereo vision foot type method for three-dimensional measurement according to claim 4 based on color coding, feature exist
In the progress multicolour pattern diamond shape array characteristic point detection obtains the image coordinate of dense feature point, matches intensive feature
Point the step of include：
It uses preset Filtering Template using each characteristic point as anchor point, makees convolution operation with image, size will be special according to response
Sign point is classified, and the type code value of characteristic point is obtained；
Location feature point color code regions；
Diamond shape array color is decoded based on HCL color spaces, obtains the color code value of characteristic point；
The complete code value of characteristic point is worth in conjunction with the type code value and color code；
Code value constraint, limit restraint and the disparity range constraint of binding characteristic point, match intensive characteristic point.
6. the multi-view stereo vision foot type method for three-dimensional measurement according to claim 4 based on color coding, feature exist
In, progress multicolour pattern diamond shape array characteristic point detection, further include before the step of matching intensive characteristic point：
Three-dimensional correction carried out to input picture according to the camera inside and outside parameter, it is gray processing, gaussian filtering, contrast stretching, straight
Side's figure equalization and expansion process.
7. the multi-view stereo vision foot type method for three-dimensional measurement according to claim 5 based on color coding, feature exist
In described matched according to the camera inside and outside parameter and foot type dense feature point as a result, the step of carrying out model construction packet
It includes：
After carrying out three-dimensional point cloud splicing using camera inside and outside parameter and Feature Points Matching relationship tectonic transition matrix, sparse three are obtained
Dimension module；
Using Poisson surface algorithm for reconstructing, the dense model of foot type is obtained.
8. the multi-view stereo vision foot type method for three-dimensional measurement according to claim 7 based on color coding, feature exist
In described matched according to the camera inside and outside parameter and foot type dense feature point as a result, carrying out the step of foot type parameter measurement
Suddenly include：
Generate footprint figure；
Foot type characteristic point is positioned on the footprint figure, back projection determines its three-dimensional position to foot model；
According to feature girth constructing definitions auxiliary plane, feature girth is calculated.
9. a kind of multi-view stereo vision foot type three-dimension measuring system based on color coding, which is characterized in that including memory, place
Reason device and the computer program being stored on the memory, are realized such as when the computer program is run by the processor
The step of claim 1-8 any one of them methods.
10. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium
Program, the step of method as described in any one of claim 1-8 is realized when the computer program is run by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810074452.2A CN108305286B (en) | 2018-01-25 | 2018-01-25 | Color coding-based multi-view stereoscopic vision foot type three-dimensional measurement method, system and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810074452.2A CN108305286B (en) | 2018-01-25 | 2018-01-25 | Color coding-based multi-view stereoscopic vision foot type three-dimensional measurement method, system and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108305286A true CN108305286A (en) | 2018-07-20 |
| CN108305286B CN108305286B (en) | 2021-09-07 |
Family
ID=62866558
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810074452.2A Active CN108305286B (en) | 2018-01-25 | 2018-01-25 | Color coding-based multi-view stereoscopic vision foot type three-dimensional measurement method, system and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108305286B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109584356A (en) * | 2018-11-23 | 2019-04-05 | 东南大学 | A kind of decoded more view reconstructing methods of M-array image adaptive local window |
| CN109859228A (en) * | 2019-01-18 | 2019-06-07 | 深圳壹账通智能科技有限公司 | Recommend the method, apparatus and computer equipment of shoes money type based on image recognition |
| CN110288713A (en) * | 2019-07-03 | 2019-09-27 | 北京机械设备研究所 | A kind of quick three-dimensional model reconstruction method and system based on multi-vision visual |
| CN110751145A (en) * | 2019-10-18 | 2020-02-04 | 南京大学 | Multi-view image acquisition and preprocessing system and method under complex illumination condition |
| CN110751731A (en) * | 2019-10-16 | 2020-02-04 | 光沦科技(深圳)有限公司 | 3D reconstruction method and system for structured light |
| WO2020114035A1 (en) * | 2018-12-04 | 2020-06-11 | 中国科学院自动化研究所 | Three-dimensional feature extraction method and apparatus based on machine vision |
| CN111402341A (en) * | 2020-03-10 | 2020-07-10 | 创新奇智(广州)科技有限公司 | Camera parameter determination method and device, electronic equipment and readable storage medium |
| CN111739106A (en) * | 2020-08-12 | 2020-10-02 | 成都极米科技股份有限公司 | Corner point coding method, calibration method, device, electronic equipment and storage medium |
| CN112617809A (en) * | 2020-12-24 | 2021-04-09 | 新拓三维技术(深圳)有限公司 | Footprint area calculation method and system |
| CN113170094A (en) * | 2018-12-05 | 2021-07-23 | 腾讯美国有限责任公司 | Method and apparatus for geometric smoothing |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1544883A (en) * | 2003-11-25 | 2004-11-10 | 浙江大学 | Three-dimensional foot type measuring and modeling method based on specific grid pattern |
| CN101658347A (en) * | 2009-09-24 | 2010-03-03 | 浙江大学 | Method for obtaining dynamic shape of foot model |
| CN102445165A (en) * | 2011-08-05 | 2012-05-09 | 南京航空航天大学 | Stereo vision measurement method based on single-frame color coding grating |
| CN103591906A (en) * | 2012-08-13 | 2014-02-19 | 上海威塔数字科技有限公司 | A method for carrying out three dimensional tracking measurement on a moving object through utilizing two dimensional coding |
| CN104126989A (en) * | 2014-07-30 | 2014-11-05 | 福州大学 | Foot surface three-dimensional information obtaining method based on multiple RGB-D cameras |
| WO2016051416A1 (en) * | 2014-09-30 | 2016-04-07 | Fitfully Ltd. | System and method of 3d modeling and virtual fitting of 3d objects |
-
2018
- 2018-01-25 CN CN201810074452.2A patent/CN108305286B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1544883A (en) * | 2003-11-25 | 2004-11-10 | 浙江大学 | Three-dimensional foot type measuring and modeling method based on specific grid pattern |
| CN101658347A (en) * | 2009-09-24 | 2010-03-03 | 浙江大学 | Method for obtaining dynamic shape of foot model |
| CN102445165A (en) * | 2011-08-05 | 2012-05-09 | 南京航空航天大学 | Stereo vision measurement method based on single-frame color coding grating |
| CN103591906A (en) * | 2012-08-13 | 2014-02-19 | 上海威塔数字科技有限公司 | A method for carrying out three dimensional tracking measurement on a moving object through utilizing two dimensional coding |
| CN104126989A (en) * | 2014-07-30 | 2014-11-05 | 福州大学 | Foot surface three-dimensional information obtaining method based on multiple RGB-D cameras |
| WO2016051416A1 (en) * | 2014-09-30 | 2016-04-07 | Fitfully Ltd. | System and method of 3d modeling and virtual fitting of 3d objects |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109584356A (en) * | 2018-11-23 | 2019-04-05 | 东南大学 | A kind of decoded more view reconstructing methods of M-array image adaptive local window |
| WO2020114035A1 (en) * | 2018-12-04 | 2020-06-11 | 中国科学院自动化研究所 | Three-dimensional feature extraction method and apparatus based on machine vision |
| CN113170094A (en) * | 2018-12-05 | 2021-07-23 | 腾讯美国有限责任公司 | Method and apparatus for geometric smoothing |
| CN109859228A (en) * | 2019-01-18 | 2019-06-07 | 深圳壹账通智能科技有限公司 | Recommend the method, apparatus and computer equipment of shoes money type based on image recognition |
| CN110288713A (en) * | 2019-07-03 | 2019-09-27 | 北京机械设备研究所 | A kind of quick three-dimensional model reconstruction method and system based on multi-vision visual |
| CN110288713B (en) * | 2019-07-03 | 2022-12-23 | 北京机械设备研究所 | Rapid three-dimensional model reconstruction method and system based on multi-view vision |
| CN110751731A (en) * | 2019-10-16 | 2020-02-04 | 光沦科技(深圳)有限公司 | 3D reconstruction method and system for structured light |
| CN110751731B (en) * | 2019-10-16 | 2023-03-24 | 光沦科技(深圳)有限公司 | 3D reconstruction method and system for structured light |
| CN110751145A (en) * | 2019-10-18 | 2020-02-04 | 南京大学 | Multi-view image acquisition and preprocessing system and method under complex illumination condition |
| CN111402341A (en) * | 2020-03-10 | 2020-07-10 | 创新奇智(广州)科技有限公司 | Camera parameter determination method and device, electronic equipment and readable storage medium |
| CN111739106B (en) * | 2020-08-12 | 2020-12-25 | 成都极米科技股份有限公司 | Corner point coding method, calibration method, device, electronic equipment and storage medium |
| CN111739106A (en) * | 2020-08-12 | 2020-10-02 | 成都极米科技股份有限公司 | Corner point coding method, calibration method, device, electronic equipment and storage medium |
| CN112617809A (en) * | 2020-12-24 | 2021-04-09 | 新拓三维技术(深圳)有限公司 | Footprint area calculation method and system |
| CN112617809B (en) * | 2020-12-24 | 2024-05-24 | 新拓三维技术(深圳)有限公司 | Foot print area calculation method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108305286B (en) | 2021-09-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108305286A (en) | Multi-view stereo vision foot type method for three-dimensional measurement, system and medium based on color coding | |
| CN107945268B (en) | A kind of high-precision three-dimensional method for reconstructing and system based on binary area-structure light | |
| CN111462206B (en) | Monocular structure light depth imaging method based on convolutional neural network | |
| Furukawa et al. | Accurate camera calibration from multi-view stereo and bundle adjustment | |
| CN106091984B (en) | A kind of three dimensional point cloud acquisition methods based on line laser | |
| Waechter et al. | Virtual rephotography: Novel view prediction error for 3D reconstruction | |
| CN105005755B (en) | Three-dimensional face identification method and system | |
| CN106600686A (en) | Three-dimensional point cloud reconstruction method based on multiple uncalibrated images | |
| CN105205866B (en) | City threedimensional model fast construction method based on point off density cloud | |
| CN106910242A (en) | The method and system of indoor full scene three-dimensional reconstruction are carried out based on depth camera | |
| CN104820991B (en) | A kind of multiple soft-constraint solid matching method based on cost matrix | |
| CN105654547B (en) | Three-dimensional rebuilding method | |
| US20050088515A1 (en) | Camera ring for three-dimensional (3D) surface imaging | |
| CN106327532A (en) | Three-dimensional registering method for single image | |
| CN107274483A (en) | A kind of object dimensional model building method | |
| CN111563952B (en) | Method and system for realizing stereo matching based on phase information and spatial texture characteristics | |
| CN108648264A (en) | Underwater scene method for reconstructing based on exercise recovery and storage medium | |
| CN107977996A (en) | Space target positioning method based on target calibrating and positioning model | |
| Vu et al. | Efficient hybrid tree-based stereo matching with applications to postcapture image refocusing | |
| CN112884880B (en) | Line laser-based honey pomelo three-dimensional modeling device and method | |
| CN110096993A (en) | The object detection apparatus and method of binocular stereo vision | |
| Häne et al. | Hierarchical surface prediction | |
| CN107610219A (en) | The thick densification method of Pixel-level point cloud that geometry clue perceives in a kind of three-dimensional scenic reconstruct | |
| CN104796624A (en) | Method for editing and propagating light fields | |
| CN106323190A (en) | Depth measurement range-customizable depth measurement method and system for obtaining depth image |
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 |