CN1266452C - Composite coding multiresolution three-dimensional digital imaging method - Google Patents

Abstract

The invention discloses a method for realizing multi-resolution three-dimensional digital imaging by using composite coding, that is, a three-dimensional digital technology based on the combination of dot matrix projection and fringe projection. The method adopts an imaging device composed of a digital projection lighting transmitter, an image sensing receiver and an image processor. The exit pupil of the transmitter and the entrance pupil of the image sensing receiver are used to form a triangulation system; first, the 3D digital imaging of dot matrix projection has no phase blur characteristics, and the lower resolution 3D depth image of the object is obtained by dot matrix encoding. Then project the fringes to the surface of the object, use the characteristics of high spatial phase modulation accuracy, and combine the obtained low-resolution depth image of the object to perform phase unwrapping, and further obtain a finer multi-resolution three-dimensional digital image of the measured object. The invention combines the advantages of three-dimensional digital imaging methods with two different coding modes of dot matrix projection and fringe projection, has high precision of three-dimensional digital imaging, and has strong universality for curved surfaces with complex topologies.

Description

Multi-resolution 3D Digital Imaging Method with Composite Encoding

technical field

The invention relates to a compound coding multi-resolution three-dimensional digital imaging method, which belongs to the three-dimensional digital imaging technology.

Background technique

In the active 3D sensing technology based on triangulation, based on the traditional triangulation, the spatial modulation of the structured illumination beam by the surface shape of the 3D object changes the angle of the imaging beam, that is, changes the imaging spot on the detector. For the position on the array, the distance is calculated by determining the position of the imaging light spot and the geometric parameters of the optical path of the system. Existing technologies include methods using single-beam point-structured illumination and line-structured illumination using sheet-like beams, and phase measurement profilometry, including phase shift profilometry, Fourier transform profilometry, and spatial phase detection profilometry. The phase measurement profilometry is finally attributed to the triangulation method, but in different measurement techniques, different methods are used to extract the geometric parameters required in the triangulation calculation from the observed light field.

The method based on spatial lattice projection is to project a two-dimensional spatial lattice onto the measured object to form surface structured light illumination, and a complete three-dimensional digital image of the measured object can be obtained by one imaging, thus overcoming the existing triangular The shortcomings of the single-beam point-structured illumination and the line-structured illumination of the sheet-like beam in the active three-dimensional sensing technology of the method must be scanned point by point or line by line to obtain a complete three-dimensional digital image, which improves the sampling efficiency. Compared with phase measurement profilometry, the dot matrix encoding method calculates the depth of the object directly by determining the position of the imaging dot matrix, and there is no problem of phase ambiguity and error propagation. However, the accuracy of this sensing technology is directly affected by the measurement accuracy of the imaging light spot position. The sensitivity of the detector and the geometric distortion of the imaging system will affect its measurement accuracy. Therefore, its measurement accuracy is not as good as that of phase measurement profilometry.

Phase measurement profilometry adopts fringe image encoding, and the depth information of the object is encoded in the carrier fringe. By calculating the folding phase and phase unwrapping, combined with the structural parameters of the optical system, the depth image of the object is obtained indirectly. Compared with direct geometric quantity measurement, phase measurement profilometry has higher accuracy, up to one hundredth of a fringe period; in addition, the phase measurement method is not sensitive to changes in background, contrast and noise. However, when there is a discontinuous area or a large change in the surface gradient of the measured object, or there is an information blind area, the phase unwrapping of phase measurement profilometry becomes a very difficult problem, such as phase ambiguity and error propagation, resulting in incomplete 3D data. Complete 3D data cannot be obtained.

There are three comparable technical literatures:

(1) Invention patent: ZL 02131096.3.

(2) Richard McBain, "high speed laser triangulation measurements of shape and thickness", US Patent 6,466,305.

(3) ul R, Yoder JR., "topography measuring apparatus", US Patent 4,902,123.

Contents of the invention

The object of the present invention is to provide a method for realizing multi-resolution three-dimensional digital imaging by using composite coding, which improves the confidence of measurement results and high measurement accuracy.

The present invention is realized through the following technical scheme, adopting an imaging device including a digital projection lighting transmitter, an image sensor receiver and an image processor, and utilizing dot matrix coding and stripe coding to realize a multi-resolution three-dimensional digital imaging method, which Characterized by the following processes:

1. The digital projection lighting transmitter projects two-dimensional dot matrix graphic lighting objects with different densities, and deflects the image sensing receiver so that its optical axis is on the same line as the center of the projected dot matrix lighting field. The exit pupil of the digital projection lighting transmitter, the entrance pupil of the image sensing receiver and the center of the lighting field form a triangle, and the connection line between the exit pupil of the digital projection lighting transmitter and the entrance pupil of the image sensing receiver is Baseline, which forms several triangles with the centers of all lattices projected on the object, forming a triangulation system; according to the principle of affine transformation, the coded lattices projected on the reference plane and the surface of the measured object respectively, after coordinate rotation, Translating and perspective projection transformation, respectively obtain the analytic formula of the spatial coordinates of the coding lattice on the surface of the reference plane and the measured object on the plane of the image receiving sensor, and further obtain the correspondence between the reference plane and the two lattices on the surface of the measured object The relationship between the position difference between the points in the x direction of the imaging plane and the depth value of the measured object at the corresponding point; thus the digital projection lighting transmitter projects a two-dimensional spatial lattice structured light illumination, and the image receiving sensor obtains a reference plane The dot matrix image and the dot matrix image projected on the surface of the measured object, according to the known geometric parameters of the optical path of the system, the image processor calculates the position difference between the corresponding points of the dot matrix image in the x direction, and then according to the position difference and the measured object The relationship between the depth values of the measured objects at corresponding points is used to calculate the depth image of the object.

2. The digital projection lighting transmitter projects striped structured light to illuminate the surface of the object to be measured. The projected striped structured light is a single frequency strip or a sequence of time-series frequency-variable strips. For single-frequency fringe structured light, the phase mapping method is used, combined with the low-resolution depth image of the object already obtained by lattice projection, to perform phase unwrapping, and further obtain a finer multi-resolution three-dimensional digital image. For the time series variable frequency fringe sequence structured light projection, the coded fringe intensity distribution map sequence is obtained respectively, and the phase map of each frequency fringe intensity distribution map is calculated by using the "phase shift algorithm", and the 2π uncertainty is eliminated in each phase map, so that Obtain the depth image with progressive resolution for the scene; at this time, the function of lattice projection is to use the object depth value obtained by it to determine the boundary condition of the frequency conversion fringe sequence, effectively saving the two-dimensional encoded fringe image required in the phase unwrapping process , which increases the real-time performance of the phase unwrapping algorithm; and further obtains multi-resolution three-dimensional digital images of higher precision and finer objects.

Lattice coding is a 3D digitization technology based on affine transformation. Stripe coding is a 3D digitization technology of phase mapping. The combination of the two is a composite coding technology. The depth of the object is determined on the "thicker" outline, and the absolute distribution of the phase in the phase map can be determined thereby to solve the problem of phase ambiguity; and, when there is an illumination blind area and an imaging blind area, since the lattice projection is based on the reference plane and The position difference of the "dot matrix pair" formed by the dot matrix image on the surface of the object is used to calculate the depth value of the object. The "dot matrix pair" is independent of each other. The depth value at the discrete point in the information blind area, but it will not affect the surrounding area, will not cause error propagation, and can still get an approximate three-dimensional depth image; and the phase profilometry based on fringe projection, not only these blind areas cannot be obtained Correct phase value, and cause error propagation, seriously affect the phase unwrapping of adjacent areas, and cannot obtain the three-dimensional depth image of the measured object.

The technical scheme based on dot matrix projection three-dimensional digital imaging will be further described in detail below in conjunction with the accompanying drawings:

Dot matrix projection 3D digital imaging technology is a 3D digital technology established on the basis of affine transformation. Fig. 1 is a schematic diagram of the structure of dot matrix imaging projected on a reference plane. Point P is the exit pupil of the digital projection lighting emitter, Po is the projection optical axis, and the intersection of projection rays PP ₁ , PP ₂ and PP ₃ with the reference plane R determines the position of the projected two-dimensional space lattice on the reference plane. Point I is the entrance pupil of the image sensing receiver, Io is the imaging optical axis, and the angle between it and the projection optical axis is α, Ip is the symmetric plane of the image sensing receiver plane relative to point I, according to the principle of perspective projection , the imaging position of the lattice on the image sensing receiver can be obtained by calculating the intersection of the imaging ray and the plane _Ip .

In the dot matrix projection imaging system, the imaging process of the dot matrix projected on the reference plane can be divided into three steps:

Firstly, the object coordinate system xyz is positively rotated around the y axis by an angle of α, transformed into a coordinate system x′y′z′, so that the z′o′ axis of the new coordinate system coincides with the imaging optical axis Io;

Then, the coordinate system x′y′z′ is negatively translated by L along the z′ axis, so that the entrance pupil I of the image sensing receiver is the coordinate origin, and the transformed coordinate system is x _p y _p z _p ;

Finally, do a perspective projection in the coordinate system x _p y _p z _p , the perspective projection plane I _p is perpendicular to the z _p axis, and at the position F (focal distance) away from the origin, it is relative to the image sensing receiver plane to x _p o _p y _p symmetry.

Then, after the above steps 1 and 2, the coordinates of the reference plane R (z=0) in the x _p y _p z _p coordinate system can be expressed as:

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; x</span>}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}\\ {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}\\ {\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; L</span>}{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}\mathrm{<span\; class="notranslate">\; \&Delta;x</span>}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Where α is the angle between the projection optical axis Po and the imaging optical axis Io; L is the distance between the entrance pupil I of the image sensing receiver and the center o of the illumination field; Δx is the point projected on the reference plane R The spacing of the array along the x direction; k _x is an integer, which is the coefficient multiplied by Δx, and is used to represent the x coordinate values of different points in the object coordinate system.

Then do perspective projection under the coordinate system x _p y _p z _p to determine the position of the lattice on the plane I _p :

$\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}}{{\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; p</span>}\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; f</span>}}<span\; class="notranslate">\; \&DoubleRightArrow;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; p</span>}\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; x</span>}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}{\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; x</span>}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

where F is the focal length of the imaging system.

For a three-dimensional measured object, as shown in FIG. 2 , it is a structural schematic diagram of a lattice imaging projected on a three-dimensional object on an arbitrary curved surface. The same projection rays PP ₁ , PP ₂ and PP ₃ as in Figure 1 intersect with the surface of the measured object and points B ₁ , B ₂ and B ₃ , these intersection points determine the projected two-dimensional space lattice on the object surface S spatial location. The dot matrix projected on the surface S of the object also goes through the above three steps and is imaged on the image sensing receiver. At this time, the coordinates of the lattice on S in the x _p y _p z _p coordinate system can be obtained:

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; x</span>}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; z</span>}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}\\ {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}\\ {\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; x</span>}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; z</span>}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

Then do perspective projection under the coordinate system x _p y _p z _p , the position of the point matrix on the surface of the object on the perspective projection plane I _p can be obtained:

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}}{{\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}}{\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; x</span>}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; z</span>}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}{\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; x</span>}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; z</span>}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

According to formulas (2) and (4), the position difference Δx _p in the x direction between the “lattice pairs” formed by corresponding points in the two lattices can be obtained:

$\mathrm{<span\; class="notranslate">\; \&Delta;</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; p</span>}\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

The calculation expression of Δx _p includes the relationship between it and the depth value of the object at this point, so as to obtain the calculation expression of the depth value of the object. When the parameters of the optical system are known, the displacement in the x direction between all "lattice pairs" can be calculated by combining the dot matrix image of the reference plane and the dot matrix image on the surface of the measured object, and the calculated object surface The depth values of all discrete points on the object can be obtained to obtain the depth image of the object.

The composite coding method based on the lattice-based affine transformation and the fringe-based phase mapping three-dimensional digitization technology has a significant effect in solving the problem of three-dimensional digitization of objects with surface discontinuities or objects with a large surface gradient; and the composite coding method can be Obtain the carrier image of multiple encoding methods, and use this method to obtain multiple phase images to construct a multi-resolution three-dimensional digital image of the object; and to a certain extent, it can alleviate the incompleteness of the data caused by the lighting blind area and imaging blind area, which makes the phase unwrapping impossible. The problem. In addition, since the problems of phase ambiguity and error propagation are effectively overcome, the testable depth range and testable object range of this composite coding technology are larger than those of the single carrier technology.

Compared with existing methods and technologies, the present invention has the following advantages:

1) Combining the advantages of three-dimensional digital imaging methods based on two different encoding methods, dot matrix projection and fringe projection. It eliminates the problems of phase ambiguity and error propagation, improves the confidence of the measurement results, ensures high measurement accuracy, and alleviates the problem of incomplete data caused by information blind spots to a certain extent.

2) In order to meet different test needs, various forms of composite codes can be projected, including: dot matrix plus single frequency stripes and dot matrix plus time-series frequency-variable stripe sequences, and the fusion of data obtained by different sensing methods to achieve multiple resolutions, 3D digital imaging with different precision;

3) The testable depth range and testable object range of this composite coding technology are greater than that of the single carrier technology.

Description of drawings

Figure 1 is a schematic diagram of dot matrix imaging projected on a reference plane.

Fig. 2 is a schematic diagram of dot matrix imaging projected on any three-dimensional object.

Fig. 3 is a schematic structural diagram of a composite coded three-dimensional imaging system.

Figure 4 is a step-like object simulated by calculation.

Fig. 5 is a three-dimensional digital image of the object in Fig. 4 obtained by dot matrix projection method.

Fig. 6 is (a) expanded phase diagram of the object in Fig. 4 obtained by fringe projection method; (b) three-dimensional digital image.

Fig. 7 is a three-dimensional digital image of the object in Fig. 4 obtained by combining dot matrix projection with fringe projection using the composite encoding method proposed by the present invention.

Detailed ways

According to the above-mentioned method, the device for realizing composite encoding and multi-resolution three-dimensional digital imaging mainly includes a digital projection lighting transmitter, an image sensing receiver and an image processor. Said digital projection lighting transmitter can be a digital liquid crystal projection device (LCD projector), a digital micromirror projection device (DMD projector) or a silicon substrate liquid crystal projection device (LCOS projector), which can be conveniently provided by a computer image processing system. Generate two-dimensional dot matrix and fringe graphics and write them into a digital projection device; said image sensing receiver includes an optical imaging lens and a photodetector, and the optical imaging lens can be an imaging lens or lens group with a fixed focal length or a variable focal length, Binary optical imaging system, diffraction element imaging system, microscopic imaging system; said photodetection device can be charge coupled device, liquid crystal device, spatial light modulation device, CMOS device or digital camera. The image processor is a combination of a digital signal processor and a programmable application-specific integrated circuit, or a combination of a general-purpose image processing card and a computer. It is characterized in that the digital projection lighting transmitter projects two-dimensional spatial lattice and fringe pattern respectively, and the image sensing receiver receives the dot matrix and fringe code pattern modulated by the depth information of the measured object, and then performs corresponding decoding by the image processor Obtain a multi-resolution 3D depth image of an object.

The embodiment will be further described below in conjunction with the accompanying drawings.

Fig. 3 is a schematic structural diagram of a composite coded three-dimensional imaging system. As shown in the figure, the exit pupil P of the projection mirror 102 of the digital projection lighting transmitter 101, the entrance pupil I of the imaging lens 104 of the image sensing receiver 103 and the center o of the illumination field are located in the same plane and form a triangle, A triangulation system is formed. The two-dimensional lattice generated by the computer or digital signal processor of the image processor 105 is respectively projected on the reference plane 106 and the surface S of the object 107, and the angle between the projection optical axis Po and the imaging optical axis Io is α, expressed as PI The connecting line between is the baseline, which forms a number of triangles with the centers of all lattices projected on the object, each of which can form a triangulation system. The regular lattice on the reference plane and the deformed lattice representing the depth information of any free surface object shape are respectively received by the image sensing receiver 103 and sent to the image processor 105. By comparing the reference plane and the surface of the measured object The position differences of the dot matrix images formed by the upper dot matrix on the image sensing receiver correspond to the dot matrix, and the depth value of the measured object is calculated. Then, the digital projection lighting transmitter 101 continues to project the single frequency fringe generated by the image processor 105 to the surface 107 of the measured object, and then the image sensing receiver 103 receives the carrier fringe pattern and sends it to the image processor 105, combining the passing points The low-resolution three-dimensional depth value of the measured object obtained by the array projection method is used to perform phase unwrapping on the carrier fringe pattern. Blur and error propagation problems caused by information blind areas such as shadows can be effectively overcome, so as to obtain a complete three-dimensional digital image of the object. Another composite encoding method is dot matrix plus time-series frequency-variable fringe sequence. After obtaining the depth value of the object by means of dot-matrix projection, the digital projection lighting transmitter 101 can also project a time-series frequency-conversion fringe sequence to the surface of the measured object 107, through The known depth value of the object, combined with the system structure parameters, can calculate the total phase difference of the field width corresponding to the first phase jump, and also obtain the corresponding number of fringe cycles in the field of view, thus determining the timing frequency conversion The boundary conditions of the fringe sequence phase unwrapping algorithm reduce the number of iterations and improve the effectiveness of the algorithm.

Figure 4 is a step-like object simulated by calculation. The height of the step is equal to 1.1 times of the equivalent wavelength corresponding to the projected fringe when the fringe projection method is used to obtain the depth image of the object in FIG. 6 .

Fig. 5 is a three-dimensional digital image of the object in Fig. 4 obtained by using the lattice projection method. The number of dot matrix in the dot matrix graphics is 64×64. For the area where the gradient of the surface of the object changes greatly, due to the insufficient sampling rate of the projected dot matrix, it is impossible to obtain accurate results in this part of the area. As shown in the figure, only one gradient can be obtained. The result of the slow change is that the high-frequency information is lost, so that the area with a large gradient change is smoothed.

Fig. 6 is a three-dimensional digital image of an object acquired by fringe projection method. Figure 6(a) is the unfolded phase diagram obtained after phase unwrapping the carrier fringe pattern. Since the height of the step-like object in Fig. 4 is greater than the equivalent wavelength corresponding to the projected fringe, there is a phase ambiguity of 2π, so we cannot get The correct depth image, as shown in Figure 6(b), is the depth image calculated from Figure 6(a). The existence of phase blur makes the object depth map calculated from the unfolded phase image deviate, as shown in Figure 6(b) The height of the steps in the shown depth image does not match the actual height.

Fig. 7 is a three-dimensional digital image of the object in Fig. 4 obtained by using the composite coding method proposed by the present invention, combining the two coding methods of lattice projection and fringe projection, and using the phase unwrapping algorithm. First perform dot-matrix coding imaging to obtain a lower-resolution depth image of the object, then perform fringe coding imaging, and perform phase expansion according to the depth value of the object obtained by the dot-matrix projection method, as shown in Figure 6(a) The phase map has marked precise areas of different phase distributions, and the lighter colored areas in the figure are actually areas where phase jumps exist. At this time, the depth value of the object has been obtained by the method of dot matrix projection. Although the resolution is low, it still reflects the real distribution of the depth of the object from the thicker outline. The depth value obtained by dot matrix projection , combined with the conversion of the structural parameters of the system into phase values, the phase value obtained by lattice projection is used as the "control point" to control the process of phase unwrapping when the carrier fringe pattern is phase unwrapped. The steps are as follows: when the phase unwrapping reaches the position of these "control points", compare the phase value at this time with the phase value obtained by the lattice projection, if the error between the two is within the allowable range, the phase unwrapping continues; otherwise, it is considered that there is phase ambiguity, Calculate the phase difference between the two, and correct the phase value of the current point, as well as the phase values of all points returning to the edge of the "jump area" along the original phase unwrapping path, to solve the phase ambiguity problem; continue the phase unwrapping process, for each control Points are processed according to the above principles until all points in the area of the image that need to be phase unwrapped are processed. The results are shown in Figure 7. The phase ambiguity problem in Figure 6 has been resolved, and the gradient of the step-shaped object becomes steeper, and its depth image is more consistent with the actual result.

Claims (1)

1. A multi-resolution three-dimensional digital imaging method with composite coding, which adopts an imaging device including a digital projection lighting transmitter, an image sensor receiver and an image processor, and utilizes dot matrix coding and stripe coding to realize multi-resolution three-dimensional digital imaging. Its characteristics is to include the following processes: 1) The digital projection lighting transmitter projects two-dimensional dot matrix graphic lighting objects with different densities, deflects the image sensing receiver so that its optical axis is on the same line as the center of the projected dot matrix lighting field, and the output of the digital projection lighting transmitter The pupil, the entrance pupil of the image sensing receiver and the center of the illumination field form a triangle, taking the line between the exit pupil of the digital projection lighting transmitter and the entrance pupil of the image sensing receiver as the baseline, and it is projected on the object The centers of all lattices constitute several triangles, forming a triangulation system; according to the principle of affine transformation, the coded lattices projected on the reference plane and the surface of the measured object respectively, after coordinate rotation, translation and perspective projection transformation, respectively Obtain the analytical formula of the spatial coordinates of the coding lattice on the surface of the reference plane and the measured object on the image receiving sensor plane, and further obtain the imaging plane x between the corresponding points of the reference plane and the surface of the measured object. The relationship between the position difference in the direction and the depth value of the measured object at the corresponding point; thus the digital projection lighting transmitter projects a two-dimensional spatial lattice structured light illumination, and the image receiving sensor obtains the lattice image of the reference plane and projects it on the The dot matrix image on the surface of the measured object, according to the known geometric parameters of the optical path of the system, the image processor calculates the position difference between the corresponding points of the dot matrix image in the x direction, and then according to the position difference and the depth of the measured object at the corresponding point The relationship between the values calculates the depth image of the object; 2) The digital projection lighting transmitter projects the striped structured light to irradiate the surface of the measured object. The projected striped structured light is a single frequency strip, or a sequence of time-series frequency-variable strips; for a single frequency striped structured light, use phase mapping method, combined with The low-resolution depth image of the object that has been obtained by lattice projection is phase-unwrapped to further obtain a finer multi-resolution three-dimensional digital image; for the time-series frequency-variable fringe sequence structured light projection, the encoded fringe intensity distribution map sequence is obtained respectively, using " Phase Shift Algorithm" calculates the phase diagram of each frequency fringe intensity distribution diagram, and eliminates the 2π uncertainty in each phase diagram, so as to obtain the depth image of the scene with progressive resolution; at this time, the function of lattice projection is to use The object depth value obtained by it determines the boundary conditions of the variable frequency fringe sequence, which effectively saves the two-dimensional coded fringe pattern required in the phase unwrapping process and increases the real-time performance of the phase unwrapping algorithm; further obtains higher precision and finer object multi-resolution 3D digital image.

Images