CN108596980A - Circular target vision positioning precision assessment method, device, storage medium and processing equipment - Google Patents
Circular target vision positioning precision assessment method, device, storage medium and processing equipment Download PDFInfo
- Publication number
- CN108596980A CN108596980A CN201810273909.2A CN201810273909A CN108596980A CN 108596980 A CN108596980 A CN 108596980A CN 201810273909 A CN201810273909 A CN 201810273909A CN 108596980 A CN108596980 A CN 108596980A
- Authority
- CN
- China
- Prior art keywords
- circular target
- error
- discrete
- curve
- circular
- 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
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of circular target vision positioning precision assessment method, device, storage medium and processing equipments.Method includes:Set original state of the circular target relative to camera;By the continuous circular curve projection of circular target to the plane of delineation, elliptic curve is obtained;Grid discretization is carried out to elliptic curve according to image resolution ratio, simulation superposition is carried out to image error using monte carlo method, generates the discrete pixels point set curve with random error;Segmental arc extraction, ellipse fitting and the identification of space circle pose are carried out to discrete pixels point set curve, obtain the deviation spatial pose state of circular target;The margin of error that the deviation spatial pose state and original state of circular target are obtained by comparison, positioning accuracy is determined according to the margin of error.The present invention solves the coupling of multiple error source in circular target vision positioning and is difficult to detach the difficulty of analysis and qualitative assessment, can be to provide important support to the selection of circular target feature and utilization in monocular vision design of measuring system.
Description
Technical field
The present invention relates to vision positioning technical field more particularly to a kind of circular target vision positioning precision assessment method,
Device, storage medium and processing equipment.
Background technology
Monocular vision pose measurement is as a kind of important pose measurement means, with simple in structure, accuracy of measurement is high
The advantages that, the known geometrical model for measuring target, such as the point of target, straight line, curve feature are needed when measurement.Wherein circle is a kind of
Important curvilinear characteristic has important application to pass through especially in target positioning in fields such as target identification, tracking, positioning
Determining for target can be realized by single circular feature or additional a small amount of characteristic information in the size of calibration circular target in advance
Position, and due to the round symmetry characteristic of itself, part random error can be balanced so that positioning accuracy can generally be better than being based on
The localization method of point feature or line feature.The fields such as industry, space flight, medical treatment it is a large amount of in, if it is possible to dexterously using circle
Shape feature carries out visual perception and the system of measurement designs, and will play most important work to the simplification of whole system and stable operation
With.
Due in the application in the fields such as industry, space flight and medical treatment, it will usually because of mission requirements difference, to visual perception and survey
The precision of amount system proposes different requirements, and this requires can be to the perception and survey of vision system in the system design scheme stage
Accuracy of measurement is assessed.And the vision measurement system based on circular feature, due to the complexity of projection, precision is difficult to lead to
It crosses analytic method to be assessed, and measures the error close-coupled of the error and image procossing that resolve, it is difficult to which separation considers that this makes
The qualitative assessment for obtaining measurement accuracy is more difficult.
Invention content
The technical problem to be solved by the present invention is in view of the deficiencies of the prior art, provide a kind of circular target vision positioning
Precision assessment method, device, storage medium and processing equipment.
The technical solution that the present invention solves above-mentioned technical problem is as follows:A kind of circular target vision positioning accuracy evaluation side
Method includes the following steps:
Set original state of the circular target relative to camera;
By the imaging process of perspective projection principle simulation circular target, the continuous circular curve of the circular target is thrown
Shadow obtains elliptic curve to the plane of delineation;
Grid discretization is carried out to the elliptic curve according to image resolution ratio, discrete projection point is obtained, described discrete
Simulation superposition is carried out to image error using monte carlo method on subpoint, generates the discrete pixels point set with random error
Curve;
Segmental arc extraction, ellipse fitting and the identification of space circle pose are carried out to the discrete pixels point set curve, obtained round
The deviation spatial pose state of target;
The margin of error that the deviation spatial pose state and original state of circular target are obtained by comparison, according to the error
Amount determines positioning accuracy.
Another technical solution that the present invention solves above-mentioned technical problem is as follows:A kind of circular target vision positioning accuracy evaluation
Device, including:
Setup unit, for setting original state of the circular target relative to camera;
Projecting cell, for the imaging process by perspective projection principle simulation circular target, by the circular target
Continuous circular curve projection obtains elliptic curve to the plane of delineation;
Discrete unit obtains discrete projection for carrying out grid discretization to the elliptic curve according to image resolution ratio
Point carries out simulation superposition to image error using monte carlo method on the discrete projection point, generates and carry random error
Discrete pixels point set curve;
Positioning unit, for carrying out segmental arc extraction, ellipse fitting and space circle pose to the discrete pixels point set curve
Identification, obtains the deviation spatial pose state of circular target;
Error calculation unit, the mistake of deviation spatial pose state and original state for obtaining circular target by comparison
Residual quantity determines positioning accuracy according to the margin of error.
Another technical solution that the present invention solves above-mentioned technical problem is as follows:A kind of storage medium is stored thereon with calculating
Machine program realizes the method described in said program when the program is executed by processor.
Another technical solution that the present invention solves above-mentioned technical problem is as follows:A kind of processing equipment, the processing equipment packet
It includes:
One or more processors;
Memory, for storing one or more programs,
When one or more of programs are executed by one or more of processors so that one or more of processing
Device realizes the method as described in said program.
The beneficial effects of the invention are as follows:With existing precision analysis compared with appraisal procedure, the present invention need not derive fixed
The explicit expression relationship of position precision and circular target between orientation, but to the error source in target imaging and image procossing
It is simulated, the positioning accuracy under statistical error existence condition.It is difficult to solve multiple error source coupling in circular target vision positioning
Can be the selection in monocular vision design of measuring system to circular target feature and profit with the difficulty of separation analysis and qualitative assessment
With offer important support.
Description of the drawings
Fig. 1 is the circular target vision positioning precision assessment method flow chart that the embodiment of the present invention proposes;
Fig. 2 is the schematic diagram of difference original state corresponding circular target and camera in the embodiment of the present invention;
Fig. 3 is the signal to the description of the spatial position of circular target, posture and size sampling process in the embodiment of the present invention
Figure;
Fig. 4 is the schematic diagram that space circular target is projected to the plane of delineation in the embodiment of the present invention;
Fig. 5 is the process for the discrete pixels point set curve that continuous segmental arc is transformed to superposition random error in the embodiment of the present invention
Schematic diagram;
Fig. 6 is that the elliptic curve based on discrete pixels point set is fitted schematic diagram in the embodiment of the present invention;
Fig. 7 is the circular target vision positioning accuracy evaluation device block diagram that the embodiment of the present invention proposes.
In attached drawing, parts list represented by the reference numerals are as follows:
1, camera, 2, circular target.
Specific implementation mode
The principle and features of the present invention will be described below with reference to the accompanying drawings, and the given examples are served only to explain the present invention, and
It is non-to be used to limit the scope of the present invention.
Fig. 1 gives a kind of schematic stream of circular target vision positioning precision assessment method provided in an embodiment of the present invention
Cheng Tu.As shown in Figure 1, this method includes:
S101, original state of the setting circular target relative to camera;
S102, by the imaging process of perspective projection principle simulation circular target, by the continuous circular of the circular target
Curve projection obtains elliptic curve to the plane of delineation;
S103 carries out grid discretization to the elliptic curve according to image resolution ratio, discrete projection point is obtained, described
Simulation superposition is carried out to image error using monte carlo method on discrete projection point, generates the discrete pixels with random error
Point set curve;
S104 carries out segmental arc extraction, ellipse fitting and space circle pose to the discrete pixels point set curve and identifies, obtains
The deviation spatial pose state of circular target;
S105 obtains the margin of error of the deviation spatial pose state and original state of circular target by comparison, according to institute
It states the margin of error and determines positioning accuracy.
It should be noted that original state of multigroup circular target relative to camera can be set in step S101, it is right
Step S102 to S105 is executed respectively in every group of original state, respectively obtains the corresponding positioning of each group of circular target original state
Error.As shown in Fig. 2, the corresponding circular target 2 of every group of original state relative to camera 1 have different distance, different directions and
Different sizes.
The method that above-described embodiment provides, need not derive the explicit table of positioning accuracy and circular target between orientation
Up to relationship, but target imaging and the error source in image procossing are simulated, the positioning accurate under statistical error existence condition
Degree.It solves the coupling of multiple error source in circular target vision positioning to be difficult to detach the difficulty of analysis and qualitative assessment, can be monocular
In vision measurement system design important support is provided with utilization to the selection of circular target feature.
Optionally, as one embodiment of the invention, S101, original state of the setting circular target relative to camera, tool
Body includes:
Centered on camera photocentre, using viewing field of camera angle as boundary, the envelope that four planes of camera photocentre are constituted was determined
Close pyramid region, to the center location of circular target in the closing pyramid region uniform sampling, to the normal direction of circular target
The uniform sampling within the scope of predetermined angle sets the size of circular target, and acquisition includes circular target spatial pose and size
Original state, wherein circular target spatial pose includes spatial position and the normal direction of circular target.Wherein, the method for circular target
To including yaw angle and pitch angle.
Fig. 3 is the schematic diagram described to the spatial position of circular target, posture and size sampling process.First to round mesh
Closing pyramid is carried out cutting division, cut surface by target center location uniform sampling in closing pyramid according to certain interval
Sampled point of the intersection point as center location, setting coordinate is (x, y, z).Secondly, to any one position sampled point (x, y,
Z), the normal direction of the circular target is sampled according to two dimensions of yaw angle α and pitch angle β, to yaw angle α [0,
180 °] uniform sampling in range, to pitch angle β, uniform sampling, five dimensions for obtaining (x, y, z, α, β) are adopted in [0,90 °] range
Sample state.It considers further that the setting of circular target size, i.e., value is set in a certain range to the radius R of circular target, formed
A large amount of sextuple space sample states (R, x, y, z, α, β).
In the embodiment, setting camera is that static, circular target moves in viewing field of camera, this will be moved camera in practice
Static mobile relationship carries out reverse setting with circular target, and this aspect is by the setting office of camera and circular target relative pose
Limit indicates in viewing field of camera so that it is more intuitive, clear to traverse all possible camera and circular target relative pose;It is another
Aspect, more conducively according to the difference of the settings such as position, posture and size, analyze camera to circular target measurement accuracy with away from
From, the variation of posture and size.
Optionally, as one embodiment of the invention, S102 passes through the imaging of perspective projection principle simulation circular target
Journey obtains elliptic curve by the continuous circular curve projection of the circular target to the plane of delineation.In the embodiment, round mesh
Known to target spatial position, normal direction and size.
Fig. 4 presents the schematic diagram that space circular target is projected to the plane of delineation.Two equal sized circles in space
Shape target, which can project in the plane of delineation, forms same elliptical image, therefore utilizes image ellipse equation and circular target size
It solves circular target pose and two groups of solutions can be obtained.With reference to opposite positions of the Fig. 2 between circular target camera in S102 steps
It sets and is introduced with the solution process of shaft orientation relation.First, circular target is imaged in the picture meets perspective projection principle, if u
For the point in image on oval camber line, then meet following formula:
U=MpXc=Mp[R(θ)-R(θ)t]Xw, (2-1)
Wherein, u=(u, v, 1)TIndicate on curve that any point projects to the point coordinates of camera image plane, Ω=(t,
θ) indicate translation and rotation of the camera coordinates system relative to world coordinate system, t=[tx ty tz] indicate camera coordinates system relative to
The translation of world coordinate system, θ=[α, beta, gamma] indicate that rotation of the camera coordinates system relative to world coordinate system, R (θ) indicate camera
Spin matrix of the coordinate system relative to world coordinate system, Xc=(Xc,Yc,Zc)T, Xw=(Xw,Yw,Zw,1)TRound mesh is indicated respectively
Mark homogeneous coordinates of any point in camera coordinates system and world coordinate system, Mp∈R3×3It indicates from lunar surface observation point to image
Perspective projection transformation matrix,
Wherein, f indicates the camera focus as unit of pixel, (u0,v0) it is image principal point coordinate.Hypothetical world coordinate system
It establishes at the center of target circle, Z-direction is directed toward camera photocentre perpendicular to disc, then θ=[α, β, 0], so as to
Wherein, c α, c β, s α, s β indicate that cos α, cos β, sin α, sin β ,@indicate to define in the matrix before equation respectively
For the subsequent symbolic indication of equation, i.e. internal matrix corresponding element is equal.
If the radius size of circular target is r0, then the circular curve equation in world coordinate system be
Enable x=Xw-tx, y=Yw-ty, z=Zw-tz=-tz, then can obtain circular target the subpoint u of the plane of delineation table
It reaches:
Wherein, C1=a1tx+b1ty+c1tz, C2=a2tx+b2ty+c2tz, C3=a3tx+c3tz。
Simultaneous (2-3) and (2-4) can be obtained:
Wherein, u '=u-u0, v '=v-v0.Formula (2-5) is brought into formula (2-3) to obtain:
A0u′2+B0u′v′+C0v′2+D0u′+E0v′+F0=0, (2-6)
Wherein A0、B0、C0、D0、E0And F0Respectively:
The dihydric phenol curvilinear equation of formula (2-6) description corresponding continuous curve in the plane of delineation is to project to image to put down
The elliptic curve in face.
In the embodiment, by deriving three-dimensional space curve to the projection relation of two dimensional surface, space circle mesh is formd
Be marked on the full curve equation projected in two dimensional image, to by space circle target imaging and discretization expression in chance error
Difference setting, which is limited in two dimensional image plane, to be carried out, and problem is effectively simplified, to be imaged the simulation preparatory condition of random error.
Optionally, as one embodiment of the invention, S103 carries out grid according to image resolution ratio to the elliptic curve
Discretization carries out simulation superposition to image error using monte carlo method, generates the discrete pixels point set with random error
Curve specifically includes:The plane of delineation is divided into a grids of (M-1) × (N-1) according to image resolution ratio M × N, is single with grid
Member carries out segment processing to the elliptic curve, the elliptic curve is divided into multiple segmental arc units, by every segmental arc unit
Midpoint be defined as circular target continuous circular curve discrete projection point, on the discrete projection point utilize Monte Carlo
Analogy method is superimposed the random error for meeting probability distribution, and correspondence is assigned in four pixels around the discrete projection point,
Generate the discrete pixels point set curve with random error.
Fig. 5 is the process schematic for the discrete pixels point set curve that continuous segmental arc is transformed to superposition random error.In figure
The grid cell that adjacent four pixels surround in image when grid, entire image are divided into (M-1) × (N- according to resolution ratio
1) grid.Segment processing is carried out to elliptic curve using grid as unit, elliptic curve is divided into multiple small segmental arc units.
Its implementation is:1) plane of delineation is divided into the grid of (M-1) × (N-1) according to image resolution ratio, wherein M is image slices
Plain line number, N are image pixel columns, and the region that adjacent rows and adjacent column surround is 1 grid cell;2) as 1. the of Fig. 5 walks,
Segment processing is carried out to elliptic curve using grid as unit, the small segmental arc fallen into single grid cell is a segmental arc list
Elliptic curve is divided into the small segmental arc unit of elliptic curve by member, and the midpoint of every small segmental arc is defined as circular curve
Discrete projection point.The endpoint of each small segmental arc unit can solve in the following manner:
1) in the plane of delineation, every grid straight line is described using a linear equation, for example, lateral grid linear equation is
U=i (i ∈ [1, M]), intersects with elliptic curve, and intersection point is located between two adjacent longitudinal grids, it is assumed that in v ∈ [j1,j1
+ 1) between, wherein j1∈[1,N];Longitudinal grid linear equation is v=j (j ∈ [1, N]), is intersected with elliptic curve, intersection point
Between two adjacent lateral grids, it is assumed that in u ∈ [i1,i1+ 1) between, wherein i1∈[1,M]。
2) overlapping relation of computation grid linear equation and the elliptic curve equation of formula (2-6) description, by u=i and v
=j is updated in formula (2-6), the quadratic equation with one unknown about v and u can be obtained, respectively:
C0v′2+[B0(i-u0)+E0]v′+A0(i-u0)2+D0(i-u0)+F0=0,
A0u′2+[B0(j-v0)+D0]u′+C0(j-v0)2+E0(j-v0)+F0=0.
2 groups of solutions, as two intersection points of grid straight line and elliptic curve can be respectively obtained by solving above-mentioned equation.Convert i and j
Value, acquire all grid straight lines and elliptical intersection point point set.
3) all intersection points are searched for from any one intersection point, successively along elliptic curve, between two adjoining nodes
Segmental arc be a small segmental arc unit, adjoining nodes are two endpoints of small segmental arc unit.
Meeting a certain probability distribution (such as normal distribution) using Monte-carlo Simulation Method superposition on discrete projection point
Random error, correspondence are assigned in four pixels around discrete projection point, generate the discrete pixels point set with random error
Curve, its implementation are as follows:Consider the gray scale (or brightness) of single pixel in image by pixel position deviation, photoperceptivity
And the influence of dark noise, to the gray scales of four pixels around plane of delineation discrete projection point value as follows:
1) as 2. the of Fig. 5 walks, the influence of pixel position deviation is considered, as discrete projection point (ui,vi) position deviation it is full
Sufficient normal distribution P (x, y) generates random sequence (the Δ xi of error according to normal distribution P (x, y)j,Δyij), by (Δ xij,Δ
yij) the corresponding discrete projection point (ui, vi) that is added to obtains new discrete projection point position (xi,yi)=(ui,vi)+(Δxij,Δ
yij)。
2) as 3. the of Fig. 5 walks, by new discrete projection point (xi,yi) at brightness according to bilinear transformation relation allocation
To four pixels of surrounding, new discrete projection point (x is rememberedi,yi) gray scale value be Ii, the coordinate of four pixels of surrounding is
(ui,v)、(ui+1,v)、(ui,v+1)、(ui+1,vi+ 1), then the gray scale value of four pixels is respectively:(ui+1-xi)(vi+
1-yi)Ii、(xi-ui)(vi+1-yi)Ii、(ui+1-xi)(yi-vi)Ii、(xi-ui)(yi-vi)Ii。
3) as 4. the of Fig. 5 walks, consider the influence of pixel photoperceptivity and dark noise, gray scale is carried out respectively to four pixels
Randomness superposition and decaying, setting dark noise w meets [0, Ib] normal distribution in range, photosensitive attenuation rate η meets [ηr,1]
Be uniformly distributed, then the gray scale value of each pixel becomes:wi1+ηi1(ui+1-xi)(vi+1-yi)Ii、wi2+ηi2(xi-ui)(vi+
1-yi)Ii、wi3+ηi3(ui+1-xi)(yi-vi)Ii、wi4+ηi4(xi-ui)(yi-vi)Ii。
1. the of Fig. 5 walks, indicate that using grid be unit to elliptic curve progress segment processing, fall into single grid cell
Small segmental arc be a segmental arc unit, elliptic curve is divided into the small segmental arc unit of elliptic curve, every small segmental arc
Midpoint be defined as the discrete projection point of circular curve;
2. the of Fig. 5 walks, indicate the translation after discrete projection point overlay error, the stain in grid is obtained after translation;
3. the of Fig. 5 walks, indicate that distribution of the stain in grid to four pixels around, four pixels distances are intermediate
Point is closer, and brightness is higher (corresponding spot size is bigger);
4. the of Fig. 5 walks, indicate around four pixel brightness carry out the randomness superposition and decaying of gray scale, brightness respectively
The size of corresponding points changes after change.
Based on above-mentioned steps, the discrete pixels point set curve with random error, the main body of superposition of error are finally obtained
Now in the gray scale value of pixel.Discrete point set is refined thus, phase is retained using image border method for tracing
The larger pixel of gray scale value in adjacent pixel rejects the smaller pixel of gray scale value, obtains the refinement edge point set of single pixel.
Optionally, as one embodiment of the invention, S104, to the discrete pixels point set curve carry out segmental arc extraction,
Ellipse fitting and the identification of space circle pose, the deviation spatial pose state for obtaining circular target include:
Edge single pixel refinement is carried out to the discrete pixels point set curve, obtains the refinement edge point set of single pixel;It is right
The refinement edge point set of the single pixel carries out ellipse fitting, obtains the elliptic curve equation being fitted in the plane of delineation;Based on institute
It states elliptic curve equation to establish using camera photocentre as vertex, by the ellipse conical surface normal equation of elliptic curve, utilizes the ellipse cone
Face normal equation seeks two sections identical with the circular target size, and circular target phase is determined according to described two sections
For two groups of spatial poses of camera, according to the original state of known circular target, it is round mesh to choose small one group of difference
Deviation spatial pose state.
Edge single pixel refinement is carried out to the discrete pixels point set curve in S103 steps first in S104, thinning method is
The pixel that gray scale value is larger in adjacent pixel is retained using image border method for tracing, rejects the smaller pixel of gray scale value,
Obtain the refinement edge point set of single pixel.
Then it carries out oval feature extraction respectively again and positions two steps with the space circular target based on oval feature, after
The two steps are introduced successively in face.
Fig. 6 is the elliptic curve fitting schematic diagram based on discrete pixels point set.It is only gived in figure quasi- for elliptic curve
The some discrete pixel of conjunction.Discrete pixels are clicked using least square curve fitting algorithm in the embodiment and carry out oval intend
It closes, thought is the ensemble average to all pixels for participating in fitting, and fitting precision can reach sub-pixel.To discrete picture
Vegetarian refreshments carries out the method that least square fitting seeks elliptic equation characterising parameter and is described as follows.
Elliptic equation can be described by the conic section of belt restraining, and general type is:
Enable Θ=[A B C D E F], u=[x2 2xy y22x 2y 1], the coefficient Θ in above-mentioned formula can basis
Scale factor difference is generated without array solution, does not influence the value of u.In order to realize the uniqueness of Θ solutions, a scale may specify
The factor, such as F=1 or | | Θ | |=1.Fitting solution procedure provides as follows.
It enablesI=1,2 ..., N indicate the ith pixel point on image, then root
Optimization problem can be obtained according to formula (4-1):
WhereinIndicate all pixels point set, H is by AC-B2The constraint matrix that > 0 is obtained, definition
For:
Optimization problem in formula (4-2) is solved, by introducing Lagrange factor λ, by the optimization of belt restraining
Problem is converted into unconstrained problem, and can be obtained to quadratic objective function progress derivative operation:
Enable S=UTU considers the particularity of U, S and H, we carry out the partitioning of matrix and obtain:
According to above-mentioned piecemeal relationship, matrix equation can be rebuild:
S1Θ1+S2Θ2=λ H1Θ1, (4-5)
By formula (4-6) it is found that working as S3For nonsingular matrix when, Θ2It is represented by Θ1Linear transformation, i.e.,Carrying it into formula (4-5) can obtain:
Due to H1For nonsingular matrix, formula (4-7) can be write as:
Equality constraint in formula (4-4) is rewritten as:
Θ1H1Θ1=1. (4-9)
It enablesThen ellipse fitting Solve problems, which are converted into, solves M Θ1=λ Θ1The problem of, i.e.,
The problem of seeking the characteristic value of matrix M.In view of Θ1H1Θ1The constraint of=1 pair of characteristic value nonnegativity takes the non-negative characteristic value of M
Corresponding feature vector is as Θ1Solution.According toAcquire Θ2, to which obtain elliptic curve general equation is
Number Θ are to get to the characterising parameter A, B, C, D, E, F of conic section elliptic equation.
The spatial position of location algorithm resolving circular target and the algorithm of normal direction based on oval feature provide as follows:It is oval
Camber line point ujMeet elliptic equation in the plane of delineation:
It is converted into the form of matrix:
The curved surface side for the elliptic cone that camera photocentre and space circle are formed can be obtained by bringing formula (2-1) into formula (4-11)
Journey:
It enablesThen elliptic cone surface equation can be write:
Under camera coordinates system, the expression formula of circular cone is more complicated, the constant description of cutting planes and calculating.Therefore first by phase
Machine coordinate system space is transformed into normed space and calculates, in result of calculation reconvert to camera coordinates system.Rotary course is fixed oval
Origin is bored still at camera optical center.It is symmetrical matrix it is found that there are orthogonal matrix P to obtain Q diagonalization by Q:
PTQP=diag (λ1,λ2,λ3),
Wherein λ1,λ,2λ3For the characteristic value of Q, then the point in new coordinate space can be obtainedIt is Z to obtain rotary shaftcsThe standard ellipse of axis bores surface equation:
λ1Xcs+λ2Ycs+λ3Zcs=0. (4-14)
By formula (4-14) it is found that three coefficient lambdas1,λ2,λ3It is inevitable that there are two jack per lines, and with another contrary sign.Due to λ1,
λ2,λ3The problem of with the positive and negative value of eigenvectors matrix P so that λ1,λ2,λ3There is multigroup solution.If the characteristic value of Q is κ1,κ2,κ3,
It is v with only corresponding feature vector1,v2,v3.If κ1,κ2Jack per line and | κ1| > | κ2|, then λ1=κ1, λ2=κ2, λ3=κ3.Enable P=
[p,1p2,p3], ifThen p1=v1, otherwise p1=-v1。p2=v2, p2=v2×v3。
Based on the elliptic cone surface equation in normed space, the normal vector of plane is where can acquiring center location and circle:
The result of formula (4-15) and formula (4-16) is transformed into the spatial position that camera coordinate system obtains circular target
And normal direction:
Relative position and shaft orientation relation as between circular target camera.
In the embodiment, least square fitting elliptic curve is used based on discrete pixels point in image, it can be by whole
The average mode of body reduces the random error in the expression of curve imaging discreteization so that the position essence of elliptic curve in the picture
Degree reaches sub-pixel.On the other hand, by way of elliptic cone Analytical Expression, the fixed ruler intersected with elliptic cone is sought
Very little circular cross-section, and then the method for obtaining circular target spatial position are Analytical Solution method, do not introduce uncertain miss
Difference, therefore the positioning accuracy from plane of delineation elliptic curve inverting extraterrestrial target pose can be improved to the full extent.
Optionally, as one embodiment of the invention, S105 obtains the deviation spatial pose shape of circular target by comparison
The margin of error of state and original state determines positioning accuracy according to the margin of error.
Optionally, as an embodiment of the present invention, further include the steps that repeating preset times S103 and S104, to obtain
Multiple margins of error of the deviation spatial pose state and original state of circular target are obtained, and calculate the mathematics of the multiple margin of error
It is expected that using the mathematic expectaion as the corresponding position error of current spatial pose.
Specifically, the mathematic expectaion for calculating the margin of error is the corresponding position error of current spatial pose.Assuming that carrying out M time
Simulation calculates, each volumetric position error epWith normal direction attitude error enComputational methods be respectively:
Wherein,With Δ nkThe circular target spatial position and method that kth time spatial pose recognizer obtains are indicated respectively
To error,Indicate the circular target spatial position that kth time space circle pose recognizer obtains,Indicate kth time space
The circular target space normal direction that circle pose recognizer obtains,For initially set circular target spatial position,Initially set
Fixed circular target space normal direction.
In the embodiment, by obtaining the margins of error multiple enough, various ellipse fittings and position can be largely covered
The error condition of appearance positioning can be averaged to various random error situations, obtain by calculating the mathematic expectaion of the margin of error
The mean error that spatial pose calculates.By the margin of error of the enough quantity of analog selection, it can realize that spatial pose averagely misses
The accurate calculating of difference.
The circular target vision positioning precision provided according to embodiments of the present invention is provided above in association with Fig. 1 to Fig. 6
Appraisal procedure.Circular target vision positioning accuracy evaluation device provided in an embodiment of the present invention is described in detail with reference to Fig. 7.It should
Device includes setup unit, projecting cell, discrete unit, positioning unit and error calculation unit.
Wherein, original state of the setup unit setting circular target relative to camera;Projecting cell passes through perspective projection original
The imaging process of reason simulation circular target obtains ellipse by the continuous circular curve projection of the circular target to the plane of delineation
Curve;Discrete unit carries out grid discretization according to image resolution ratio to the elliptic curve, discrete projection point is obtained, described
Simulation superposition is carried out to image error using monte carlo method on discrete projection point, generates the discrete pixels with random error
Point set curve;Positioning unit carries out segmental arc extraction, ellipse fitting and space circle pose to the discrete pixels point set curve and identifies,
Obtain the deviation spatial pose state of circular target;Error calculation unit obtains the deviation spatial pose of circular target by comparison
The margin of error of state and original state determines positioning accuracy according to the margin of error.
It should be noted that original state of multigroup circular target relative to camera can be arranged in setup unit, for every
Group original state calls projecting cell, discrete unit, positioning unit and error calculation unit respectively, respectively obtains each group of circle
The corresponding position error of target original state.Wherein, the corresponding circular target of every group of original state has difference relative to camera
Distance, different directions and different sizes.
The device that above-described embodiment provides, need not derive the explicit table of positioning accuracy and circular target between orientation
Up to relationship, but target imaging and the error source in image procossing are simulated, the positioning accurate under statistical error existence condition
Degree.It solves the coupling of multiple error source in circular target vision positioning to be difficult to detach the difficulty of analysis and qualitative assessment, can be monocular
In vision measurement system design important support is provided with utilization to the selection of circular target feature.
Optionally, as one embodiment of the invention, setup unit, which has, to be used for, centered on camera photocentre, with camera
Field angle is boundary, the closing pyramid region that four planes of camera photocentre are constituted is determined, to the center location of circular target
The uniform sampling in the closing pyramid region, to the normal direction of circular target within the scope of predetermined angle uniform sampling, setting circle
The size of shape target, acquisition include the original state of circular target spatial pose and size, wherein circular target spatial pose packet
Include spatial position and the normal direction of circular target.
Optionally, as one embodiment of the invention, discrete unit, which has, to be used for, according to image resolution ratio M × N by image
Plane is divided into a grids of (M-1) × (N-1), and segment processing is carried out to the elliptic curve using grid as unit, will be described ellipse
Circular curve is divided into multiple segmental arc units, the midpoint of every segmental arc unit is defined as the continuous circular curve of circular target from
Subpoint is dissipated, meets the random error of probability distribution using Monte-carlo Simulation Method superposition on the discrete projection point, it is right
It should be assigned in four pixels around the discrete projection point, generate the discrete pixels point set curve with random error.
Optionally, as one embodiment of the invention, discrete unit, which has, to be used for:
Discrete projection point (ui,vi) position deviation meet probability distribution P (x, y), according to probability distribution P (x, y) generate mistake
Random sequence (the Δ x of differenceij,Δyij), by (Δ xij,Δyij) be added to corresponding discrete projection point (ui,vi) obtain it is new discrete
Subpoint position (xi,yi)=(ui,vi)+(Δxij,Δyij)。
By new discrete projection point (xi,yi) at brightness according to bilinear transformation relation allocation to four pixels of surrounding,
Remember new discrete projection point (xi,yi) gray scale value be Ii, the coordinate of four pixels of surrounding is (ui,vi)、(ui+1,vi)、
(ui,vi+1)、(ui+1,vi+ 1), then the gray scale value of four pixels is respectively:(ui+1-xi)(vi+1-yi)Ii、(xi-ui)
(vi+1-yi)Ii、(ui+1-xi)(yi-vi)Ii、(xi-ui)(yi-vi)Ii。
Four pixels are carried out with the randomness superposition and decaying of gray scale respectively, setting dark noise w meets [0, Ib] in range
Probability distribution, photosensitive attenuation rate η meets [ηr, 1] be uniformly distributed, then the gray scale value of each pixel becomes:wi1+ηi1(ui+
1-xi)(vi+1-yi)Ii、wi2+ηi2(xi-ui)(vi+1-yi)Ii、wi3+ηi3(ui+1-xi)(yi-vi)Ii、wi4+ηi4(xi-ui)(yi-
vi)Ii。
Optionally, as one embodiment of the invention, positioning unit is specifically used for:To the discrete pixels point set curve into
The single pixel refinement of row edge, obtains the refinement edge point set of single pixel;The refinement edge point set of the single pixel is carried out oval
Fitting obtains the elliptic curve equation being fitted in the plane of delineation;It is top to be established with camera photocentre based on the elliptic curve equation
Point, the ellipse conical surface normal equation by elliptic curve, are sought and the circular target size using the ellipse conical surface normal equation
Identical two sections determine two group spatial poses of the circular target relative to camera, according to known according to described two sections
Circular target original state, choose deviation spatial pose state of the difference small one group for round purpose.
Optionally, as another embodiment of the present invention, error calculation unit is additionally operable to repeat to call discrete unit and positioning
Unit to obtain the deviation spatial pose state of circular target and multiple margins of error of original state, and calculates the multiple mistake
The mathematic expectaion of residual quantity, using the mathematic expectaion as the corresponding position error of current spatial pose.
The embodiment of the present invention also provides a kind of storage medium, is stored thereon with computer program, which is held by processor
The circular target vision positioning precision assessment method that above-described embodiment provides is realized when row.
In this embodiment, computer readable storage medium, the program realize that above-described embodiment carries when being executed by processor
The step of circular target vision positioning precision assessment method of confession, therefore any one of the proposition of the embodiment with foregoing invention
Whole advantageous effects of circular target vision positioning precision assessment method, details are not described herein.
The embodiment of the present invention also provides a kind of processing equipment, and the processing equipment includes:One or more processors;Storage
Device, for storing one or more programs, when one or more of programs are executed by one or more of processors so that
One or more of processors realize the circular target vision positioning precision assessment method provided such as above-described embodiment.
In the embodiment, processing equipment includes processor, and processor is for executing the computer program stored in memory
The step of circular target vision positioning precision assessment method that Shi Shixian such as above-described embodiments provide, therefore with foregoing invention
Whole advantageous effects of the circular target vision positioning precision assessment method for any one that embodiment proposes, details are not described herein.
It is apparent to those skilled in the art that for convenience of description and succinctly, the dress of foregoing description
The specific work process with unit is set, can refer to corresponding processes in the foregoing method embodiment, details are not described herein.
In several embodiments provided herein, it should be understood that disclosed device and method can pass through it
Its mode is realized.For example, the apparatus embodiments described above are merely exemplary, for example, the division of unit, only
A kind of division of logic function, formula that in actual implementation, there may be another division manner, such as multiple units or component can combine or
Person is desirably integrated into another system, or some features can be ignored or not executed.
The unit illustrated as separating component may or may not be physically separated, and be shown as unit
Component may or may not be physical unit, you can be located at a place, or may be distributed over multiple networks
On unit.Some or all of unit therein can be selected according to the actual needs to realize the mesh of the embodiment of the present invention
's.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, it can also
It is that each unit physically exists alone, can also be during two or more units are integrated in one unit.It is above-mentioned integrated
The form that hardware had both may be used in unit is realized, can also be realized in the form of SFU software functional unit.
It, can if integrated unit is realized in the form of SFU software functional unit and when sold or used as an independent product
To be stored in a computer read/write memory medium.Based on this understanding, technical scheme of the present invention substantially or
Say that all or part of the part that contributes to existing technology or the technical solution can embody in the form of software products
Out, which is stored in a storage medium, including some instructions are used so that a computer equipment
(can be personal computer, server or the network equipment etc.) executes all or part of each embodiment method of the present invention
Step.And storage medium above-mentioned includes:It is USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random
Access various Jie that can store program code such as memory (RAM, Random Access Memory), magnetic disc or CD
Matter.
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all the present invention spirit and
Within principle, any modification, equivalent replacement, improvement and so on should all be included in the protection scope of the present invention.
Claims (14)
1. a kind of circular target vision positioning precision assessment method, which is characterized in that include the following steps:
Set original state of the circular target relative to camera;
By the imaging process of perspective projection principle simulation circular target, the continuous circular curve projection of the circular target is arrived
The plane of delineation obtains elliptic curve;
Grid discretization is carried out to the elliptic curve according to image resolution ratio, discrete projection point is obtained, in the discrete projection
Simulation superposition is carried out to image error using monte carlo method on point, it is bent to generate the discrete pixels point set with random error
Line;
Segmental arc extraction, ellipse fitting and the identification of space circle pose are carried out to the discrete pixels point set curve, obtain circular target
Deviation spatial pose state;
The margin of error of the deviation spatial pose state and original state of circular target is obtained by comparison, it is true according to the margin of error
Determine positioning accuracy.
2. according to the method described in claim 1, it is characterized in that, original state of the setting circular target relative to camera
Including:
Centered on camera photocentre, using viewing field of camera angle as boundary, the closing rib that four planes of camera photocentre are constituted was determined
Bore region, to the center location of circular target in the closing pyramid region uniform sampling, to the normal direction of circular target pre-
If uniform sampling in angular range, the size of circular target is set, it includes the initial of circular target spatial pose and size to obtain
State, wherein circular target spatial pose includes spatial position and the normal direction of circular target.
3. according to the method described in claim 1, it is characterized in that, described carry out the elliptic curve according to image resolution ratio
Grid discretization obtains discrete projection point, and mould is carried out to image error using monte carlo method on the discrete projection point
Quasi- superposition, generating the discrete pixels point set curve with random error includes:
The plane of delineation is divided into a grids of (M-1) × (N-1) according to image resolution ratio M × N, is unit to described ellipse using grid
Circular curve carries out segment processing, and the elliptic curve is divided into multiple segmental arc units, and the midpoint of every segmental arc unit is defined
It is folded using Monte-carlo Simulation Method on the discrete projection point for the discrete projection point of the continuous circular curve of circular target
Fill it up with the random error of sufficient probability distribution, correspondence is assigned in four pixels around the discrete projection point, generate with
The discrete pixels point set curve of chance error difference.
4. according to the method described in claim 3, it is characterized in that, utilizing Monte Carlo simulation side on the discrete projection point
Method is superimposed the random error for meeting probability distribution, and correspondence is assigned in four pixels around the discrete projection point, generates band
There is the discrete pixels point set curve of random error to include:
Discrete projection point (ui,vi) position deviation meet probability distribution P (x, y), generate error according to probability distribution P (x, y)
Random sequence (Δ xij,Δyij), by (Δ xij,Δyij) be added to corresponding discrete projection point (ui,vi) obtain new discrete projection
Point position (xi,yi)=(ui,vi)+(Δxij,Δyij);
By new discrete projection point (xi,yi) at brightness according to bilinear transformation relation allocation to four pixels of surrounding, note is new
Discrete projection point (xi,yi) gray scale value be Ii, the coordinate of four pixels of surrounding is (ui,vi)、(ui+1,vi)、(ui,vi
+1)、(ui+1,vi+ 1), then the gray scale value of four pixels is respectively:(ui+1-xi)(vi+1-yi)Ii、(xi-ui)(vi+1-
yi)Ii、(ui+1-xi)(yi-vi)Ii、(xi-ui)(yi-vi)Ii;
Four pixels are carried out with the randomness superposition and decaying of gray scale respectively, setting dark noise w meets [0, Ib] probability in range
Distribution, photosensitive attenuation rate η meet [ηr, 1] be uniformly distributed, then the gray scale value of each pixel becomes:wi1+ηi1(ui+1-xi)
(vi+1-yi)Ii、wi2+ηi2(xi-ui)(vi+1-yi)Ii、wi3+ηi3(ui+1-xi)(yi-vi)Ii、wi4+ηi4(xi-ui)(yi-vi)
Ii。
5. method according to any one of claims 1 to 4, which is characterized in that described to the discrete pixels point set curve
Carrying out segmental arc extraction, ellipse fitting and the identification of space circle pose, the deviation spatial pose state for obtaining circular target includes:
Edge single pixel refinement is carried out to the discrete pixels point set curve, obtains the refinement edge point set of single pixel;
Ellipse fitting is carried out to the refinement edge point set of the single pixel, obtains the elliptic curve equation being fitted in the plane of delineation;
It is established using camera photocentre as vertex, by the ellipse conical surface normal equation of elliptic curve, profit based on the elliptic curve equation
Two sections identical with the circular target size are sought with the ellipse conical surface normal equation, are determined according to described two sections
Two group spatial poses of the circular target relative to camera, according to the original state of known circular target, choose difference it is small one
Group is the deviation spatial pose state of round purpose.
6. method according to any one of claims 1 to 4, which is characterized in that further include repeat preset times by
Grid discretization is carried out to the elliptic curve according to image resolution ratio, obtains discrete projection point, it is sharp on the discrete projection point
Simulation superposition is carried out to image error with monte carlo method, generates the discrete pixels point set curve with random error, and
Segmental arc extraction, ellipse fitting and the identification of space circle pose are carried out to the discrete pixels point set curve, obtain the inclined of circular target
The step of difference space position and posture, to obtain the deviation spatial pose state of circular target and multiple margins of error of original state,
And the mathematic expectaion of the multiple margin of error is calculated, using the mathematic expectaion as the corresponding position error of current spatial pose.
7. a kind of circular target vision positioning accuracy evaluation device, which is characterized in that including:
Setup unit, for setting original state of the circular target relative to camera;
Projecting cell, for the imaging process by perspective projection principle simulation circular target, by the continuous of the circular target
Circular curve projects to the plane of delineation, obtains elliptic curve;
Discrete unit, for, to elliptic curve progress grid discretization, obtaining discrete projection point according to image resolution ratio,
Simulation superposition is carried out to image error using monte carlo method on the discrete projection point, is generated discrete with random error
Pixel point set curve;
Positioning unit, for carrying out segmental arc extraction, ellipse fitting and the identification of space circle pose to the discrete pixels point set curve,
Obtain the deviation spatial pose state of circular target;
Error calculation unit, the error of deviation spatial pose state and original state for obtaining circular target by comparison
Amount, positioning accuracy is determined according to the margin of error.
8. device according to claim 7, which is characterized in that the setup unit, which has, to be used for, in being with camera photocentre
The heart determined the closing pyramid region that four planes of camera photocentre are constituted, to circular target using viewing field of camera angle as boundary
Center location uniform sampling in the closing pyramid region, uniformly adopts the normal direction of circular target within the scope of predetermined angle
Sample, sets the size of circular target, and acquisition includes the original state of circular target spatial pose and size, wherein circular target
Spatial pose includes spatial position and the normal direction of circular target.
9. device according to claim 7, which is characterized in that the discrete unit, which has, to be used for, according to image resolution ratio M
The plane of delineation is divided into a grids of (M-1) × (N-1) by × N, and segment processing is carried out to the elliptic curve using grid as unit,
The elliptic curve is divided into multiple segmental arc units, the midpoint of every segmental arc unit is defined as to the continuous circular of circular target
The discrete projection point of curve meets the random of probability distribution on the discrete projection point using Monte-carlo Simulation Method superposition
Error, correspondence are assigned in four pixels around the discrete projection point, generate the discrete pixels point set with random error
Curve.
10. device according to claim 8, which is characterized in that the discrete unit, which has, to be used for:
Discrete projection point (ui,vi) position deviation meet probability distribution P (x, y), generate error according to probability distribution P (x, y)
Random sequence (Δ xij,Δyij), by (Δ xij,Δyij) be added to corresponding discrete projection point (ui,vi) obtain new discrete projection
Point position (xi,yi)=(ui,vi)+(Δxij,Δyij);
By new discrete projection point (xi,yi) at brightness according to bilinear transformation relation allocation to four pixels of surrounding, note is new
Discrete projection point (xi,yi) gray scale value be Ii, the coordinate of four pixels of surrounding is (ui,vi)、(ui+1,vi)、(ui,vi
+1)、(ui+1,vi+ 1), then the gray scale value of four pixels is respectively:(ui+1-xi)(vi+1-yi)Ii、(xi-ui)(vi+1-
yi)Ii、(ui+1-xi)(yi-vi)Ii、(xi-ui)(yi-vi)Ii;
Four pixels are carried out with the randomness superposition and decaying of gray scale respectively, setting dark noise w meets [0, Ib] probability in range
Distribution, photosensitive attenuation rate η meet [ηr, 1] be uniformly distributed, then the gray scale value of each pixel becomes:wi1+ηi1(ui+1-xi)
(vi+1-yi)Ii、wi2+ηi2(xi-ui)(vi+1-yi)Ii、wi3+ηi3(ui+1-xi)(yi-vi)Ii、wi4+ηi4(xi-ui)(yi-vi)
Ii。
11. according to claim 7 to 10 any one of them device, which is characterized in that the positioning unit is specifically used for:
Edge single pixel refinement is carried out to the discrete pixels point set curve, obtains the refinement edge point set of single pixel;
Ellipse fitting is carried out to the refinement edge point set of the single pixel, obtains the elliptic curve equation being fitted in the plane of delineation;
It is established using camera photocentre as vertex, by the ellipse conical surface normal equation of elliptic curve, profit based on the elliptic curve equation
Two sections identical with the circular target size are sought with the ellipse conical surface normal equation, are determined according to described two sections
Two group spatial poses of the circular target relative to camera, according to the original state of known circular target, choose difference it is small one
Group is the deviation spatial pose state of round purpose.
12. according to claim 7 to 10 any one of them device, which is characterized in that error calculation unit is additionally operable to repeat to adjust
With discrete unit and positioning unit, to obtain the deviation spatial pose state of circular target and multiple margins of error of original state,
And the mathematic expectaion of the multiple margin of error is calculated, using the mathematic expectaion as the corresponding position error of current spatial pose.
13. a kind of storage medium, is stored thereon with computer program, which is characterized in that the program is realized when being executed by processor
Such as method according to any one of claims 1 to 6.
14. a kind of processing equipment, which is characterized in that the processing equipment includes:
One or more processors;
Memory, for storing one or more programs,
When one or more of programs are executed by one or more of processors so that one or more of processors are real
Now such as method according to any one of claims 1 to 6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810273909.2A CN108596980B (en) | 2018-03-29 | 2018-03-29 | Circular target visual positioning precision evaluation method and device, storage medium and processing equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810273909.2A CN108596980B (en) | 2018-03-29 | 2018-03-29 | Circular target visual positioning precision evaluation method and device, storage medium and processing equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108596980A true CN108596980A (en) | 2018-09-28 |
CN108596980B CN108596980B (en) | 2021-12-07 |
Family
ID=63624022
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810273909.2A Active CN108596980B (en) | 2018-03-29 | 2018-03-29 | Circular target visual positioning precision evaluation method and device, storage medium and processing equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108596980B (en) |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109760107A (en) * | 2019-01-22 | 2019-05-17 | 广东工业大学 | A kind of robot localization Accuracy Assessment based on monocular vision |
CN109949367A (en) * | 2019-03-11 | 2019-06-28 | 中山大学 | A kind of visual light imaging localization method based on circular projection |
CN110378970A (en) * | 2019-07-08 | 2019-10-25 | 武汉理工大学 | A kind of monocular vision deviation detecting method and device for AGV |
CN110674810A (en) * | 2019-08-30 | 2020-01-10 | 苏州悦谱半导体有限公司 | Optimization algorithm applied to semiconductor optical CCD vision |
CN110782492A (en) * | 2019-10-08 | 2020-02-11 | 三星(中国)半导体有限公司 | Pose tracking method and device |
CN111210413A (en) * | 2020-01-02 | 2020-05-29 | 北京机科国创轻量化科学研究院有限公司 | Pose detection method in motion process of wire feeding mechanism |
CN111380503A (en) * | 2020-05-29 | 2020-07-07 | 电子科技大学 | Monocular camera ranging method adopting laser-assisted calibration |
CN111780745A (en) * | 2020-06-29 | 2020-10-16 | 南京航空航天大学 | Short arc ellipse fitting optimization method for deep space exploration optical navigation |
CN112255869A (en) * | 2020-11-03 | 2021-01-22 | 成都景中教育软件有限公司 | Parameter-based three-dimensional graph dynamic projection implementation method |
CN112365531A (en) * | 2020-10-13 | 2021-02-12 | 西安理工大学 | Reliability evaluation method for ellipse detection result of automatic scrolling system |
CN112381880A (en) * | 2020-11-27 | 2021-02-19 | 航天科工智能机器人有限责任公司 | Binocular vision pose estimation method based on circle features |
CN112505671A (en) * | 2021-02-07 | 2021-03-16 | 北方工业大学 | Millimeter wave radar target positioning method and device under GNSS signal missing environment |
CN113064272A (en) * | 2021-03-04 | 2021-07-02 | 武汉大学 | Optical free-form surface construction method and system under semi-discrete optimal transmission |
CN115493499A (en) * | 2021-12-30 | 2022-12-20 | 北京航天飞行控制中心 | Method and system for assembling cylinder or cylinder-like body |
US11610330B2 (en) | 2019-10-08 | 2023-03-21 | Samsung Electronics Co., Ltd. | Method and apparatus with pose tracking |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104715471A (en) * | 2014-01-03 | 2015-06-17 | 杭州海康威视数字技术股份有限公司 | Target positioning and tracking method and device |
CN104729481A (en) * | 2015-03-12 | 2015-06-24 | 北京空间飞行器总体设计部 | Cooperative target pose precision measurement method based on PNP perspective model |
CN107167116A (en) * | 2017-03-13 | 2017-09-15 | 湖北汽车工业学院 | A kind of visible detection method of space circular arc pose |
-
2018
- 2018-03-29 CN CN201810273909.2A patent/CN108596980B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104715471A (en) * | 2014-01-03 | 2015-06-17 | 杭州海康威视数字技术股份有限公司 | Target positioning and tracking method and device |
CN104729481A (en) * | 2015-03-12 | 2015-06-24 | 北京空间飞行器总体设计部 | Cooperative target pose precision measurement method based on PNP perspective model |
CN107167116A (en) * | 2017-03-13 | 2017-09-15 | 湖北汽车工业学院 | A kind of visible detection method of space circular arc pose |
Cited By (25)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109760107A (en) * | 2019-01-22 | 2019-05-17 | 广东工业大学 | A kind of robot localization Accuracy Assessment based on monocular vision |
CN109760107B (en) * | 2019-01-22 | 2022-04-12 | 九天创新(广东)智能科技有限公司 | Monocular vision-based robot positioning accuracy evaluation method |
CN109949367A (en) * | 2019-03-11 | 2019-06-28 | 中山大学 | A kind of visual light imaging localization method based on circular projection |
CN109949367B (en) * | 2019-03-11 | 2023-01-20 | 中山大学 | Visible light imaging positioning method based on circular projection |
CN110378970A (en) * | 2019-07-08 | 2019-10-25 | 武汉理工大学 | A kind of monocular vision deviation detecting method and device for AGV |
CN110378970B (en) * | 2019-07-08 | 2023-03-10 | 武汉理工大学 | Monocular vision deviation detection method and device for AGV |
CN110674810A (en) * | 2019-08-30 | 2020-01-10 | 苏州悦谱半导体有限公司 | Optimization algorithm applied to semiconductor optical CCD vision |
CN110674810B (en) * | 2019-08-30 | 2023-04-18 | 苏州悦谱半导体有限公司 | Optimization method applied to semiconductor optical CCD vision |
CN110782492B (en) * | 2019-10-08 | 2023-03-28 | 三星(中国)半导体有限公司 | Pose tracking method and device |
US11610330B2 (en) | 2019-10-08 | 2023-03-21 | Samsung Electronics Co., Ltd. | Method and apparatus with pose tracking |
CN110782492A (en) * | 2019-10-08 | 2020-02-11 | 三星(中国)半导体有限公司 | Pose tracking method and device |
CN111210413B (en) * | 2020-01-02 | 2023-10-03 | 北京机科国创轻量化科学研究院有限公司 | Pose detection method in movement process of wire feeding mechanism |
CN111210413A (en) * | 2020-01-02 | 2020-05-29 | 北京机科国创轻量化科学研究院有限公司 | Pose detection method in motion process of wire feeding mechanism |
CN111380503A (en) * | 2020-05-29 | 2020-07-07 | 电子科技大学 | Monocular camera ranging method adopting laser-assisted calibration |
CN111780745A (en) * | 2020-06-29 | 2020-10-16 | 南京航空航天大学 | Short arc ellipse fitting optimization method for deep space exploration optical navigation |
CN112365531B (en) * | 2020-10-13 | 2022-08-12 | 西安理工大学 | Reliability evaluation method for ellipse detection result of automatic scrolling system |
CN112365531A (en) * | 2020-10-13 | 2021-02-12 | 西安理工大学 | Reliability evaluation method for ellipse detection result of automatic scrolling system |
CN112255869B (en) * | 2020-11-03 | 2021-09-14 | 成都景中教育软件有限公司 | Parameter-based three-dimensional graph dynamic projection implementation method |
CN112255869A (en) * | 2020-11-03 | 2021-01-22 | 成都景中教育软件有限公司 | Parameter-based three-dimensional graph dynamic projection implementation method |
CN112381880A (en) * | 2020-11-27 | 2021-02-19 | 航天科工智能机器人有限责任公司 | Binocular vision pose estimation method based on circle features |
CN112505671B (en) * | 2021-02-07 | 2021-05-25 | 北方工业大学 | Millimeter wave radar target positioning method and device under GNSS signal missing environment |
CN112505671A (en) * | 2021-02-07 | 2021-03-16 | 北方工业大学 | Millimeter wave radar target positioning method and device under GNSS signal missing environment |
CN113064272A (en) * | 2021-03-04 | 2021-07-02 | 武汉大学 | Optical free-form surface construction method and system under semi-discrete optimal transmission |
CN115493499A (en) * | 2021-12-30 | 2022-12-20 | 北京航天飞行控制中心 | Method and system for assembling cylinder or cylinder-like body |
CN115493499B (en) * | 2021-12-30 | 2024-04-19 | 北京航天飞行控制中心 | Cylinder or cylinder-like assembly method and system |
Also Published As
Publication number | Publication date |
---|---|
CN108596980B (en) | 2021-12-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108596980A (en) | Circular target vision positioning precision assessment method, device, storage medium and processing equipment | |
CN110443885B (en) | Three-dimensional human head and face model reconstruction method based on random human face image | |
JP6681729B2 (en) | Method for determining 3D pose of object and 3D location of landmark point of object, and system for determining 3D pose of object and 3D location of landmark of object | |
CN107179768B (en) | Obstacle identification method and device | |
CN105913489B (en) | A kind of indoor three-dimensional scenic reconstructing method using plane characteristic | |
CN106780619B (en) | Human body size measuring method based on Kinect depth camera | |
JP6740033B2 (en) | Information processing device, measurement system, information processing method, and program | |
KR102068419B1 (en) | Method, apparatus and computer readable medium for adjusting point cloud data collection trajectory | |
US7190826B2 (en) | Measuring the location of objects arranged on a surface, using multi-camera photogrammetry | |
CN105069743A (en) | Detector splicing real-time image registration method | |
CN111145227B (en) | Iterative integral registration method for space multi-view point cloud of underground tunnel | |
CN110136178B (en) | Three-dimensional laser point cloud registration method and device based on endpoint fitting | |
CN108154104A (en) | A kind of estimation method of human posture based on depth image super-pixel union feature | |
Kroemer et al. | Point cloud completion using extrusions | |
Hu et al. | Pipe pose estimation based on machine vision | |
CN113393524A (en) | Target pose estimation method combining deep learning and contour point cloud reconstruction | |
CN107025647A (en) | Distorted image evidence collecting method and device | |
CN112581511A (en) | Three-dimensional reconstruction method and system based on approximate vertical scanning point cloud rapid registration | |
CN110136179B (en) | Three-dimensional laser point cloud registration method and device based on straight line fitting | |
EP4004488A1 (en) | Improvements in or relating to photogrammetry | |
CN110458951A (en) | A kind of the modeling data acquisition methods and relevant apparatus of power grid shaft tower | |
Będkowski et al. | Towards terrestrial 3D data registration improved by parallel programming and evaluated with geodetic precision | |
CN115937466A (en) | Three-dimensional model generation method, system and storage medium integrating GIS | |
Chen et al. | The trajectory planning system for spraying robot based on k-means clustering and NURBS curve optimization | |
CN105740851A (en) | Three-dimensional face automatic positioning method and curved surface registration method and system |
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 |