CN104463863B - The scaling method and system of movement interference field based on the projection of time heterodyne - Google Patents

Abstract

The invention relates to a calibration method of motion interference field based on time heterodyne projection. First, two cameras are used to collect the deformed fringe pattern of the stereo target and the time parameter calibration plate fixedly placed in the field of view, and then the time parameter is calibrated accordingly and the modulation degree image of the stereo target is obtained, and then After normalizing the modulation degree image, detect the stereo target and two marker points, and use the marker points on the stereo target as a reference to perform spatial phase expansion on the folded phase images of the two cameras to obtain the stereoscopic images collected by the two cameras. The continuous phase distribution map of the target surface, and finally interpolate sub-pixel points on the continuous phase distribution map of one of the cameras, and then search for the corresponding sub-pixel points on the continuous phase distribution map of the other camera, and reconstruct isophase points accordingly Based on the three-dimensional data of , the isophase surface is fitted, and its folding phase and plane parameters are recorded. The invention can calibrate the moving interference field, so that the moving interference field can be used like a stable interference field.

Description

Calibration method and system for motion interference field based on temporal heterodyne projection

technical field

The invention belongs to the field of three-dimensional digital imaging, and in particular relates to a calibration method and system for a motion interference field based on temporal heterodyne projection.

Background technique

Dynamic surface imaging and measurement has a wide range of needs in the fields of assembly line inspection, military, experimental mechanics, and somatosensory games. At the same time, phase-assisted 3D imaging has the advantages of non-contact, fast speed, high precision, and high data density, and has been widely used in reverse engineering, quality control, defect detection, and cultural entertainment. Using phase-assisted means to perform dynamic surface imaging can significantly improve imaging accuracy and resolution, and phase is a physical quantity that cannot be directly observed. Phase shift algorithm is an important means to obtain phase information from fringe structured light. When using the phase-shift algorithm for dynamic surface imaging, motion has a greater impact on imaging, and reducing the number of deformed fringe images required for imaging is an important means to reduce this impact. Using the folded fringe map with a single spatial frequency can effectively reduce the number of images required for imaging, but it must use two constraints in the corresponding point search. In the system using a projector as a structured light generating device, since the projector can be calibrated by using the camera calibration method, only two cameras are needed to provide sufficient constraints. With the moving fringe interference field generated by temporal heterodyne interferometry, the use of two cameras is not enough to provide sufficient constraints.

Contents of the invention

The technical problem to be solved by the present invention is to provide a calibration method and system of motion interference field based on temporal heterodyne projection, so as to ensure that when two cameras and a single spatial frequency fringe are used, there are sufficient constraints to determine the corresponding point. The present invention is achieved like this:

A kind of calibration method of motion interference field based on temporal heterodyne projection, comprises the following steps:

Step A, fixedly place the three-dimensional target and the time parameter calibration plate with one marker point respectively in the field of view, and use two cameras to synchronously collect several frames of deformed fringe images containing the three-dimensional target and the time parameter calibration plate , at the same time, according to the deformed fringe images of each frame collected by each camera, use the corresponding phase shift algorithm to calculate the folded phase map and modulation degree image of each camera, and record the time parameter calibration board of the folded phase map of each camera The coordinates and phase values of the marker points; at the same time, normalize the modulation degree images of each camera, and perform image segmentation on the normalized modulation degree images to remove the background;

Step B, edge detection is performed on the modulation degree images of the two divided cameras, and the stereo target and two marker points are extracted. At the same time, based on the marker points on the stereo target, the spatial phase unfolds the folding of the two cameras. Phase map, to obtain the continuous phase distribution map of the three-dimensional target surface collected by the two cameras;

Step C, calculate the gradient of the phase along the x-axis of the image in the continuous phase distribution diagram of one of the cameras, use the gradient to obtain the interpolated phase interval, and interpolate sub-pixel points in the horizontal direction in each row, and at the same time, search for the other camera's The sub-pixel points corresponding to the interpolated sub-pixel points in the continuous phase distribution map, and reconstruct the three-dimensional data of the isophase points accordingly;

Step D, using the reconstructed three-dimensional data of isophase points to fit isophase surfaces, and recording the folding phase and plane parameters of said isophase surfaces.

Further, before solving the folded phase images and modulation degree images of each frame of deformed fringe images collected by each camera, the step of filtering the several frames of deformed fringe images is also included;

The method of calculating the folded phase map and modulation image of each camera is a phase shift algorithm;

The image segmentation method is a threshold segmentation method.

Further, the edge detection adopts Canny operator;

The method for extracting the three-dimensional target and the two marker points is as follows: performing ellipse fitting on the edge detection image data obtained by the edge detection, and judging and extracting the three-dimensional marker by the relative position of the fitting error and the abscissa of the data. Target with two marker points.

Further, said step C includes the following steps:

Step C1, using one line of data on the continuous phase distribution map of one of the cameras, use the least squares method to fit the gradient of the phase along the abscissa direction of the image, and obtain the interpolated phase interval by the following formula:

in Represents the gradient of the phase along the abscissa, φ _plug represents the phase interval of the calculated interpolation, and round represents the rounding operation;

Step C2, according to the interpolated phase interval, interpolate sub-pixel points along each row on the continuous phase distribution map of the selected camera to obtain a new continuous phase distribution map;

In step C3, the sub-pixel point corresponding to the interpolated sub-pixel point is found in the continuous phase distribution map of another camera through the method of auxiliary epipolar line constraint, and the three-dimensional data of the isophase point is reconstructed accordingly.

Further, the step D fits the equiphase surface through the following formula:

In the formula, x _n , y _n , z _n are the three-dimensional coordinate values of space points, n is the number of equiphase points, a, b, c are the reciprocals of the intercepts of equiphase planes on the three coordinate axes.

A calibration system for motion interference field based on temporal heterodyne projection, comprising:

The deformed fringe image acquisition unit is used to use two cameras to synchronously collect a number of frames of a three-dimensional target and a time parameter calibration board that are fixedly placed in the field of view and each have a mark point, including the three-dimensional target and the time parameter calibration board. The deformed fringe pattern;

The image processing unit is used to calculate the folded phase map and modulation degree image of each camera according to the deformed fringe map of each frame collected by each camera, and record the time parameter calibration of the folded phase map of each camera The coordinates and phase values of the marker points on the board; at the same time, normalize the modulation degree images of each camera, and perform image segmentation on the normalized modulation degree images to remove the background;

The edge detection and phase unwrapping unit is used to perform edge detection on the modulation degree images of the two divided cameras, and extract the three-dimensional target and two marker points. At the same time, based on the marker points on the three-dimensional target, the spatial phase Unfold the folded phase diagrams of the two cameras to obtain the continuous phase distribution diagram of the three-dimensional target surface collected by the two cameras;

The image interpolation and three-dimensional reconstruction unit is used to calculate the gradient of the phase along the x-axis of the image in the continuous phase distribution diagram of one of the cameras, use the gradient to obtain the interpolated phase interval, and interpolate sub-pixel points along the horizontal direction in each row, and at the same time, Search for the sub-pixel point corresponding to the interpolated sub-pixel point in the continuous phase distribution map of another camera, and reconstruct the 3D data of the isophase point accordingly;

The isophase surface fitting unit is configured to use the reconstructed three-dimensional data of the isophase points to fit the isophase surface, and record the folding phase and plane parameters of the isophase surface.

Further, the image processing unit includes:

An image filtering module, which is used to carry out Gaussian filtering to the deformed fringe images of each frame collected by the two cameras;

The phase demodulation module is used to perform phase demodulation on the filtered deformed fringe images of each frame of each camera, obtain the folded phase map and modulation degree image of each camera, and record the time parameters of the folded phase map of each camera The coordinates and phase values of the marker points on the calibration plate;

The modulation degree segmentation module is used to normalize the modulation degree images of each camera, and perform image segmentation on the normalized modulation degree images to remove the background.

Further, the edge detection and phase unwrapping unit includes:

The edge detection module is used to carry out edge detection to the modulation degree images of the two cameras after segmentation according to the Canny operator;

The target recognition module is used to perform ellipse fitting on the detected edge detection image, and judge and extract the three-dimensional target and two marker points through the relative position of the fitting error and the abscissa of the data;

The phase unwrapping module is used to unfold the folded phase images of the two cameras in spatial phase based on the marker points on the three-dimensional target, so as to obtain the continuous phase distribution image of the surface of the three-dimensional target collected by the two cameras.

Further, the image interpolation and three-dimensional reconstruction unit includes:

The sub-pixel interpolation unit is used to use one line of data on the continuous phase distribution map of one of the cameras to fit the gradient of the phase along the abscissa direction of the image by the least square method, and obtain the interpolated phase interval by the following formula:

in Represents the gradient of the phase along the abscissa, φ _plug represents the phase interval of the calculated interpolation, and round represents the rounding operation;

The three-dimensional reconstruction unit is used to interpolate sub-pixel points along each row on the continuous phase distribution map of the selected camera according to the phase interval of the interpolation to obtain a new continuous phase distribution map; and find another The sub-pixel points corresponding to the interpolated sub-pixel points in the continuous phase distribution map of the first camera are used to reconstruct the three-dimensional data of the isophase points.

Further, the isophase surface fitting unit fits the isophase surface through the following formula:

In the formula, x _n , y _n , z _n are the three-dimensional coordinate values of space points, n is the number of equiphase points, a, b, c are the reciprocals of the intercepts of equiphase planes on the three coordinate axes.

The calibration system provided by the present invention has high reliability. Since the least squares method is used to fit isophase surfaces, ordinary objects can be used instead of high-precision stereo targets in the calibration process, which greatly reduces the need for stereo targets. The requirement of processing accuracy reduces the production cost of the three-dimensional target. At the same time, because the mark points on the time parameter calibration plate record the phase information related to the time parameter, the calibration system can calibrate the interference field in motion, which makes the motion interference field Can be used the same as Stabilizing Interfering Field.

Description of drawings

Figure 1: Schematic flow chart of the calibration method of the motion interference field based on temporal heterodyne projection provided by the present invention;

Figure 2: Schematic diagram of the calibration system of the motion interference field based on temporal heterodyne projection provided by the present invention;

Figure 3: Schematic diagram of the structure of the deformed fringe pattern acquisition unit in the above calibration system;

Fig. 4: A schematic diagram of the principle of using two CCD cameras to find corresponding points provided by the embodiment.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments.

Fig. 1 shows the calibration method flow of the motion interference field based on temporal heterodyne projection provided by the present invention. According to Fig. 1, the calibration method specifically includes the following steps:

Step A, fixedly place the three-dimensional target and the time parameter calibration plate with one marker point respectively in the field of view, and use two cameras to synchronously collect several frames of deformed fringe images containing the three-dimensional target and the time parameter calibration plate, and at the same time , according to the deformed fringe images of each frame collected by each camera, use the corresponding phase shift algorithm to calculate the folded phase map and modulation degree image of each camera, and record the mark points on the time parameter calibration board of each camera’s folded phase map coordinates and phase values; at the same time, normalize the modulation degree images of each camera, and perform image segmentation on the normalized modulation degree images to remove the background.

It should be noted that a pattern with two concentric circles, one large and one small, can be used as the marking point, wherein the inner circle is white, and the ring between the outer circle and the inner circle is black for easy identification. The "time parameter calibration board" is mainly used to calibrate the time parameters. Mark points can be pasted on the fixed white board as a time parameter calibration board. The fixed white board can make it more convenient to judge the edge of the image, and the position of the fixed white board is different during calibration and imaging. It can be moved so that the phase change brought by the self-calibration time parameter when the object is imaged. The placement positions of the three-dimensional target and the fixed whiteboard need to separate the three-dimensional target and the fixed whiteboard in the images collected by the two cameras, so as to facilitate subsequent extraction of the three-dimensional target. Since only the position calibrated by the stereo target can have equi-phase plane constraints, under the premise of ensuring the separation of the stereo target and the fixed whiteboard, the size of the stereo target should occupy the field of view as much as possible, so as to facilitate the imaging of the object surface. At the same time, the size of the marker points needs to be limited. Since the ring data between the outer circle and the inner circle are discarded after the modulation threshold is divided, the difference between the outer circle diameter and the inner circle diameter cannot exceed one space period. In order to facilitate subsequent spatial phase unwrapping and avoid error transmission. In the embodiment of the present invention, before calculating the folded phase map and the modulation degree image of each camera, it may also include the step of filtering several frames of deformed fringe images collected by each camera. The filtering method may be Gaussian filtering, and the phase The phase-shifting algorithm demodulates the phase to solve the folded phase map and modulation image of each frame of the deformed fringe image collected by each camera. The threshold segmentation method can be used to segment the modulation image to remove the background and extract the three-dimensional target and time parameters. Calibration board.

Step B, edge detection is performed on the modulation degree images of the two divided cameras, and the stereo target and two marker points are extracted. At the same time, based on the marker points on the stereo target, the spatial phase unfolds the folding of the two cameras. Phase map, to obtain the continuous phase distribution map of the stereo target surface collected by the two cameras.

It should be noted that when extracting three-dimensional targets and marker points, grayscale, shape, and position criteria can be used to remove stray edges through grayscale criteria, and the error of ellipse fitting can be used to distinguish marker points from Three-dimensional target, by fitting the position of the center of the circle, different marker points can be distinguished and the three-dimensional target can be distinguished from the time parameter calibration plate. In the embodiment of the present invention, the Canny operator can be used to perform edge detection on the modulation degree images of the two divided cameras, and the ellipse fitting can be performed on the edge detection image data obtained by the edge detection, and the fitting error and the data The relative position of the abscissa of the three-dimensional target is judged and extracted, and the three-dimensional target is used as the reference point, and the space phase unfolds the folded phase diagram of the two cameras to obtain the three-dimensional target of the two cameras Continuous phase profile of a surface. When a pattern with two concentric circles, one large and one small, is used as the mark point, after the two mark points are extracted, the folding phase of the inner circle part of the mark point on the time parameter calibration board can be recorded, and the three-dimensional target can be used to The marked points of the three-dimensional target are used as a reference, and the spatial phase unfolds the folded phase maps of the three-dimensional target surface collected by the two cameras, so as to obtain the continuous phase distribution map of the three-dimensional target surface collected by the two cameras. It should be noted that the reason for phase unwrapping is that the phase unwrapping must be used to search for corresponding points with dual cameras so that the phase can monotonically increase in the lateral direction, and the auxiliary epipolar constraints can uniquely determine the corresponding points.

Step C, calculate the gradient of the phase along the x-axis of the image in the continuous phase distribution diagram of one of the cameras, use the gradient to obtain the interpolated phase interval, and interpolate sub-pixel points in the horizontal direction in each row, and at the same time, search for the other camera's The sub-pixel points corresponding to the interpolated sub-pixel points in the continuous phase distribution map are used to reconstruct the three-dimensional data of the isophase points.

In the embodiment of the present invention, step C specifically includes the following steps:

Step C1, using one line of data on the continuous phase distribution map of one of the cameras, use the least squares method to fit the gradient of the phase along the abscissa direction of the image, and obtain the interpolated phase interval by the following formula:

in Represents the gradient of the phase along the abscissa, φ _plug represents the phase interval of the calculated interpolation, and round represents the rounding operation;

Step C2, according to the interpolated phase interval, interpolate sub-pixel points along each row on the continuous phase distribution map of the selected camera to obtain a new continuous phase distribution map;

In step C3, the sub-pixel point corresponding to the interpolated sub-pixel point is found in the continuous phase distribution map of another camera through the method of auxiliary epipolar line constraint, and the three-dimensional data of the isophase point is reconstructed accordingly.

It should be noted that the purpose of equally dividing the space period is to reduce the number of phase values of the isophase surface and to facilitate the calculation when searching for corresponding points.

Step D, using the reconstructed 3D data of isophase points to fit isophase surfaces, and recording folded phase and plane parameters of isophase surfaces.

In the embodiment of the present invention, the three-dimensional data of the reconstructed isophase point can be used to fit the isophase surface through the following formula:

In the formula, x _n , y _n , z _n are the three-dimensional coordinate values of space points, n is the number of equiphase points, a, b, c are the reciprocals of the intercepts of equiphase planes on the three coordinate axes.

It should be noted that the recorded phase must be a folded phase, because the folded phase value is used in the corresponding point search process, so the continuous phase value exceeds the value range.

Fig. 2 shows the calibration system structure of the motion interference field based on temporal heterodyne projection provided by the embodiment of the present invention. According to Fig. 2, the calibration system includes:

Deformed fringe image acquisition unit 1, used to utilize two cameras to synchronously acquire a plurality of frames of a three-dimensional target and a time parameter calibration board that are fixedly placed in the field of view and respectively provided with a mark point, including the three-dimensional target and the time parameter calibration board deformed fringe graph;

The image processing unit 2 is used to calculate the folded phase map and the modulation degree image of each camera according to the deformed fringe map of each frame collected by each camera using the corresponding phase shift algorithm, and record the time parameters of the folded phase map of each camera The coordinates and phase values of the marker points on the calibration board; at the same time, normalize the modulation degree images of each camera, and perform image segmentation on the normalized modulation degree images to remove the background;

The edge detection and phase unwrapping unit 3 is used to perform edge detection on the modulation degree images of the two divided cameras, and extract the three-dimensional target and two marker points. At the same time, based on the marker points on the three-dimensional target, spatial Phase unfold the folded phase diagrams of the two cameras to obtain the continuous phase distribution diagram of the three-dimensional target surface collected by the two cameras;

The image interpolation and three-dimensional reconstruction unit 4 is used to calculate the gradient of the phase along the x-axis of the image in the continuous phase distribution diagram of one of the cameras, use the gradient to obtain the interpolated phase interval, and interpolate the sub-pixel points along the horizontal direction in each row, and at the same time , search for the sub-pixel point corresponding to the interpolated sub-pixel point in the continuous phase distribution map of another camera, and reconstruct the three-dimensional data of the isophase point accordingly;

The isophase surface fitting unit 5 is configured to use the reconstructed 3D data of the isophase points to fit the isophase surface, and record the folding phase and plane parameters of the isophase surface.

As shown in Figure 3, in the embodiment of the present invention, the deformed fringe image acquisition unit 1 can adopt the following structural form:

Including a laser 101, two CCD cameras 103, an external trigger signal generator 105, a computer 106, a radio frequency signal generator 107, two acousto-optic deflectors 108, two lenses 110, two dichroic prisms 112, and two mirrors 114 , two diaphragms 116, and a microscope objective lens 118. In the embodiment of the present invention, a CCD camera is used to collect the deformed fringe image, and of course other types of cameras may also be used to collect the deformed fringe image. The radio frequency signal generator 107 is used to generate two radio frequency signals with a slight frequency difference, and the external trigger signal generator 105 is used to generate a fixed-frequency rectangular wave as an external trigger signal. The laser 101, two dichroic prisms 112, and two reflectors 114 form a Mach-Zehnder interferometer. Two acousto-optic deflectors 108 are placed in two interference arms, and two-way interference is modulated under the control of a radio frequency signal generator 107. Light, so that the frequency of the first-order diffracted light produces a small frequency difference Δf, after the two paths of interfering light are collimated by two lenses 110 respectively, other diffracted beams are filtered out by two diaphragms 116, and projected to A moving interference fringe is generated on the surface of the measured object. The spatial frequency of the interference fringes is fixed, and the gray level of any spatial point changes with time as a sinusoidal function, and the frequency of change is Δf. The external trigger signal generator 105 generates an external trigger signal with a frequency of 4Δf to control the two CCD cameras 103 to collect the deformed fringe pattern. When collecting the deformed fringe pattern, the two CCD cameras 103 can collect synchronously according to the frequency difference 4 times of the two radio frequency signals .

In the embodiment of the present invention, the image processing unit 2 includes:

An image filtering module, which is used to carry out Gaussian filtering to the deformed fringe images of each frame collected by the two cameras;

The phase demodulation module is used to perform phase demodulation on the filtered deformed fringe images of each frame of each camera, obtain the folded phase map and modulation degree image of each camera, and record the time parameters of the folded phase map of each camera The coordinates and phase values of the marker points on the calibration plate;

The modulation degree segmentation module is used to normalize the modulation degree images of each camera, and perform image segmentation on the normalized modulation degree images to remove the background.

In the embodiment of the present invention, the edge detection and phase unwrapping unit 3 includes:

The edge detection module is used to carry out edge detection to the modulation degree images of the two cameras after segmentation according to the Canny operator;

The target recognition module is used to perform ellipse fitting on the detected edge detection image, and judge and extract the three-dimensional target and two marker points through the relative position of the fitting error and the abscissa of the data;

The phase unwrapping module is used to unfold the folded phase images of the two cameras in spatial phase based on the marker points on the three-dimensional target, so as to obtain the continuous phase distribution image of the surface of the three-dimensional target collected by the two cameras.

In the embodiment of the present invention, the image interpolation and three-dimensional reconstruction unit 4 includes:

The sub-pixel interpolation unit is used to use one line of data on the continuous phase distribution map of one of the cameras to fit the gradient of the phase along the abscissa direction of the image by the least square method, and obtain the interpolated phase interval by the following formula:

in Represents the gradient of the phase along the abscissa, φ _plug represents the phase interval of the calculated interpolation, and round represents the rounding operation;

The three-dimensional reconstruction unit is used to interpolate sub-pixel points along each row on the continuous phase distribution map of the selected camera according to the phase interval of the interpolation to obtain a new continuous phase distribution map; and find another The sub-pixel points corresponding to the interpolated sub-pixel points in the continuous phase distribution map of the first camera are used to reconstruct the three-dimensional data of the isophase points.

In the embodiment of the present invention, the isophase surface fitting unit 5 uses the image interpolation and the three-dimensional data of the isophase point reconstructed by the three-dimensional reconstruction unit 3 to fit the isophase surface by the following formula:

In the formula, x _n , y _n , z _n are the three-dimensional coordinate values of space points, n is the number of equiphase points, a, b, c are the reciprocals of the intercepts of equiphase planes on the three coordinate axes.

In order to describe the present invention in detail, the process of searching for corresponding points using calibration data will be described below in conjunction with FIG. 4 . First of all, it is necessary to correct the overall change of the phase caused by the time parameters. The correction data is obtained by using the phases of the marker points on the time parameter calibration board at the calibration time _t1 and the imaging time _t2 . Specifically, it is calculated by the following method:

Ideally, the phase value of any pixel point t ₁ and t ₂ can be used to calibrate this time, but in practice, due to the influence of errors, if a pattern with two concentric circles, one large and one small, is used as the marker point , this paper can use all the pixels in the inner circle of the entire marker point to solve this phase change. The following formula is the calculation formula for solving the phase change:

b＝Mod(φ _i (t ₂ )-φ _i (t ₁ ),2π)

i=1,...,N

Among them, i is the serial number of the pixel, N is the number of the pixel, φ _i (t ₁ ), φ _i (t ₂ ) represent the phase value of the i-th pixel at time t ₁ and t ₂ respectively, Mod(φ _i (t ₂ )-φ _i (t ₁ ),2π) represents the operation of dividing φ _i (t ₂ )-φ _i (t ₁ ) by 2π and taking the remainder, is the average value of the phase changes of all inner circle pixels. The corrected equiphase surface phase data can be used to search for corresponding points, the specific process is as follows:

As shown in Figure 4, the image planes of the left and right cameras are I _l and I _r respectively, and the optical centers of the left and right cameras are O _l and O _r respectively; a point P ₅ (x,y,z) on the object plane, Imaged on the point p _l (x _l ,y _l ) of the left camera, the phase value is φ(p _l (x _l ,y _l )), according to the optical center O _l of the left camera and the pixel point p _l (x _l ,y _l ) can be connected to obtain the space straight line The equation, the space point imaged in p _l (x _l ,y _l ) must be located on the space line superior. According to the calibrated interferometric field data and the straight line The equation of can obtain the coordinates of a set of spatial points M ₁ , M ₂ , M ₃ , M ₄ and P ₅ . p _l (x _l ,y _l ) has a group of points p _r1 , p _r2 , p _r3 , p _r4 , p _r5 on the epipolar line corresponding to the right camera, and their phases are equal. The calibration parameters of the right camera reconstruct the coordinate positions of a group of points in space, and the coordinates of another group of space points N ₁ , N ₂ , N ₃ , N ₄ and P ₅ can be obtained. But in fact, the object points corresponding to this group of points in space are P ₁ , P ₂ , P ₃ , P ₄ , and P ₅ . In theory, only the coordinates of point P ₅ in the two groups of points should be coincident. Therefore, the position of point _P5 can be determined by cross-comparing the spatial distances of the two groups of points.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention should be included in the protection of the present invention. within range.

Claims (8)

1. a kind of scaling method of the movement interference field based on the projection of time heterodyne, it is characterised in that comprise the following steps：

Step A, will be respectively equipped with the three-dimensional target of an index point with time parameter scaling board fixed placement in visual field, and profit The deforming stripe figure of the three-dimensional target and time parameter scaling board is included with two some frames of camera synchronous acquisition, meanwhile, root According to each frame deforming stripe figure of every camera collection, the wrapped phase figure of every camera is calculated with adjusting using corresponding phase shift algorithm System image, and record the coordinate and phase value of the index point in the wrapped phase figure of every camera on time parameter scaling board； Meanwhile the modulation degree image of every camera is normalized, and image segmentation is carried out to the modulation degree image after normalization, go Except background；

The time parameter scaling board is the fixation blank equipped with index point, the mark of the solid target and time parameter scaling board Will point is the pattern of small one and large one two concentric circles, and wherein inner circle is white, and the cylindrical annulus between inner circle is black；It is described The difference of outer diameter of a circle and interior diameter of a circle is no more than a space periodic；

Step B, carries out edge detection to the modulation degree image of two cameras after segmentation, extracts three-dimensional target and two marks Point, meanwhile, on the basis of the index point on three-dimensional target, the wrapped phase figure of two cameras is unfolded in space phase, obtains two The continuous phase distribution map of the three-dimensional target surface of camera collection；

Step C, is calculated gradient of the phase along image x-axis in the continuous phase distribution map of wherein one camera, is obtained using gradient The phase intervals of interpolation, and in every row interpolation sub-pix point in transverse direction, meanwhile, the continuous phase point of another camera of search Sub-pix point corresponding with the sub-pix point of institute interpolation in Butut, and the three-dimensional data of equiphase point is rebuild accordingly；

The step C includes the following steps：

Step C1, using the data line on the continuous phase distribution map of wherein one camera, passes through least square fitting phase Gradient of the position along image abscissa direction, and the phase intervals of interpolation are obtained by equation below：

<mrow> <msub> <mi>&amp;phi;</mi> <mrow> <mi>p</mi> <mi>l</mi> <mi>u</mi> <mi>g</mi> </mrow> </msub> <mo>=</mo> <mfrac> <mrow> <mn>2</mn> <mi>&amp;pi;</mi> </mrow> <mrow> <mi>r</mi> <mi>o</mi> <mi>u</mi> <mi>n</mi> <mi>d</mi> <mrow> <mo>(</mo> <mfrac> <mrow> <mn>2</mn> <mi>&amp;pi;</mi> </mrow> <msub> <mo>&amp;dtri;</mo> <mi>x</mi> </msub> </mfrac> <mo>)</mo> </mrow> </mrow> </mfrac> </mrow>

WhereinRepresent gradient of the phase along abscissa, φ_plugRepresent the phase intervals of interpolation being calculated, round is represented Round up computing；

Step C2, it is sub- along each row interpolation on the continuous phase distribution map of selected camera according to the phase intervals of interpolation Pixel, obtains new continuous phase distribution map；

Step C3, by aiding in the method for epipolar-line constraint to search the sub- picture in the continuous phase distribution map of another camera with interpolation The corresponding sub-pix point of vegetarian refreshments, and the three-dimensional data of equiphase point is rebuild accordingly；

Step D, equiphase surface is fitted using the three-dimensional data of the equiphase point of reconstruction, and records the folding phase of the equiphase surface Position and plane parameter.

2. the scaling method of movement interference field as claimed in claim 1, it is characterised in that solve each frame of every camera collection The wrapped phase figure of deforming stripe figure is with before modulation degree image, further including what some frame deforming stripe figures were filtered Step；

The method of the wrapped phase figure and modulation degree image that calculate every camera is phase shift algorithm；

The method of described image segmentation is thresholding method.

3. the scaling method of movement interference field as claimed in claim 1, it is characterised in that the edge detection uses Canny Operator；

The method of extraction three-dimensional target and two index points is：The edge-detected image data obtained to the edge detection carries out Ellipse fitting, and by error of fitting and the relative position of the abscissa of data, judge and extract the three-dimensional target and two Index point.

4. the scaling method of movement interference field as claimed in claim 1, it is characterised in that the step D passes through equation below It is fitted equiphase surface：

<mrow> <mfenced open = "[" close = "]"> <mtable> <mtr> <mtd> <msub> <mi>x</mi> <mn>1</mn> </msub> </mtd> <mtd> <msub> <mi>y</mi> <mn>1</mn> </msub> </mtd> <mtd> <msub> <mi>z</mi> <mn>1</mn> </msub> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <msub> <mi>x</mi> <mi>n</mi> </msub> </mtd> <mtd> <msub> <mi>y</mi> <mi>n</mi> </msub> </mtd> <mtd> <msub> <mi>z</mi> <mi>n</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mfenced open = "[" close = "]"> <mtable> <mtr> <mtd> <mi>a</mi> </mtd> </mtr> <mtr> <mtd> <mi>b</mi> </mtd> </mtr> <mtr> <mtd> <mi>c</mi> </mtd> </mtr> </mtable> </mfenced> <mo>=</mo> <mfenced open = "[" close = "]"> <mtable> <mtr> <mtd> <mn>1</mn> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mn>1</mn> </mtd> </mtr> </mtable> </mfenced> </mrow>

X in formula_n,y_n,z_nFor the D coordinates value of spatial point, n is the number of equiphase point, and a, b, c is equiphase surface in three seats The inverse of intercept on parameter.

A kind of 5. calibration system of the movement interference field based on the projection of time heterodyne, it is characterised in that including：

Deforming stripe figure collecting unit, for being respectively equipped with one in visual field using two camera synchronous acquisition fixed placements The three-dimensional target of index point includes the change of the three-dimensional target and time parameter scaling board with some frames of time parameter scaling board Shape bar graph；

The time parameter scaling board is the fixation blank equipped with index point, the mark of the solid target and time parameter scaling board Will point is the pattern of small one and large one two concentric circles, and wherein inner circle is white, and the cylindrical annulus between inner circle is black；It is described The difference of outer diameter of a circle and interior diameter of a circle is no more than a space periodic；

Image processing unit, for each frame deforming stripe figure gathered according to every camera, is calculated using corresponding phase shift algorithm The wrapped phase figure of every camera and modulation degree image, and record in the wrapped phase figure of every camera on time parameter scaling board Index point coordinate and phase value；Meanwhile the modulation degree image of every camera is normalized, and to the tune after normalization System image carries out image segmentation, removes background；

Edge detection and phase unwrapping unit, for carrying out edge detection to the modulation degree image of two cameras after segmentation, carry Three-dimensional target and two index points are taken out, meanwhile, on the basis of the index point on three-dimensional target, two cameras are unfolded in space phase Wrapped phase figure, obtain the continuous phase distribution map of the three-dimensional target surface of two cameras collection；

Image interpolation and three-dimensional reconstruction unit, phase is along image x in the continuous phase distribution map for calculating wherein one camera The gradient of axis, the phase intervals of interpolation are obtained using gradient, and in every row interpolation sub-pix point in transverse direction, meanwhile, search Sub-pix point corresponding with the sub-pix point of institute interpolation in the continuous phase distribution map of another camera, and equiphase is rebuild accordingly The three-dimensional data of point；

Described image interpolation includes with three-dimensional reconstruction unit：

Sub-pixel interpolation unit, for the data line on the continuous phase distribution map using wherein one camera, passes through minimum Gradient of the square law fit phase along image abscissa direction, and the phase intervals of interpolation are obtained by equation below：

<mrow> <msub> <mi>&amp;phi;</mi> <mrow> <mi>p</mi> <mi>l</mi> <mi>u</mi> <mi>g</mi> </mrow> </msub> <mo>=</mo> <mfrac> <mrow> <mn>2</mn> <mi>&amp;pi;</mi> </mrow> <mrow> <mi>r</mi> <mi>o</mi> <mi>u</mi> <mi>n</mi> <mi>d</mi> <mrow> <mo>(</mo> <mfrac> <mrow> <mn>2</mn> <mi>&amp;pi;</mi> </mrow> <msub> <mo>&amp;dtri;</mo> <mi>x</mi> </msub> </mfrac> <mo>)</mo> </mrow> </mrow> </mfrac> </mrow>

WhereinRepresent gradient of the phase along abscissa, φ_plugRepresent the phase intervals of interpolation being calculated, round is represented Round up computing；

Three-dimensional reconstruction unit, for the phase intervals according to interpolation, along every on the continuous phase distribution map of selected camera Row interpolation sub-pix point, obtains new continuous phase distribution map；And by aiding in the method for epipolar-line constraint to search another phase Sub-pix point corresponding with the sub-pix point of interpolation in the continuous phase distribution map of machine, and three dimensions of equiphase point are rebuild accordingly According to；

Equiphase surface fitting unit, for being fitted equiphase surface using the three-dimensional data for the equiphase point rebuild, and described in record The wrapped phase and plane parameter of equiphase surface.

6. the calibration system of movement interference field as claimed in claim 5, it is characterised in that described image processing unit includes：

Image filtering module, for carrying out gaussian filtering to each frame deforming stripe figure of two camera collections；

Phase demodulation modules, for carrying out phase demodulating to each frame deforming stripe figure of filtered every camera, obtain every The wrapped phase figure of camera and modulation degree image, and record the mark in the wrapped phase figure of every camera on time parameter scaling board The coordinate and phase value of will point；

Modulation degree splits module, for the modulation degree image of every camera to be normalized, and to the modulation degree after normalization Image carries out image segmentation, removes background.

7. the calibration system of movement interference field as claimed in claim 5, it is characterised in that the edge detection and phase unwrapping Unit includes：

Edge detection module, for carrying out edge detection to the modulation degree image of two cameras after segmentation according to Canny operators；

Target identification module, for carrying out ellipse fitting to the edge-detected image detected, and passes through error of fitting and data Abscissa relative position, judge and extract three-dimensional target and two index points；

Phase unwrapping module, on the basis of the index point on three-dimensional target, the folding phase of two cameras to be unfolded in space phase Bitmap, obtains the continuous phase distribution map of the three-dimensional target surface of two camera collections.

8. the calibration system of movement interference field as claimed in claim 5, it is characterised in that the equiphase surface fitting unit leads to Cross equation below fitting equiphase surface：

<mrow> <mfenced open = "[" close = "]"> <mtable> <mtr> <mtd> <msub> <mi>x</mi> <mn>1</mn> </msub> </mtd> <mtd> <msub> <mi>y</mi> <mn>1</mn> </msub> </mtd> <mtd> <msub> <mi>z</mi> <mn>1</mn> </msub> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <msub> <mi>x</mi> <mi>n</mi> </msub> </mtd> <mtd> <msub> <mi>y</mi> <mi>n</mi> </msub> </mtd> <mtd> <msub> <mi>z</mi> <mi>n</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mfenced open = "[" close = "]"> <mtable> <mtr> <mtd> <mi>a</mi> </mtd> </mtr> <mtr> <mtd> <mi>b</mi> </mtd> </mtr> <mtr> <mtd> <mi>c</mi> </mtd> </mtr> </mtable> </mfenced> <mo>=</mo> <mfenced open = "[" close = "]"> <mtable> <mtr> <mtd> <mn>1</mn> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mn>1</mn> </mtd> </mtr> </mtable> </mfenced> </mrow>

X in formula_n,y_n,z_nFor the D coordinates value of spatial point, n is the number of equiphase point, and a, b, c is equiphase surface in three seats The inverse of intercept on parameter.