CN102564347B - Object three-dimensional outline measuring device and method based on Dammann grating - Google Patents
Object three-dimensional outline measuring device and method based on Dammann grating Download PDFInfo
- Publication number
- CN102564347B CN102564347B CN201110456929.1A CN201110456929A CN102564347B CN 102564347 B CN102564347 B CN 102564347B CN 201110456929 A CN201110456929 A CN 201110456929A CN 102564347 B CN102564347 B CN 102564347B
- Authority
- CN
- China
- Prior art keywords
- image
- dimensional
- information
- point
- measuring
- 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.)
- Active
Links
Landscapes
- Length Measuring Devices By Optical Means (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides an object three-dimensional outline measuring device and method based on a Dammann grating. The device consists of a laser device, a collimating lens, the Dammann grating, a projection lens, two digital cameras, transmission lines and a computer. Laser emitted by the laser device passes through the collimating lens, the Dammann grating and the projection lens sequentially so as to form a rectangular lattice light beam to illuminate the surface of an object to be measured. Images of the measurement area of the object to be measured is collected by the two digital cameras and are input to the computer through the transmission lines; partial three-dimensional feature information of the object to be measured is obtained after the images are subjected to data processing by the computer; and measurement on full-area three-dimensional feature information of the object to be measured is realized by moving the object to be measured or changing the device position, so as to obtain object three-dimensional outline information. The object three-dimensional outline measuring method disclosed by the invention has the advantages of regular lattice, high signal to noise ratio, strong measurement adaptability, high measurement precision and the like, and the object three-dimensional outline measuring device disclosed by the invention has the advantages of simple structure, low cost, light weight, wide application field and the like. A three-dimensional outline restructuring process is conveniently simplified by the computer; the automatic measurement can be realized; and the object three-dimensional outline measuring device and method have the advantages of important practical value and prospect.
Description
Technical field
The present invention relates to object three-dimensional contour outline and measure, particularly a kind of object three-dimensional contour outline measuring set and measuring method based on Darman raster.
Background technology
Along with domestic manufacturing industry is manufactured striding forward of field to numeral, the demand that complex-curved 3 d shape is measured is also more and more urgent, object three-dimensional contour outline measuring technique at home manufacturing application is also more extensive, as fields such as Aero-Space, auto industry, clinical treatments.In many measuring three-dimensional profile technology, active optical measuring three-dimensional profile technology, due to its noncontact, harmless, the whole audience, high resolving power with the advantage such as easily be automated, has obtained people's very big concern especially.And people recognize the importance of robot vision field three-dimensional measurement equally in recent years, the research of active optical measuring three-dimensional profile technology is increased day by day.
Existing active optical measuring three-dimensional profile technology mainly contains Fourier transform, phase measurement, gray code method and color coding method etc.
Fourier transform is simple in structure and measuring accuracy is higher, but for complex-curved, can produce leak, obscure the error causing with fence effect, and easily there is aliasing and bring frequency leakage and phase unwrapping discontinuous in projection lines, cannot obtain valid data.
Thereby the image that phase measurement obtains several outs of phase by phase-shifting technique is realized accurate phase unwrapping and is obtained higher measuring accuracy, but phase-shifting technique Measuring Time is long, in measuring process, require sample and measurement mechanism position relationship strictly constant, to actual measurement, make troubles and many restrictions.
Coloud coding method is usingd coloud coding pattern or gray-coded pattern loading and the tool for transmitting as object dimensional information, by colouring information or gray scale being analyzed to decoding, obtain three-D profile data, it is more serious that the method is affected by color of object surface and shooting condition, and application scenario is limited.
In sum, in above-mentioned prior art except there is the shortcoming of system architecture complexity, and aforementioned techniques all requires harshness to measurement environment and illumination condition, has also limited its deep application in complicated lighting condition and robot vision field.
The two-value phase grating of volume coordinate modulation type by Dammann and Gortle in invention in 1971 [formerly technology 1:H.Dammann and K.Gortler, Opt.Comm.1971,3 (3): 312～315].It is as fraunhofer type optical device, and the Fraunhofer diffraction pattern that incident light wave produces through it is the isocandela hot spot of certain dot matrix number, has avoided general amplitude grating because of the uneven distribution of the caused spectrum point light intensity of sinc function intensity envelope completely.Nineteen ninety-five Zhou Changhe has provided Darman raster solution [technology 2:C.H.Zhou formerly, L.R.Liu, the Appl.Opt. of from 2 to 64 dot matrix, 1995,34 (26), 5961～5969], designed afterwards the circular Darman raster matching with most of optical systems, and at length analyzed phase place and made error and sidewall corrosion error to the impact of grating performance [technology 3:C.H.Zhou formerly, J.Jia, L.R.Liu, Opt.Lett., 2003,27 (22): 2174～2176].
Based on boring copper sheet, can obtain three dimensional surface data, its principle is that the dot matrix that utilizes boring copper sheet to produce is irradiated to behind measured object surface, utilizes digital camera take the dot matrix picture of distortion and finally solve object three-dimensional contour outline.But the problem of the method is array hole scale copper is to be difficult to formation rule dot matrix, and array of light spots is of poor quality.If the diameter in hole is larger, the luminous point projecting is too large, and measuring accuracy cannot guarantee, also the collimation of lighting source and even light has been proposed to harsh requirement simultaneously; If the diameter in copper hole is too little, projected light exists diffraction to disturb the shape that cannot guarantee to project dot matrix.Because each luminous point light intensity space distribution is uniformly, antijamming capability is low, measure poor signal to noise, Measurement Resolution is low.Most laser energies are all stopped by copper coin simultaneously, and system capacity utilization factor is very low, and measurement environment requires harsh.
Darman raster is evenly distributed in energy of light source in each order of diffraction time, and the relative tertiary location of each order of diffraction time is strictly fixing and can Exact Solution, by suitable Optical System Design, the accurate size of controlling projection luminous point, thus improve the Measurement Resolution of system.The people such as Zhang Jun have proposed the Fourier's striped profilometry based on Darman raster first, adopt Darman raster to coordinate cylindrical mirror to substitute traditional striped and produce scheme, there is capacity usage ratio high, projected fringe is evenly distributed, the advantages such as the large and clear-cut texture of ratio of compression is bright, efficiently solve traditional Fourier's the Schlieren method capacity usage ratio low, the problems such as striped ratio of compression is little, but its category that still belongs to Fourier's technology of profiling is [referring to technology 4:J.Zhang formerly, C.H.Zhou, X.X.Wang, Three Dimensional Profilometry using Dammann Grating, Appl Opt, 2009, 48 (19), 3709-3715].
Summary of the invention
The present invention seeks to propose a kind of object three-dimensional contour outline measuring set and measuring method based on Darman raster, this device has the advantages such as measuring accuracy is high, apparatus structure is simple, cost is low, dot matrix pattern is adjustable, capacity usage ratio is high, be applicable to complicated lighting environment, the reconstruct of three-D profile data computer is convenient.
Technical solution of the present invention is as follows:
A kind of object three-dimensional contour outline measuring set, be characterized in that this device has laser instrument, collimation lens, Darman raster, projecting lens, left digital camera, right digital camera, transmission line, control line and computing machine form, the annexation of each parts is: described laser emitting laser passes through collimation lens successively, Darman raster, after projecting lens, form rectangular lattice pattern and the measured object surface of throwing light on, the illuminated region of measured object is inputted computing machine and obtained measured object partial 3 d pattern information after view data processing and Point Cloud Processing by transmission line by the image of described left digital camera and right digital camera synchronous acquisition, by mobile measured object or modifier position, realize the measurement of the region-wide three-dimensional appearance information of measured object, obtain object three-dimensional contour outline information.
Described laser instrument refers to Wavelength stabilized single wavelength laser.
Described collimation lens refers to realize the optical lens that laser emitting light is realized contracting bundle collimating effect.
Described Darman raster refers to a branch of incident laser to be beamed into NxM bundle laser dot-matrix and the equally distributed Microstructure Optics element of dot matrix light intensity, and N and M are positive integer.To the diffraction pattern of the described Darman raster of described laser wavelength, be rectangular distribution rule dot matrix, lattice distance can be controlled by design Darman raster.
Described projecting lens refer to by after Damman raster splitting beam dot matrix light High frequency filter and low distortion be projected to the optical lens on measured object surface.
Described left digital camera and right digital camera are placed on respectively projection optical system both sides, form an angle with projection optical axis.Camera is taken rear and picture synchronization collection is sent to computing machine measured object simultaneously respectively.Be sent to the time-series image data of computing machine after computer picture and Point Cloud Processing, the cloud data of the profiling object surface that just can obtain is also realized object three-dimensional contour outline and is rebuild.
Described computing machine has image capture interface, laser control interface and measuring three-dimensional profile reconstruction software.
Utilize above-mentioned object three-dimensional contour outline measuring set to carry out the method for object three-dimensional contour outline measurement, comprise the following steps:
1. adjust projecting lens and measured object in suitable distance, make the rectangular distribution dot matrix light Uniform Illumination producing by described projecting lens after described Damman raster splitting beam in measured object area surfaces to be measured;
2. by described left digital camera, right digital camera, measured object is carried out to Sync image capture, the image of setting t left digital camera of the moment and right digital camera collection is respectively I
land I (t)
r(t), by the image of described computing machine, process the rear cloud data P (t) that obtains the partial contour information that comprises the tested region of object;
3. by changing testee measuring point, constantly measure and obtain the independent point cloud data acquisition { P (t) } of different parts (different measuring is t constantly).After this { P (t) } Point Cloud Processing by described computing machine, obtain the full visual angle cloud data P of object
f, and restore object three-dimensional contour outline information, complete the measurement of object three-dimensional contour outline.
Described image is processed and is comprised that synchronous filtering, characteristic matching and some cloud build three processes.
Described synchronous filtering is responsible for t
1the image I constantly gathering
l(t
1) and I
r(t
1) enter respectively four-way synchronous filtering:
1. first passage carries out high-pass filtering, and obtains the zero order light point diagram I that only contains Darman raster zero order light dot information
0 l(t
1) and I
0 r(t
1);
2. second channel carries out bandpass filtering and weeds out body surface color information and only obtain the Dammam dot chart I containing the distortion of contour of object information
d l(t
1) and I
d r(t
1);
3. third channel gray-scale filtering, carries out gray-scale filtering and obtains rejecting ground unrest but the half-tone information figure I of inclusion surface grey scale change information
g l(t
1) and I
g r(t
1);
4. four-way noise reduction filtering, carries out noise reduction filtering and obtains retaining object texture color information low noise image I
p l(t
1) and I
p r(t
1).
Four Channel Synchronous filtering obtain above-mentioned four groups of filtering datas, and above-mentioned four kinds of filtering technique concrete methods of realizings are referring to [formerly technology 5: Zhang Yujin, Image Engineering (second edition), publishing house of Tsing-Hua University, 2007].
The image pair pixel that described characteristic matching is responsible for left digital camera, right digital camera to gather in the same time carries out Corresponding matching one by one.Concrete steps comprise:
1. due to zero order light point diagram I
0 l(t
1) and I
0 r(t
1) zero level luminous point is separately unique and definite, calculates zero order light point diagram I
0 land I (t1)
0 r(t
1) zero level luminous point barycenter, obtain the luminous point center-of-mass coordinate under the computer picture coordinate system of strict coupling
with
luminous point centroid computing method is referring to [formerly technology 6:S.Ghosal, R.Mehrotra.Orthogonal moment operators for subpixel edge detection, Pattern Recognition, 1993,26 (2): 295-306].
2. calculating chart I
d l(t
1) and I
d r(t
1) in the computer graphic photo coordinate system of all luminous point barycenter under center-of-mass coordinate (μ
l(t
1), v
l(t
1)) and (μ
r(t
1), v
r(t
1)), simultaneously at figure I
g l(t
1) and I
g r(t
1) the middle yardstick invariant features information C that calculates aforementioned each luminous point neighborhood
l(μ
l, v
l) and C
r(μ
r, v
r), the method of calculating yardstick invariant features information refers to formerly formerly technology 7:D.G.Lowe of technology 7[, Distinctive Image Features from Scale-Invariant Keypoints, International Journal of Computer Vision, 200460 (2): 91-11].
3. define I
d land I (t1)
d r(t
1) in figure each luminous point barycenter to zero level luminous point centroid distance separately for point from:
With
At I
p r(t
1) find I in figure
d l(t
1) figure mid point is apart from being D
lluminous point
corresponding matching point
in conjunction with polar curve constraint condition, will be with I
d r(t
1) to take its zero level luminous point barycenter in image be the center of circle, and with radius k
1d
land k
2d
lfor the yardstick invariant features information of all luminous points of endless belt in bound is included into set { G} to be matched.4. set up characteristic information evaluation function F and threshold function table M
f, calculate
c
r(μ
r, v
r) ∈ { G}, the C that evaluation function F value is minimum
r(μ
r, v
r) corresponding luminous point is coupling luminous point
if minimum F value is still greater than M
fbe considered as that it fails to match and abandon coupling.Cycle calculations I
d l(t
1) all luminous points complete all couplings, and finally obtain computer picture planimetric coordinates matching list T (L, R), thus characteristic matching completed.
Described some cloud builds to be responsible for combining with the inside and outside parameter of left digital camera (L), right digital camera (R) in above-mentioned coordinate matching table T (L, R), utilizes trigonometry by t
1computer picture planimetric coordinates information Restore All is constantly t
1moment world coordinate system W (t
1) three-dimensional coordinate information (Xw
t1, Yw
t1, Zw
t1), thereby complete a cloud, build and obtain discrete point cloud P (t
1).Described camera inside and outside parameter acquisition methods is referring to technology 8[Z.Zhang formerly, A Flexible New Technique for Camera Calibration, IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000,22 (11): 1330-1334].
Described some cloud treatment scheme comprises image index and puts cloud process.
Described image index is responsible for moment t
mthe low noise image I obtaining by aforementioned noise reduction filtering
p land I (t)
p r(t) with synchronization t
mcloud data P (t
m) corresponding relation and set up index function In (P, I
p l, I
p r), thereby complete image index.
Described some cloud be responsible for by the local cloud data P (t) at measured object different measuring position all coordinate transform to standard world coordinate system W
sdescend and obtain standard cloud data P
f.Concrete steps comprise: 1. set t
1cloud data P (the t constantly obtaining
1) residing coordinate is standard world coordinate system Ws.T
1next is t constantly for constantly adjacent
2constantly, t
2the cloud data constantly obtaining is P (t
2), to (the I that in aforementioned synchronous filtering, four-way filtering obtains
p l(t
1), I
p l(t
2)) and (I
p r(t
1), I
p r(t
2)) color of image and half-tone information carry out global characteristics and calculate fast, and obtain the characteristic information function S of earlier figures picture
l(t
1), S
l(t
2) and S
r(t
1), S
r(t
2), find S
land S (t1)
l(t
2) same characteristic features and set up correspondence set { S
l(t
1, t
2), in like manner there is { S
r(t
1, t
2).2. utilize Feature Correspondence Algorithm and in conjunction with gradient of disparity constraint and limit constraint, calculate { S
l(t
1, t
2) middle I
p l(t
1) pixel at I
p r(t
1) in match point, in like manner calculate { S
r(t
1, t
2) middle I
p r(t
1) pixel at I
p l(t
1) in match point, by this cross-matched process, finally determine t
1and t
2some strict coincide point constantly.3. three strict coincide points can be determined a cloud P (t
1) and P (t
2) between an euclidean transformation be related to R (t
2, t
1), error causes that { X} has determined some cloud P (t
1) and P (t
2) between a plurality of euclidean transformation relations be R
1(t
2, t
1) to R
n(t
2, t
1), by all R (t
2, t
1) composition euclidean transformation set { R (t
2, t
1).4. according to vertex scheme arrow and the curvature information of a cloud, calculate P (t
1) in each summit p footnote and obtain footnote value FP (p), for P (t
2) in each summit q have footnote value FP (q).For meeting | FP (p)-FP (q) | a pair of point (p, q) of≤ε is to P (t
1) and P (t
2) between euclidean transformation be related to R (t
2, t
1) there is a ticket voting right, all tools are votable forms polling place pair set { (p, q) } to (p, q).5. according to index function In, be that polling place pair set { (p, q) } is selected to throw the euclidean transformation set { R (t that participates in ballot
2, t
1), and utilize the middle all-pair of set { (p, q) } to { R (t
2, t
1) in all R values vote, the R that score is the highest
bjust elect P (t as
1) and (t
2) euclidean transformation relation, and according to R
bcomplete discrete point cloud data P (t
2) to P (t
1) coordinate conversion of some cloud coordinate system Ws, obtain the cloud data P of data dilatation
2(t
1).6. measuring process repeat said process 1.-5., by t in follow-up measurement
3discrete point cloud data P (the t constantly obtaining
3) splice to P
2(t
1) acquisition P
3(t
1), in like manner by P (t
4) splice to P
3(t
1) obtain P
4(t
1) ... etc., last moment t the most at last
nthe discrete point cloud obtaining splices to P
n-1(t
1) in obtain P
n(t
1).Because P
n(t
1) world coordinates be standard world coordinate system W
s, P
n(t
1) be exactly the full visual angle cloud data P of object under Ws coordinate system
f.Can be to cloud data P
fcarry out standard data format output, also can carry out surface fitting to it and restore three profile informations of object.Curved surface fitting method is wherein consulted [technology 9:V.Weissa formerly, L.Andorb, G.Rennera, and T.V á radya, Advanced surface fitting techniques, Computer Aided Geometric Design, 2002,19 (1): 19-42].
Technique effect of the present invention:
The present invention is applied to Darman raster in object three-dimensional contour outline measuring set, and due to Darman raster, to have luminous point light intensity even, the high and large feature of spot definition ratio of compression of capacity usage ratio, thereby possess dot matrix rule and signal to noise ratio (S/N ratio) advantages of higher.The present invention is applied to contour of object by multiple-camera computer vision and measures, and has solved the problems such as existing surface drilling technology control accuracy requirement is high, dot matrix weak effect, measuring accuracy is low, dot matrix light area is excessive, dot matrix light light intensity is inhomogeneous, measuring accuracy low technical, capacity usage ratio is low, measurement environment requirement is harsh.The present invention distributes by the three-dimensional appearance of binocular camera images match splicing reconstruction object.Compare existing three-dimension measuring system, the present invention has the advantages such as application scenario is wide, measuring accuracy is high, weight of equipment is light, cost is low, energy consumption is little, simple in structure, enforcement is convenient.
Accompanying drawing explanation
Fig. 1 is object three-dimensional contour outline measuring set block diagram of the present invention
Fig. 2 is the light-dividing principle figure of Darman raster
Fig. 3 is dot matrix light schematic diagram of the present invention
Fig. 4 is Darman raster dot matrix illumination plaster statue schematic diagram
Fig. 5 is Computer Image Processing process flow diagram
Fig. 6 is computing machine point cloud processing flow chart
Embodiment
Below by embodiment and accompanying drawing thereof, the present invention is further described.
Consult Fig. 2, Fig. 2 is the schematic diagram of the Darman raster light splitting that adopts of the present invention, and directional light focuses on focal plane 10 through lens 4 after inciding Darman raster 3.
Consult Fig. 3, Fig. 3 is dot matrix light schematic diagram of the present invention, and Damman raster splitting beam is NxM rectangular lattice light, and this dot matrix light intensity is even and lattice distance is consistent, and wherein N and M are positive integer.
Consult Fig. 1, Fig. 1 is object three-dimensional contour outline measuring set block diagram of the present invention.As seen from the figure, the present invention is based on the object three-dimensional contour outline measuring set of Darman raster, by laser instrument 1, collimation lens 2, Darman raster 3, projecting lens 4, left digital camera L, right digital camera R, transmission line 6, control line 7 and computing machine 8 form, the annexation of each parts is: the light beam that described laser instrument 1 sends passes through collimation lens 2 successively, Darman raster 3, after projecting lens 4, form two-dimensional rectangle projection dot matrix illumination measured object surface 5, rectangular lattice pattern by described left digital camera L and the surperficial face shape modulation of right digital camera R synchronous acquisition measured target 5, through the described computing machine 8 of transmission line 6 input, this computing machine 8 is by control line 7 connecting lasers 1, described computing machine 8 has image capture interface, laser control interface and measuring three-dimensional profile reconstruction software.
Utilizing object three-dimensional contour outline measuring set of the present invention to implement people's face plaster statue measuring three-dimensional profile is example, and the measuring process of measuring method of the present invention is described as follows:
The first step:
1. adjust the distance between projecting lens 4 and measured object-people face plaster statue 5 surfaces, make two-dimensional rectangle distribution dot matrix light Uniform Illumination that laser instrument 1 emergent light produces by described projecting lens (4) after described Darman raster 3 beam splitting on people's face plaster statue 5, computing machine takes orders and first by control line 7 control laser instruments 1, sends laser, single wavelength laser that laser instrument 1 sends incides the surface of Darman raster 3 after collimation lens 2 with collimated light, collimated light beam, by diffraction beam splitting after Darman raster, forms the dot matrix beam array of two-dimensional rectangle.This dot matrix beam array is projected on plaster statue surface after by projecting lens 4, and rectangular distribution dot matrix light beam deforms after by people's face plaster statue surface modulation originally, becomes irregular dot matrix, and plaster statue Darman raster dot matrix illuminating effect as shown in Figure 4;
2. described left digital camera L and right digital camera R carries out Sync image capture to people's face plaster statue, and the right moment is referred to as initial measurement t constantly to start to take this image
1, set t
1the image that left digital camera (L), right digital camera (R) gather is constantly respectively left image I
l(t
1) and right image I
r(t
1), the image of the measuring three-dimensional profile software by described computing machine (8) obtains the cloud data P (t) of the partial contour information that comprises the tested region of object after processing;
3. by changing the measuring point of testee, repeating step 2., obtain different parts, be the different measuring independently cloud data set { P (t) } of t constantly, after the Point Cloud Processing of this cloud data set { P (t) } by the measuring three-dimensional profile software of described computing machine (8), obtain the full visual angle cloud data P of object
f, and restore object three-dimensional contour outline information, complete the measurement of object three-dimensional contour outline.
Second step: process the discrete point cloud data of this left and right image to the measured position of this initial time of rear acquisition people's face plaster statue by the image treatment scheme of computing machine.
Described image treatment scheme is:
1. to described left image I
l(t
1) and right image I
r(t
1) carrying out four-way synchronous filtering, the zero order light point diagram I that only contains Darman raster zero order light dot information is obtained in first passage high-pass filtering
0 l(t
1) and I
0 r(t
1), second channel bandpass filtering weeds out body surface acquiring color information and only contains the Dammam dot chart I of the distortion of contour of object information
d l(t
1) and I
d r(t
1), third channel gray-scale filtering obtains rejects ground unrest but the half-tone information figure I of inclusion surface grey scale change information
g l(t
1) and I
g r(t
1); Four-way carries out noise reduction filtering and obtains retaining object texture color information low noise image I
p l(t
1) and I
p r(t
1).Described left image I
l(t
1) and right image I
r(t
1) to by obtaining four groups of filtering datas after four-way synchronous filtering: zero order light point diagram I
0 l(t
1) and I
0 r(t
1), Dammam dot chart I
d l(t
1) and I
d r(t
1), half-tone information figure I
g l(t
1) and I
g r(t
1) and low noise image I
p l(t
1) and I
p r(t
1);
2. according to first three groups filtering data, calculate zero order light point diagram I
0 l(t
1) and I
0 r(t
1) zero level luminous point barycenter, obtain coupling computer picture coordinate system coordinate
with
the synchronous I that calculates
d l(t
1) and I
d r(t
1) in the computer graphic photo coordinate system coordinate (μ of all luminous point barycenter
l(t
1), v
l(t
1)) and (μ
r(t
1), v
r(t
1)), then at I
g l(t
1) and I
g r(t
1) the middle yardstick invariant features information C that calculates and obtain aforementioned luminous point neighborhood
l(μ
l, v
l) and C
r(μ
r, v
r);
3. define I
d l(t
1) and I
d r(t
1) in figure each luminous point barycenter to zero level luminous point centroid distance separately for point from:
With
At I
p r(t
1) find I in figure
d l(t
1) figure mid point is apart from being D
lluminous point
corresponding matching point
in conjunction with polar curve constraint condition, will be with I
d r(t
1) to take its zero level luminous point barycenter in image be the center of circle, and with radius k
1d
land k
2d
lfor the yardstick invariant features information of all luminous points of endless belt in bound is included into set G to be matched;
4. set up characteristic information evaluation function F, calculate
c
r(μ
r, v
r) ∈ G, the C that evaluation function F value is minimum
r(μ
r, v
r) corresponding luminous point is coupling luminous point
if minimum F value is still greater than evaluation function threshold value M
fbe considered as that it fails to match and abandon this time coupling.Cycle calculations I
d l(t
1) all luminous points complete all couplings, and finally obtain computer picture planimetric coordinates matching list T (L, R), thus characteristic matching completed;
5. the inside and outside parameter with camera L and R in conjunction with coordinate matching table T (L, R), utilizes trigonometry by t
1computer picture planimetric coordinates information Restore All is constantly t
1moment world coordinate system W (t
1) three-dimensional coordinate information (Xw
t1, Yw
t1, Zw
t1), thereby complete a cloud, build and obtain discrete point cloud P (t
1).
1. above-mentioned-5. Computer Image Processing flow process refers to Fig. 5.
The 3rd step, uniform speed slow moves apparatus of the present invention, and the different parts by rectangular lattice light beam projecting to plaster statue obtains t by described left digital camera L and right digital camera R
2the left image I of plaster statue constantly
l(t
2) and right image I
r(t
2) image pair, repeat second step and obtain t
2discrete point cloud data P (the t at the moment new tested position of plaster statue
2), and put cloud treatment scheme by P (t by computing machine
2) be spliced to P (t
1) in.
Described computing machine point cloud treatment scheme is:
1. set up moment t in filtering data
mlow noise image I
p land I (t)
p r(t) with synchronization t
mcloud data P (t
m) corresponding relation and set up index function In (P, I
p l, I
p r);
2. set initial time t
1cloud data P (the t obtaining
1) residing coordinate is standard world coordinate system Ws;
3. according to adjacent moment t
1and t
2(I
p l(t
1), I
p l(t
2)) and (I
p r(t
1), I
p r(t
2)) color that comprises of image and half-tone information carry out global characteristics and calculate fast and obtain adjacent moment t
1and t
2the characteristic information function S of low noise image
l(t
1), S
l(t
2) and S
r(t
1), S
r(t
2);
4. find S
land S (t1)
l(t
2) same characteristic features and set up correspondence set { S
l(t
1, t
2), in like manner there is { S
r(t
1, t
2), in like manner there is { S
r(t
1, t
2), utilize Feature Correspondence Algorithm and in conjunction with gradient of disparity constraint and limit constraint, calculate { S
l(t
1, t
2) middle I
p l(t
1) pixel at I
p r(t
1) in match point, in like manner calculate { S
r(t
1, t
2) middle I
p r(t
1) pixel at I
p l(t
1) in match point, by this cross-matched process, finally determine t
1and t
2some strict coincide point { X} constantly.Three strict coincide points can be determined a cloud P (t
1) and P (t
2) between an euclidean transformation be related to R (t
2, t
1), error causes that { X} has determined some cloud P (t
1) and P (t
2) between a plurality of euclidean transformation relations be R
1(t
2, t
1) to R
n(t
2, t
1), by all R (t
2, t
1) composition euclidean transformation set { R (t
2, t
1);
5. according to vertex scheme arrow and the curvature information of a cloud, calculate P (t
1) in each summit p footnote and obtain footnote value FP (p), for P (t
2) in each summit q have footnote value FP (q).For meeting | FP (p)-FP (q) | a pair of point (p, q) of≤ε is to P (t
1) and P (t
2) between euclidean transformation be related to R (t
2, t
1) there is a ticket voting right, all tools are votable forms polling place pair set { (p, q) } to (p, q);
6. according to index function In, be that polling place pair set { (p, q) } is selected to throw the euclidean transformation set { R (t that participates in ballot
2, t
1), and utilize the middle all-pair of set { (p, q) } to { R (t
2, t
1) in all R values vote, the R that score is the highest
bjust elect P (t as
1) and (t
2) euclidean transformation relation, and according to R
bcomplete discrete point cloud data P (t
2) to P (t
1) coordinate conversion of some cloud coordinate system Ws, obtain the cloud data P of data dilatation
2(t
1).
1. above-mentioned-6. computing machine point cloud treatment scheme refers to Fig. 6.
The 4th step: constantly change the measuring point of plaster statue, repeat abovementioned steps in measuring process, by t in follow-up measurement
3discrete point cloud data P (the t constantly obtaining
3) splice to P
2(t
1) acquisition P
3(t
1), in like manner by P (t
4) splice to P
3(t
1) obtain P
4(t
1) ... etc., last moment t the most at last
nthe discrete point cloud obtaining splices to P
n-1(t
1) in obtain P
n(t
1).Because P
n(t
1) world coordinates be standard world coordinate system W
s, P
n(t
1) be exactly the full visual angle cloud data P of plaster statue under standard world coordinate system Ws
f.Can be to cloud data P
fcarry out standard data format output, also can carry out surface fitting to it and restore three profile informations of plaster statue.
Object three-dimensional contour outline based on Darman raster dot matrix light of the present invention is measured structure, there is measurement strong adaptability, measuring accuracy is high, measuring speed is fast, the advantages such as apparatus structure is simple, cost is low, lightweight, application scenario is wide, be convenient to the simplification of computerized algorithm implementation procedure and reconstructing 3 D contour process, can realize automatic measurement, there is important practical value and prospect.
Claims (7)
1. a measuring method of utilizing object three-dimensional contour outline measuring set to carry out object three-dimensional contour outline, this object three-dimensional contour outline measuring set based on Darman raster is by laser instrument (1), collimation lens (2), Darman raster (3), projecting lens (4), left digital camera (L), right digital camera (R), transmission line (6), control line (7) and computing machine (8) form, the annexation of each parts is: the light beam that described laser instrument (1) sends passes through collimation lens (2) successively, Darman raster (3), after projecting lens (4), form two-dimensional rectangle projection dot matrix illumination measured object surface (5), rectangular lattice pattern by described left digital camera (L) and (5) 3 d shape modulation of right digital camera (R) synchronous acquisition measured target surface, through transmission line (6), input described computing machine (8) and carry out data processing, this computing machine (8) is by control line (7) connecting laser (1), described computing machine (8) has image capture interface, laser control interface and measuring three-dimensional profile reconstruction software, it is characterized in that the method comprises the following steps:
1. adjust the distance between projecting lens (4) and measured object surface (5), make after Darman raster (3) beam splitting, by described projecting lens (4), to form on two-dimensional rectangle projection dot matrix beam lighting measured object surperficial (5) described in laser instrument (1) emergent light;
2. by described left digital camera (L), right digital camera (R), measured object is carried out to Sync image capture, the image of setting the t moment left digital camera (L) and right digital camera (R) collection is respectively I
land I (t)
r(t) the described computing machine (8) of input, the cloud data P (t) of the partial contour information that acquisition comprises the tested region of object after the image of the measuring three-dimensional profile software of this computing machine (8) is processed;
3. by changing the measuring point of testee, repeating step 2., obtain different parts, it is the different measuring independently cloud data set { P (t) } of t constantly, after the Point Cloud Processing of this cloud data set { P (t) } by the measuring three-dimensional profile software of described computing machine (8), obtain the full visual angle cloud data P of object
f, and restore object three-dimensional contour outline information, complete the measurement of object three-dimensional contour outline.
2. the measuring method of object three-dimensional contour outline according to claim 1, is characterized in that the flow process of described Computer Image Processing comprises: image synchronization filtering, Image Feature Matching and picture point cloud build.
3. the measuring method of object three-dimensional contour outline according to claim 2, is characterized in that described image synchronization filtering comprises that four-way image filtering synchronously carries out:
1. first passage high-pass filtering, obtains the zero order light point diagram I that only contains Darman raster zero order light dot information
0 l(t
1) and I
0 r(t
1);
2. second channel bandpass filtering, weeds out body surface acquiring color information and only contains the Dammam dot chart I of the distortion of contour of object information
d l(t
1) and I
d r(t
1);
3. third channel gray-scale filtering, obtains and rejects ground unrest but the half-tone information figure I of inclusion surface grey scale change information
g l(t
1) and I
g r(t
1);
4. four-way noise reduction filtering, obtains and retains object texture color information low noise image I
p l(t
1) and I
p r(t
1);
Described first passage filtering data, second channel filtering data and third channel filtering data are for characteristic matching link in image treatment scheme, and four-way filtering data is for some cloud treatment scheme.
4. the measuring method of object three-dimensional contour outline according to claim 3, is characterized in that described Image Feature Matching comprises following steps:
1. calculate zero order light point diagram I
0 l(t
1) and I
0 r(t
1) zero level luminous point barycenter, obtain the zero order light point diagram I of computer picture coordinate system
0 l(t
1) and I
0 r(t
1) luminous point center-of-mass coordinate
with
2. calculate object texture color information low noise image I
p l(t
1) and I
p r(t
1) in the computer graphic photo coordinate system coordinate of all luminous point barycenter
with
simultaneously at half-tone information figure I
g l(t
1) and I
g r(t
1) the middle yardstick invariant features information C that calculates and obtain described luminous point neighborhood
l(μ
l, ν
l) and C
r(μ
r, ν
r);
3. define low noise image I
p l(t
1) and I
p r(t
1) in figure each luminous point barycenter to zero level luminous point centroid distance separately for point from:
With
At I
p r(t
1) find I in figure
d l(t
1) figure mid point is apart from being D
lluminous point
corresponding matching point
in conjunction with polar curve constraint condition, will be with I
d r(t
1) to take its zero level luminous point barycenter in image be the center of circle, and with radius k
1d
land k
2d
lfor the yardstick invariant features information of all luminous points of endless belt in bound is included into set G to be matched;
4. set up characteristic information evaluation function F and evaluation function threshold value M
f, calculate
evaluation function F value is minimum
corresponding luminous point is coupling luminous point
if minimum F value is still greater than M
fbe considered as that it fails to match and abandon this time coupling; Cycle calculations I
d l(t
1) all luminous points complete all couplings, and finally obtain computer picture planimetric coordinates matching list T (L, R), thus characteristic matching completed.
5. the measuring method of object three-dimensional contour outline according to claim 4, it is characterized in that it is by by coordinate matching table T (L that described some cloud builds, R) combine with the inside and outside parameter of described left digital camera (L), right digital camera (R), utilize the three-dimensional coordinate information (Xw that trigonometry is t moment world coordinate system W (t) by t computer picture planimetric coordinates information Restore All constantly
t, Y w
t, Z w
t), thereby complete a cloud, build and obtain discrete point cloud data P (t).
6. the measuring method of object three-dimensional contour outline according to claim 1, is characterized in that described Point Cloud Processing comprises image index and some cloud process.
7. the measuring method of object three-dimensional contour outline according to claim 6, is characterized in that described image index is to set up t constantly
mlow noise image I
p l(t
m) and I
p r(t
m) and synchronization t
mcloud data P (t
m) corresponding relation, and set up index function In (P, I
p l, I
p r).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110456929.1A CN102564347B (en) | 2011-12-30 | 2011-12-30 | Object three-dimensional outline measuring device and method based on Dammann grating |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110456929.1A CN102564347B (en) | 2011-12-30 | 2011-12-30 | Object three-dimensional outline measuring device and method based on Dammann grating |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102564347A CN102564347A (en) | 2012-07-11 |
| CN102564347B true CN102564347B (en) | 2014-11-26 |
Family
ID=46410417
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110456929.1A Active CN102564347B (en) | 2011-12-30 | 2011-12-30 | Object three-dimensional outline measuring device and method based on Dammann grating |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102564347B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109341581A (en) * | 2018-10-31 | 2019-02-15 | 暨南大学 | The three-dimensional measurement mould group and working method of lateral parity combination Darman raster |
Families Citing this family (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103542803B (en) * | 2013-09-30 | 2016-05-11 | 中国科学院上海光学精密机械研究所 | Based on the synchronous phase shift interference device of Darman raster |
| CN104251670A (en) * | 2014-08-28 | 2014-12-31 | 合肥斯科尔智能科技有限公司 | Multi-view raster 3D (three-dimensional) scanner |
| CN104296681B (en) * | 2014-10-16 | 2016-12-07 | 浙江大学 | Dimensional topography method for sensing based on laser dot-matrix mark |
| CN104296691B (en) * | 2014-10-28 | 2017-04-19 | 电子科技大学 | Aircraft engine tail profile measurement device and method based on machine vision |
| CN105698707B (en) * | 2014-11-28 | 2018-08-31 | 上海航天动力科技工程有限公司 | A kind of grating three-dimensional shape measuring apparatus |
| CN104634276B (en) * | 2015-02-12 | 2018-08-07 | 上海图漾信息科技有限公司 | Three-dimension measuring system, capture apparatus and method, depth computing method and equipment |
| CN104776815B (en) * | 2015-03-23 | 2018-04-17 | 中国科学院上海光学精密机械研究所 | A kind of color three dimension contour outline measuring set and method based on Darman raster |
| CN106152946B (en) * | 2015-03-31 | 2019-08-02 | 酷派软件技术（深圳）有限公司 | A kind of method and terminal measuring object length |
| CN104991255A (en) * | 2015-06-26 | 2015-10-21 | 郎一宁 | Visual principle-based multipoint laser range radar |
| CN105674912B (en) * | 2016-01-26 | 2018-02-13 | 中国科学院上海光学精密机械研究所 | With reference to more finishing tool color three dimensional measurement apparatus and method of Darman raster |
| FR3055698B1 (en) * | 2016-09-08 | 2018-08-17 | Safran Aircraft Engines | METHOD FOR CONTROLLING THE CONFORMITY OF THE PROFILE OF A CURVED SURFACE OF AN ELEMENT OF A TURBOMACHINE |
| CN108267097B (en) * | 2017-07-17 | 2020-07-10 | 先临三维科技股份有限公司 | Three-dimensional reconstruction method and device based on binocular three-dimensional scanning system |
| CN111758013A (en) * | 2017-10-27 | 2020-10-09 | 精密自动化股份有限公司 | System and method for collecting data and information of surface properties of an object |
| US10871262B2 (en) * | 2017-11-02 | 2020-12-22 | Spring Rainbow Optics Co., Ltd | Structured light illumination module |
| CN108303038B (en) * | 2017-12-21 | 2020-02-21 | 天津大学 | Reflection type surface shape measuring method and device based on two-dimensional optical dot matrix |
| CN108492358A (en) * | 2018-02-14 | 2018-09-04 | 天目爱视（北京）科技有限公司 | A kind of 3D 4 D datas acquisition method and device based on grating |
| CN108564017A (en) * | 2018-04-04 | 2018-09-21 | 北京天目智联科技有限公司 | A kind of biological characteristic 3D 4 D datas recognition methods and system based on grating camera |
| CN109388893B (en) * | 2018-10-18 | 2022-05-03 | 哈尔滨工业大学 | Noise optimization method for three-way and four-way pipelines |
| EP3667624A1 (en) * | 2018-12-14 | 2020-06-17 | Siemens Healthcare GmbH | Method of determining an illumination effect of a volumetric dataset |
| CN109739027B (en) * | 2019-01-16 | 2021-07-27 | 北京华捷艾米科技有限公司 | Light spot array projection module and depth camera |
| CN112102496B (en) * | 2020-09-27 | 2024-03-26 | 安徽省农业科学院畜牧兽医研究所 | Cattle physique measurement method, model training method and system |
| CN113218336B (en) * | 2021-03-31 | 2023-01-31 | 博众精工科技股份有限公司 | Grating plate, projection device for three-dimensional reconstruction of structured light and measurement device |
| CN114841965B (en) * | 2022-04-30 | 2023-08-01 | 中建三局第一建设工程有限责任公司 | Steel structure deformation detection method and device, computer equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101451826A (en) * | 2008-12-17 | 2009-06-10 | 中国科学院上海光学精密机械研究所 | Object three-dimensional profile measuring device and method |
| CN102155925A (en) * | 2011-01-07 | 2011-08-17 | 中国科学院上海光学精密机械研究所 | Three-dimensional surface topography measuring device based on one-dimensional Dammann grating |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009031150A (en) * | 2007-07-27 | 2009-02-12 | Omron Corp | Three-dimensional shape measuring device, three-dimensional shape measurement method, three-dimensional shape measurement program, and record medium |
-
2011
- 2011-12-30 CN CN201110456929.1A patent/CN102564347B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101451826A (en) * | 2008-12-17 | 2009-06-10 | 中国科学院上海光学精密机械研究所 | Object three-dimensional profile measuring device and method |
| CN102155925A (en) * | 2011-01-07 | 2011-08-17 | 中国科学院上海光学精密机械研究所 | Three-dimensional surface topography measuring device based on one-dimensional Dammann grating |
Non-Patent Citations (1)
| Title |
|---|
| JP特开2009-31150A 2009.02.12 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109341581A (en) * | 2018-10-31 | 2019-02-15 | 暨南大学 | The three-dimensional measurement mould group and working method of lateral parity combination Darman raster |
| CN109341581B (en) * | 2018-10-31 | 2020-12-11 | 暨南大学 | Three-dimensional measurement module of transverse odd-even combined Dammann grating and working method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102564347A (en) | 2012-07-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102564347B (en) | Object three-dimensional outline measuring device and method based on Dammann grating | |
| CN101451826B (en) | Object three-dimensional profile measuring device and method | |
| CN104634276B (en) | Three-dimension measuring system, capture apparatus and method, depth computing method and equipment | |
| US10152800B2 (en) | Stereoscopic vision three dimensional measurement method and system for calculating laser speckle as texture | |
| CN102410811B (en) | Method and system for measuring parameters of bent pipe | |
| CN101813461B (en) | Absolute phase measurement method based on composite color fringe projection | |
| CN100520287C (en) | Method and apparatus for measuring shape of an object | |
| CN102884397B (en) | Structured-light measuring method and system | |
| CN103196370B (en) | Measuring method and measuring device of conduit connector space pose parameters | |
| CN201344792Y (en) | Object three-dimensional profile measuring device | |
| CN103776390A (en) | Three-dimensional natural texture data scanning machine and multi-view-field data splicing method | |
| US10648789B2 (en) | Method for monitoring linear dimensions of three-dimensional objects | |
| CN105674912B (en) | With reference to more finishing tool color three dimensional measurement apparatus and method of Darman raster | |
| WO2006013635A1 (en) | Three-dimensional shape measuring method and apparatus for the same | |
| CN106155299B (en) | A kind of pair of smart machine carries out the method and device of gesture control | |
| CN101701822B (en) | Star tracking method of star sensor based on correlation of optical joint and transformation | |
| CN105574812B (en) | Multi-angle three-dimensional data method for registering and device | |
| CN102183214A (en) | Method for optically detecting large-aperture aspherical mirror structure | |
| CN105241397A (en) | Real-time measuring splicing method and device based on structured light | |
| CN104634277A (en) | Photographing device, photographing method, three-dimensional measuring system, depth calculation method and depth calculation device | |
| EP1459035B1 (en) | Method for determining corresponding points in stereoscopic three-dimensional measurements | |
| CN101482398B (en) | Fast three-dimensional appearance measuring method and device | |
| CN113551611B (en) | Stereo vision measuring method, system, equipment and storage medium for large-size moving object | |
| CN104296681B (en) | Dimensional topography method for sensing based on laser dot-matrix mark | |
| CN103884294A (en) | Wide view field type infrared light three-dimensional morphology measurement method and device thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |