CN106846469A - The method and apparatus that feature based point tracking is reconstructed three-dimensional scenic by focusing storehouse - Google Patents

Abstract

The invention discloses the method and apparatus that a kind of feature based point tracking is reconstructed three-dimensional scenic by focusing storehouse, including：The forward model that storehouse is focused on by three-dimensional scenic generation is set up, three-dimensional scenic characteristic point is given and is focused on the geometrical relationship of storehouse；The characteristic point for focusing on each image in storehouse is extracted, the coordinate of the characteristic point that the match is successful is followed the trail of, track of the characteristic point in storehouse is focused on is obtained；Set up the inverse model that three-dimensional scenic is reconstructed by focusing storehouse：By the characteristic point that the match is successful, the equation group on characteristic point three-dimensional coordinate is set up, the three-dimensional coordinate of characteristic point is obtained by solving equation group, reconstruct three-dimensional scenic, and realize that three-dimensional geometry is measured.Focusing storehouse of the invention is to fix detector, by the collection that focusing storehouse is completed along optical axis mobile lens, using the solution of the present invention, the three-dimensionalreconstruction under camera shooting visual angle can be realized, accurate three-dimensional structure information can be provided for virtual reality and dimensional measurement.

Description

The method and apparatus that feature based point tracking is reconstructed three-dimensional scenic by focusing storehouse

Technical field

The present invention relates to computer vision and digital image processing field, more particularly to a kind of feature based point tracking is by gathering Coke pile stack reconstructs three-dimensional scenic method and apparatus.

Background technology

The photo that traditional camera shoots be the light that sends of three-dimensional scenic by lens on two-dimensional detector light intensity it is tired Plus and, so cause radiation direction information and depth information of scene to be compressed.It is research three emerging in recent years to calculate photography The frontier of imaging is tieed up, it is one of method of three-dimensional scenic reconstruct in calculating photography to focus on storehouse.It is one group poly- to focus on storehouse The burnt imaging sequence shot in different imaging planes or using different parameters, with abundant three-dimensional information.In computer vision In, mainly carry out reconstruct and the dimensional measurement of three-dimensional scenic to focus on the degree for defocusing and focusing on of storehouse.It is existing by focusing on The method of storehouse reconstruct three-dimensional scenic is the reconstructing method based on Focus field emission array (could also say that zoom method), it is necessary to calculate focusing The Focus field emission array of each image of storehouse.

The content of the invention

It is an object of the invention to provide a kind of feature based point tracking by focusing storehouse reconstruct three-dimensional scenic method and Device, the present invention uses tracing characteristic points to provide effective means by focusing storehouse reconstruct three-dimensional scenic, how visual is different from Feel reconstruct three-dimensional scenic method, the present invention provide feature based point tracking by focusing storehouse reconstruct three-dimensional scenic method and Device, it is no longer necessary to estimate camera internal reference and outer ginseng, disclosure satisfy that and shoot the angle of visual field (Field Of to camera in the prior art View, FOV) under three-dimensional scenic reconstruct demand.

To achieve the above object, the present invention provides the side that a kind of feature based point tracking is reconstructed three-dimensional scenic by focusing storehouse Method, methods described includes：The forward model that storehouse is focused on by three-dimensional scenic generation is set up, three-dimensional scenic characteristic point is obtained and is focused on The relation of storehouse；The characteristic point of each image is extracted from the focusing storehouse, storehouse is being focused on just according to three-dimensional scenic generation Drill the coordinate that the characteristic point that the match is successful in storehouse is focused on described in model tracking；Set up the three-dimensional coordinate of the characteristic point that the match is successful With the equation group of three-dimensional scenic, the three-dimensional coordinate that equation group obtains the characteristic point that the match is successful is solved, reconstruct three-dimensional scenic.

Further, the forward model of the foundation is：

The three-dimensional scenic characteristic point with focus on storehouse relation be：

Wherein, (S_x,S_y) it is the object plane of scene, f (S_x,S_y) it is point (S_x,S_y) corresponding irradiation level, depth (S_x,S_y) be The depth map of three-dimensional scenic, d_nTo focus on the n-th imaging lens of storehouse to object plane (S_x,S_y) distance, d '_nTo focus on storehouse I ＆ lt imaging lens to detector distance, d '₁To focus on the 1st imaging lens of storehouse to the distance of detector.

Further, the three-dimensional coordinate and the equation group of three-dimensional scenic for setting up the characteristic point that the match is successful, solution side Journey group obtains the three-dimensional coordinate of the characteristic point that the match is successful, specifically includes：Set up the characteristic point abscissa and three-dimensional that the match is successful The equation group of the depth map of scene, by solving least square problem, obtains the characteristic point abscissa that the match is successful；Set up matching The equation group of the depth map of successful characteristic point ordinate and three-dimensional scenic, by solving least square problem, obtains matching into The characteristic point ordinate of work(.

Further, the equation group of the depth map of the characteristic point abscissa and three-dimensional scenic：

Further, the equation group of the depth map of the characteristic point ordinate and three-dimensional scenic is：

The present invention also provides the device that a kind of feature based point tracking is reconstructed three-dimensional scenic by focusing storehouse, described device bag Include：Module is built, the forward model of storehouse is focused on by three-dimensional scenic generation for setting up, three-dimensional scenic characteristic point is given and is focused on The relation of storehouse；Acquisition module, the characteristic point for extracting each image from the focusing storehouse, follows the trail of the focusing storehouse In the characteristic point that the match is successful；Reconstructed module, for the characteristic point that the match is successful according to acquisition module acquisition, builds The three-dimensional coordinate and the equation group of three-dimensional scenic of the vertical characteristic point that the match is successful, solve equation group and obtain the characteristic point that the match is successful Three-dimensional coordinate, reconstruct three-dimensional scenic.

Further, the forward model of the structure module foundation is：

The three-dimensional scenic characteristic point with focus on storehouse relation be：

Wherein, (S_x,S_y) it is the object plane of scene, f (S_x,S_y) it is point (S_x,S_y) corresponding irradiation level, depth (S_x,S_y) be The depth map of three-dimensional scenic, d_nTo focus on the n-th imaging lens of storehouse to object plane (S_x,S_y) distance, d '_nTo focus on storehouse I ＆ lt imaging lens to detector distance, d '₁To focus on the 1st imaging lens of storehouse to the distance of detector.

Further, the reconstructed module is specifically included：First relation unit, it is horizontal for setting up the characteristic point that the match is successful The equation group of the depth map of coordinate and three-dimensional scenic, by solving least square problem, obtains the horizontal seat of characteristic point that the match is successful Mark；Second relation unit, for the foundation characteristic point ordinate that the match is successful and the equation group of the depth map of three-dimensional scenic, passes through Least square problem is solved, the characteristic point ordinate that the match is successful is obtained；Reconfiguration unit, for according to first relation unit The characteristic point abscissa that the match is successful is obtained, and second relation unit obtains the characteristic point ordinate that the match is successful, Reconstruct three-dimensional scenic.

Further, the equation group of the depth map of the characteristic point abscissa and three-dimensional scenic：

Further, the equation group of the depth map of the characteristic point ordinate and three-dimensional scenic is：

The scheme that the present invention is provided, by the relation for building three-dimensional scenic characteristic point with focus on storehouse, follows the trail of and focuses on storehouse In the characteristic point that the match is successful, and track of the tracking feature point in storehouse is focused on, and then realize camera and shoot under visual field Three-dimensional scenic is reconstructed, and can provide accurate three-dimensional structure information for virtual reality and dimensional measurement.

Brief description of the drawings

Fig. 1 is that the feature based point tracking for providing according to a first embodiment of the present invention reconstructs three-dimensional scenic side by focusing storehouse The schematic flow sheet of method.

Fig. 2 is the schematic diagram that the generation for providing according to a first embodiment of the present invention focuses on storehouse.

Fig. 3 is the n-th imaging process schematic diagram of the focusing storehouse for providing according to a first embodiment of the present invention.

Fig. 4 is that the feature based point tracking for providing according to a second embodiment of the present invention is filled by focusing storehouse reconstruct three-dimensional scenic The structural representation put.

Specific embodiment

In the accompanying drawings, same or similar element is represented or with same or like function using same or similar label Element.Embodiments of the invention are described in detail below in conjunction with the accompanying drawings.

In the description of the invention, term " " center ", " longitudinal direction ", " transverse direction ", "front", "rear", "left", "right", " vertical ", The orientation or position relationship of the instruction such as " level ", " top ", " bottom " " interior ", " outward " are to be closed based on orientation shown in the drawings or position System, is for only for ease of the description present invention and is described with simplified, must have rather than the device or element for indicating or imply meaning Specific orientation, with specific azimuth configuration and operation, therefore it is not intended that limiting the scope of the invention.

As shown in figure 1, the feature based point tracking that the present embodiment is provided reconstructs three-dimensional scenic method bag by focusing storehouse Include：

Step 101, sets up the forward model that storehouse is focused on by three-dimensional scenic generation, obtains three-dimensional scenic characteristic point and focuses on The relation of storehouse.

In the step, it is one group of imaging sequence for focusing on different imaging planes or using different parameters to shoot to focus on storehouse Row, the focusing storehouse in present embodiment is one group focuses on the imaging sequence of different imaging planes.Present embodiment is with along light Axle mobile lens, while fixed detector complete focus on storehouse collection as a example by, illustrated to setting up forward model.By detection Object (object is made up of multiple object points) focuses on the image sequence of different imaging surfaces, these image sequence shapes in device collection scene Into focusing storehouse.It is many imaging planes imaging to the three-dimensional scenic under certain visual angle due to focusing on storehouse, it is thus possible to real Existing camera shoots the three-dimensional scenic reconstruct under the angle of visual field (FOV), can provide accurate three-dimensional for virtual reality and dimensional measurement Structural information.Certainly, the present invention can also focus on storehouse using the collection of other methods.

Consider to shoot the three-dimensional scenic under visual field in camera, (f (S can be used_x,S_y),depth(S_x,S_y)) description three-dimensional Scene.Wherein, (S_x,S_y) it is the object plane of scene.f(S_x,S_y) it is point (S_x,S_y) corresponding irradiation level.depth(S_x,S_y) it is three-dimensional The depth map of scene.

Fig. 2 is illustrated that the present embodiment forms the schematic diagram for focusing on storehouse, as shown in Fig. 2 the left side of Fig. 2 represents scene, Right side represents detector, intermediate representation lens.Fixed detector, along optical axis mobile lens, the 1st time illustrated in such as Fig. 2 into Picture ... n-th is imaged.As lens are along the movement of optical axis, the object plane of the corresponding scene of detector is also correspondingly moved along optical axis The 1st object plane ... n-th object plane that be dynamic, being illustrated in such as Fig. 2.

Fig. 3 is the schematic diagram of the n-th imaging process for focusing on storehouse, as shown in figure 3, (S_x,S_y) for scene object plane, (x, Y) it is n-th imaging surface, (S_x,S_y) dotted line Yu (x, y) between the plane where n-th imaging len.d_nTo focus on storehouse N-th imaging len is to object plane (S_x,S_y) distance, d '_nTo focus on the i ＆ lt imaging len of storehouse to the distance of detector, d′₁To focus on the 1st imaging lens of storehouse to the distance of detector.

According to lens imaging principle, and lens imaging process is expressed using point spread function, set up and generated by three-dimensional scenic Focus on storehouse forward model be：

Wherein,It is point (S_x,S_y) node expansion function on imaging surface (x, y).X and y be imaging surface abscissa and Ordinate, physical significance is object point (S_x,S_y) the position coordinates in imaging surface.

In the present embodiment, for simplified model, it is considered to which the corresponding aperture of lens stop is approximately aperture, and imaging process is regarded as Pinhole imaging system, can so cause to calculate simple with simplified model, therefore, node expansion functionIt is impulse function δ functions, then ties Close the forward model stated：

It is as follows with the relation for focusing on storehouse to obtain three-dimensional scenic characteristic point：

And then obtain three-dimensional scenic characteristic point (S_x,S_y) focus on storehouse in equation of locus.

Step 102, extracts all of feature in each image from the focusing storehouse of the forward model described by step 101 Point, the forward model for focusing on storehouse according to three-dimensional scenic generation follows the trail of the coordinate of the characteristic point that the match is successful.

In practical application, it is possible to use (Scale-invariant feature transform, Scale invariant is special for SIFT Levy conversion), ORB (ORiented Brief), SURF carry out the extraction of characteristic point and match.Matching refers to same object point and exists Image space is different in different focusing storehouse pictures, and matching is exactly to find these points.

In specific implementation, it is contemplated that SIFT keeps stronger robustness to the visual angle change of image, noise, it is possible to locate The matching problem in the case of translation, rotation, affine transformation between reason image, the present embodiment performs the figure for focusing on storehouse using SIFT As matching, the characteristic point of the image that the match is successful is obtained.

In the step, all characteristic points in each image are extracted in the focusing storehouse obtained from step 101：feature_ 1, feature_2 ..., feature_M, follow the trail of and focus on all feature point coordinates that the match is successful, characteristic point in storehouse Feature_m is in the coordinate of n-th imaging surface for focusing on storehouse

Step 103：The inverse model that three-dimensional scenic is reconstructed by focusing storehouse is set up, the three of the characteristic point that the match is successful are obtained Dimension coordinate, reconstructs three-dimensional scenic.

It refers to set up the three-dimensional of characteristic point that the match is successful to sit to set up by the inverse model of focusing storehouse reconstruct three-dimensional scenic The equation group of mark and three-dimensional scenic.

The step method is specific as follows：

To the characteristic point feature_m that the match is successful, set up on characteristic point abscissa S_xWith the depth map of three-dimensional scenic depth(S_x,S_y) equation group：

Set up on characteristic point ordinate S_yWith the depth map depth (S of three-dimensional scenic_x,S_y) equation group：

On characteristic point abscissa S_xWith the depth map depth (S of three-dimensional scenic_x,S_y) equation group, be expressed as AX=b₁, X=(S_x,depth(S_x,S_y))^TCan be by solving least square problem min | | AX | |₂Obtain S_xWith depth (S_x,S_y).It is same to close In characteristic point ordinate S_yWith depth (S_x,S_y) equation group, be expressed as AY=b₂, Y=(S_y,depth(S_x,S_y))^T Can be by solving least square problem min | | AY | |₂Obtain S_y。

In practical application, the three-dimensional scenic that generation camera is shot under visual field (FOV) includes three-dimensional point cloud atlas or three-dimensional rendering Figure.

The method also includes：The distance between measurement characteristic point, there is provided the geometry range finding of three-dimensional scenic.

The method that the feature based point tracking that the present invention is provided is reconstructed three-dimensional scenic by focusing storehouse, by building three dimensional field Scape characteristic point and the relation for focusing on storehouse, follow the trail of the characteristic point for focusing on that the match is successful in storehouse, and calculate the spy that the match is successful Three-dimensional coordinate a little is levied, and then realizes the three-dimensional scenic reconstruct that camera is shot under visual field, can be that virtual reality and geometry are surveyed Amount provides accurate three-dimensional structure information.

Referring to Fig. 4, a kind of feature based point tracking is the embodiment of the invention provides by focusing storehouse reconstruct three-dimensional scenic Device, described device includes：

Module 201 is built, the forward model of storehouse is focused on by three-dimensional scenic generation for setting up, three-dimensional scenic feature is given Point and the relation for focusing on storehouse；

Acquisition module 202, the characteristic point for extracting each image from storehouse is focused on, tracking is matched into focusing on storehouse The characteristic point of work(；

Reconstructed module 203, for obtaining the characteristic point that the match is successful according to acquisition module 202, sets up what the match is successful The three-dimensional coordinate of characteristic point and the equation group of three-dimensional scenic, solve the three-dimensional coordinate that equation group obtains the characteristic point that the match is successful, Reconstruct three-dimensional scenic.

In above-mentioned implementation method, the forward model for building the foundation of module 201 is：

The three-dimensional scenic characteristic point with focus on storehouse relation be：

Wherein, (S_x,S_y) it is the object plane of scene, f (S_x,S_y) it is point (S_x,S_y) corresponding irradiation level, depth (S_x,S_y) be The depth map of three-dimensional scenic, d_nTo focus on the n-th imaging lens of storehouse to object plane (S_x,S_y) distance, d '_nTo focus on storehouse I ＆ lt imaging lens to detector distance, d '₁To focus on the 1st imaging lens of storehouse to the distance of detector.

Acquisition module 202 extracts the feature of each image using SIFT, ORB or SURF method from the focusing storehouse Point.

Reconstructed module 203 is specifically included：

First relation unit, for the foundation characteristic point abscissa that the match is successful and the equation of the depth map of three-dimensional scenic Group, by solving least square problem, obtains the characteristic point abscissa that the match is successful；

Second relation unit, for the foundation characteristic point ordinate that the match is successful and the equation of the depth map of three-dimensional scenic Group, by solving least square problem, obtains the characteristic point ordinate that the match is successful；

Reconfiguration unit for obtaining the characteristic point abscissa that the match is successful and described according to first relation unit Second relation unit obtains the characteristic point ordinate that the match is successful, reconstructs three-dimensional scenic.

The characteristic point abscissa and the equation group of the depth map of three-dimensional scenic that first relation unit is obtained：

The characteristic point ordinate and the equation group of the depth map of three-dimensional scenic that second relation unit is obtained be：

In the embodiment of the present invention, the three-dimensional scenic of the reconstruct of reconstructed module 203 includes three-dimensional point cloud atlas or three-dimensional rendering figure.

The feature based point tracking that the present invention is provided is reconstructed the device of three-dimensional scenic by focusing storehouse, by building module structure Build out three-dimensional scenic characteristic point and focus on the relation of storehouse, follow the trail of and focus on the characteristic point that the match is successful in storehouse, and calculate Three-dimensional coordinate with successful characteristic point, and then the three-dimensional scenic reconstruct that camera is shot under visual field is realized, can be virtual existing Real and dimensional measurement provides accurate three-dimensional structure information.

It is last it is to be noted that：The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations.This The those of ordinary skill in field should be understood：Technical scheme described in foregoing embodiments can be modified, or it is right Which part technical characteristic carries out equivalent；These modifications are replaced, and the essence of appropriate technical solution is departed from this Invent the spirit and scope of each embodiment technical scheme.

Claims (10)

1. a kind of method that feature based point tracking is reconstructed three-dimensional scenic by focusing storehouse, it is characterised in that methods described includes：

The forward model that storehouse is focused on by three-dimensional scenic generation is set up, three-dimensional scenic characteristic point is obtained and is focused on the relation of storehouse；

The characteristic point of each image is extracted from the focusing storehouse, is chased after according to the forward model that three-dimensional scenic generation focuses on storehouse The coordinate of the characteristic point that the match is successful in storehouse is focused on described in track；

The three-dimensional coordinate and the equation group of three-dimensional scenic of the characteristic point that the match is successful are set up, equation group is solved and is obtained what the match is successful The three-dimensional coordinate of characteristic point, reconstructs three-dimensional scenic.

2. the method for claim 1, it is characterised in that the forward model of the foundation is：

The three-dimensional scenic characteristic point with focus on storehouse relation be：

$x=-\frac{d_n^{\′}{S}_x}{depth(S_x,S_y)-(d_n^{\′}-d_1^{\′})},y=-\frac{d_n^{\′}{S}_y}{depth(S_x,S_y)-(d_n^{\′}-d_1^{\′})};$

Wherein, (S_x,S_y) it is the object plane of scene, f (S_x,S_y) it is point (S_x,S_y) corresponding irradiation level, depth (S_x,S_y) it is three-dimensional The depth map of scene, d_nTo focus on the n-th imaging lens of storehouse to object plane (S_x,S_y) distance, d '_nTo focus on the i-th of storehouse Secondary imaging lens to detector distance, d '₁To focus on the 1st imaging lens of storehouse to the distance of detector.

3. method as claimed in claim 1 or 2, it is characterised in that the three-dimensional coordinate of the foundation characteristic point that the match is successful With the equation group of three-dimensional scenic, the three-dimensional coordinate that equation group obtains the characteristic point that the match is successful is solved, specifically included：

The equation group of the depth map of the foundation characteristic point abscissa that the match is successful and three-dimensional scenic, is asked by solving least square Topic, obtains the characteristic point abscissa that the match is successful；

The equation group of the depth map of the foundation characteristic point ordinate that the match is successful and three-dimensional scenic, is asked by solving least square Topic, obtains the characteristic point ordinate that the match is successful.

4. method as claimed in claim 3, it is characterised in that the side of the depth map of the characteristic point abscissa and three-dimensional scenic Journey group：

$\left\{\begin{array}{c}depth(S_x,S_y)x_1^m+d_1^{\′}{S}_x=0\\ depth(S_x,S_y)x_2^m+d_2^{\′}{S}_x=(d_2^{\′}-d_1^{\′})x_2^m\\ \mathrm{........................}\\ depth(S_x,S_y)x_N^m+d_N^{\′}{S}_x=(d_N^{\′}-d_1^{\′})x_N^m\end{array}\right..$

5. method as claimed in claim 3, it is characterised in that the side of the depth map of the characteristic point ordinate and three-dimensional scenic Cheng Zuwei：

$\left\{\begin{array}{c}depth(S_x,S_y)y_1^m+d_1^{\′}{S}_y=0\\ depth(S_x,S_y)y_2^m+d_2^{\′}{S}_y=(d_2^{\′}-d_1^{\′})y_2^m\\ \mathrm{........................}\\ depth(S_x,S_y)y_N^m+d_N^{\′}{S}_y=(d_N^{\′}-d_1^{\′})y_N^m\end{array}\right..$

6. a kind of feature based point tracking is reconstructed the device of three-dimensional scenic by focusing storehouse, it is characterised in that described device includes：

Module is built, the forward model of storehouse is focused on by three-dimensional scenic generation for setting up, three-dimensional scenic characteristic point is given and is gathered The relation of coke pile stack；

Acquisition module, the characteristic point for extracting each image from the focusing storehouse, follows the trail of matching in the focusing storehouse Successful characteristic point；

Reconstructed module, for the characteristic point that the match is successful according to acquisition module acquisition, sets up the spy that the match is successful The equation group of three-dimensional coordinate and three-dimensional scenic a little is levied, the three-dimensional coordinate that equation group obtains the characteristic point that the match is successful is solved, weight Structure three-dimensional scenic.

7. device as claimed in claim 6, it is characterised in that the forward model that the structure module is set up is：

$g_{d_{{}^n}^{\′}}(x,y)=\∫\∫f(S_x,S_y)\·h_{d_{{}^n}^{\′}}(x+\frac{d_n^{\′}{S}_x}{depth(S_x,S_y)-(d_n^{\′}-d_1^{\′})},y+\frac{d_n^{\′}{S}_y}{depth(S_x,S_y)-(d_n^{\′}-d_1^{\′})}){\mathrm{dS}}_x{\mathrm{dS}}_y;$

The three-dimensional scenic characteristic point with focus on storehouse relation be：

$x=-\frac{d_n^{\′}{S}_x}{depth(S_x,S_y)-(d_n^{\′}-d_1^{\′})},y=-\frac{d_n^{\′}{S}_y}{depth(S_x,S_y)-(d_n^{\′}-d_1^{\′})}$

Wherein, (S_x,S_y) it is the object plane of scene, f (S_x,S_y) it is point (S_x,S_y) corresponding irradiation level, depth (S_x,S_y) it is three-dimensional The depth map of scene, d_nTo focus on the n-th imaging lens of storehouse to object plane (S_x,S_y) distance, d '_nTo focus on the i-th of storehouse Secondary imaging lens to detector distance, d '₁To focus on the 1st imaging lens of storehouse to the distance of detector.

8. device as claimed in claims 6 or 7, it is characterised in that the reconstructed module is specifically included：

First relation unit, for the foundation characteristic point abscissa that the match is successful and the equation group of the depth map of three-dimensional scenic, leads to Solution least square problem is crossed, the characteristic point abscissa that the match is successful is obtained；

Second relation unit, for the foundation characteristic point ordinate that the match is successful and the equation group of the depth map of three-dimensional scenic, leads to Solution least square problem is crossed, the characteristic point ordinate that the match is successful is obtained；

Reconfiguration unit, for obtaining the characteristic point abscissa that the match is successful, and described second according to first relation unit Relation unit obtains the characteristic point ordinate that the match is successful, reconstructs three-dimensional scenic.

9. device as claimed in claim 8, it is characterised in that the side of the depth map of the characteristic point abscissa and three-dimensional scenic Journey group：

$\left\{\begin{array}{c}depth(S_x,S_y)x_1^m+d_1^{\′}{S}_x=0\\ depth(S_x,S_y)x_2^m+d_2^{\′}{S}_x=(d_2^{\′}-d_1^{\′})x_2^m\\ \mathrm{........................}\\ depth(S_x,S_y)x_N^m+d_N^{\′}{S}_x=(d_N^{\′}-d_1^{\′})x_N^m\end{array}\right..$

10. device as claimed in claim 8, it is characterised in that the depth map of the characteristic point ordinate and three-dimensional scenic Equation group is：

$\left\{\begin{array}{c}depth(S_x,S_y)y_1^m+d_1^{\′}{S}_y=0\\ depth(S_x,S_y)y_2^m+d_2^{\′}{S}_y=(d_2^{\′}-d_1^{\′})y_2^m\\ \mathrm{........................}\\ depth(S_x,S_y)y_N^m+d_N^{\′}{S}_y=(d_N^{\′}-d_1^{\′})y_N^m\end{array}\right..$