CN101794448A - Full automatic calibration method of master-slave camera chain - Google Patents

Abstract

The invention relates to a full automatic calibration method of a master-slave camera chain in the technical field of image processing. In the invention, an image which is automatically obtained in the process of rotating a dynamic camera is spliced into a mosaic image and then the characteristic points of the image and the image of a master camera are automatically extracted and matched, thereby obtaining a calibration result between the coordinates of the pixel of the image of a static camera and the control parameter of the dynamic camera and realizing the full automatic calibration of the master-slave camera chain. The range of the visual field of the dynamic camera is represented by the mosaic spliced image, the relationship between two different imaging planes is estimated by utilizing an SURF characteristic point, only the motion trace of the dynamic camera needs to be established beforehand, the master-slave camera chain can be automatically calibrated, the automatic calibration of the master-slave camera chain can be realized under the condition that the calibration accuracy keeps better (the error rate is 3 to 5 percent), the manpower and the time are saved, and the complexity in the calibration process is lowered.

Description

The full automatic calibration method of master-slave camera chain

Technical field

What the present invention relates to is a kind of method of technical field of image processing, specifically is a kind of full automatic calibration method of master-slave camera chain.

Background technology

At present, the public place generally has been equipped with camera monitoring system, a large amount of cameras is installed is used for large-scale overlay area.Along with the development of supervisory system, also more and more to the demand of high resolving power monitoring image.The camera supervised system of principal and subordinate is a kind of scheme that addresses this problem.In such system, the video camera that (or a plurality of) fix leads one (or many) dynamically (PTZ) video camera defocus interesting target as a leader.So just promptly can obtain macroscopical monitoring image of a broader region, can obtain the high-definition picture of target again.The demarcation of master-slave camera chain is ways of addressing this issue just.Through the camera supervised system of principal and subordinate of demarcating, to pounce on when grasping interesting target when main camera, dynamic camera can be aimed at this target automatically, thereby obtains detail pictures.

Find that through literature search pertinent literature is as follows to prior art:

1, people such as X.Zhou choose some sample points from the main camera image in " A master-slave system to acquirebiometric imagery of humans at distance (a kind of in the master slave system of obtaining personage's biometric image feature at a distance) " literary composition that The First ACM International Workshop on Video Surveillance (first international video monitoring symposial of Association for Computing Machinery) is delivered, manual mobile dynamic camera, make the dynamic camera picture centre corresponding, note sample point coordinate and controlled variable then with the sample point of choosing.The dynamic camera rotational parameters of the some correspondence on other still camera images can obtain by the parameter that has obtained is carried out interpolation.But the adjustment dynamic camera parameter that this Technology Need is manual, therefore consuming time and inconvenient large-scale application.

2, this technology was estimated the relation of main camera and dynamic camera by calculating two homography matrixs between the camera review to A.W.Senior etc. in " AcquiringMulti-Scale Images by Pan-Tilt-Zoom Control and Automatic Multi-Camera Calibration (demarcating the acquisition multi-scale image automatically by PTZ control and multiple-camera) " that The seventh IEEE Workshops on Application of ComputerVision (the 7th computer vision of international institute of electrical and electronic engineers is used symposial) is delivered in 2005.But this technology also needs manual adjustment dynamic camera parameter, same consuming time and inconvenient large-scale application.

Summary of the invention

The present invention is directed to the above-mentioned deficiency of prior art, proposed a kind of full automatic calibration method of master-slave camera chain.The present invention becomes mosaic image by rotating the image mosaic that obtains automatically in the dynamic camera process, automatically this image and main camera image are carried out the extraction and the coupling of unique point again, thereby obtain the calibration result between still camera image pixel coordinate and the dynamic camera location parameter, realize the full automatic calibration master-slave camera chain.

The present invention is achieved by the following technical solutions, comprises the steps:

Step 1, preestablish the movement locus of dynamic camera, dynamic camera rotates automatically along this movement locus, and every the time t width of cloth subimage of sampling, simultaneously by the time sequencing that collects, to each width of cloth subimage label, the parameter that dynamic camera rotates during the every width of cloth subimage of record acquisition, that is: horizontally rotate angle and vertical rotation angle, and parameter and positional information that every width of cloth subimage and corresponding label thereof and corresponding dynamic camera rotate are stored in the lump.

Step 2, after the dynamic camera rotation finishes, the N width of cloth subimage that obtains is carried out arranged in groups by the camera site to be handled, obtain K group set of sub-images, two adjacent in every group of set of sub-images width of cloth subimages are carried out feature point extraction and matching treatment, it is right to obtain matching characteristic point, and then obtains the transform matrix M between two adjacent in same group of set of sub-images width of cloth subimages.

It is that dynamic camera with the subimage correspondence horizontally rotates differential seat angle and is divided into one group of set of sub-images less than the subimage of threshold value that described arranged in groups is handled, thereby obtain K group set of sub-images, and the subimage in every group is arranged according to the vertical rotation angular dimension of the dynamic camera of correspondence.

Described feature point extraction and matching treatment are finished by the SURF method, are promptly extracted the unique point of adjacent two width of cloth subimages in every group of set of sub-images by quick Hessian extraction apparatus, obtain the SURF proper vector of unique point again.

Transform matrix M in the described same group of set of sub-images between adjacent two width of cloth subimages obtains by the RANSAC method, specifically:

$\left[\begin{array}{c}{x}^{\′}\\ y^{\′}\\ z^{\′}\end{array}\right]=\left[\begin{array}{ccc}{m}_0& m_1& m_2\\ m_3& m_4& m_5\\ m_6& m_7& m_8\end{array}\right]*\left[\begin{array}{c}x\\ y\\ z\end{array}\right]\mathrm{or}{u}^{\′}=\mathrm{Mu},$

Wherein: (x ', y ', z ') and (x, y z) are the coordinate of same unique point in two adjacent width of cloth subimage coordinate systems respectively, the subimage coordinate system with the image upper left corner be initial point, x axle horizontal to the right for just, the y axle is being vertically downward for just to set up.

Step 3 according to the transformation matrix that obtains, is carried out image mosaic to every group of set of sub-images and is handled, thereby every group of set of sub-images is spliced into a width of cloth row stitching image, obtains K width of cloth row stitching image altogether.

Described image mosaic is handled, specifically: with the benchmark image of subimage middle in every group of set of sub-images for this group set of sub-images, other subimages in this group set of sub-images are converted in the benchmark image plane coordinate system according to the transformation matrix that obtains, the pixel in the coincidence zone of image is carried out the heavy gray scale splicing of cum rights to be handled, gray scale to the pixel in the non-coincidence zone of image remains unchanged, thereby every group of set of sub-images is spliced into a width of cloth row stitching image.

The gray scale splicing that described cum rights is heavy is handled, specifically:

$f_{\mathrm{res}}\left(P\right)=\frac{\underset{i=1}{\overset{W}{\mathrm{\Σ}}}{f}_i\left(P\right)d_i^n}{\underset{i=1}{\overset{W}{\mathrm{\Σ}}}{d}_i^n},$

Wherein: f _Res(P) be the pixel value that splicing back P is ordered, f _i(p) be the pixel value that P is ordered in the i width of cloth image, W is for participating in spliced image quantity, and n is a constant, d _iBe the bee-line of P point to the i width of cloth image boundary that participates in splicing.

Step 4 as an image collection, is carried out K width of cloth row stitching image successively feature point extraction, matching treatment and image mosaic to it and is handled, thereby obtain a width of cloth mosaic stitching image.

Step 5, gather a width of cloth still image by still camera, this still image and mosaic stitching image are carried out feature point extraction and matching treatment, and utilize the further search characteristics point of polar curve geometrical principle, the matching characteristic point that obtains between still image and mosaic stitching image is right, and the unique point of coupling is evenly distributed on still image and the mosaic stitching image as far as possible.

Described polar curve geometrical principle is: the corresponding point of the point of piece image on another width of cloth image are positioned on the corresponding polar curve.

Described further search characteristics point is that the unique point of still image is positioned on the corresponding polar curve in the corresponding point on the mosaic stitching image, thereby increases the quantity of unique point.

Step 6 is carried out the global calibration processing to still image and mosaic stitching image, thereby obtains the mapping relations between still image and mosaic stitching image, i.e. demarcation between principal and subordinate's video camera relation.

Described global calibration is handled, and comprises that step is:

1) any point P in still image _s(x _s, y _s) in the adjacent domain, search still image unique point, N _R(P _s) be range points P _sDistance is less than or equal to the set of all still image unique points formations of R, that is:

$N_R\left(P_s\right)=\{(P_s^1,r_1),(P_s^2,r_2),...\},$ Wherein: P _s ⁱIt is distance P _sBe r _iUnique point;

2) find N _R(P _s) each still image unique point P in the set _s ⁱCorresponding mosaic image unique point P _d ⁱ, obtain comprising P _d ⁱSubimage, and then from comprising P _d ⁱSubimage in select P _d ⁱNear the width of cloth subimage I at its subimage center, place _r ⁱ

3) search subimage I _r ⁱLabel in all set of sub-images obtains dynamic camera and is taking subimage I _r ⁱThe time residing location parameter S _i

4) to the location parameter S of dynamic camera _iCarry out interpolation processing, obtain any point P in the still image _s(x _s, y _s) with the corresponding relation of the location parameter of dynamic camera, specifically:

S＝S ₁*f ₁(r ₁，f ₂，...r _n)+S ₂*f ₂(r ₁，r ₂，...r _n)+...+S _n*f _n(r ₁，r ₂，...r _n)，

Wherein: f _iBe interpolating function.

Compared with prior art, the invention has the beneficial effects as follows: the field range of representing dynamic camera by the mosaic stitching image, utilized the relation between two different imaging planes of SURF unique point estimation, only need to formulate in advance the movement locus of good dynamic video camera, just can calibrate this master-slave camera chain automatically, can keep under the stated accuracy good conditions (error rate is 3%-5%), realize the automatic demarcation of master-slave camera chain, save manpower and time, reduced the complexity in the calibration process.

Description of drawings

Fig. 1 is an embodiment dynamic camera rotary motion trace synoptic diagram;

Wherein: (a) be the position of the dynamic camera when taking 64 width of cloth subimages; (b) be the plurality of sub image that dynamic camera is taken.

Fig. 2 is the synoptic diagram that the heavy gray scale splicing of cum rights is handled.

Fig. 3 is the mosaic stitching image that embodiment obtains.

Fig. 4 is the matching characteristic point that embodiment obtains;

Wherein: (a) be the simple matching characteristic point that utilizes the SURF method to obtain; (b) be the matching characteristic point that further utilizes the polar curve geometrical principle to obtain.

Fig. 5 is the synoptic diagram that global calibration is handled.

Fig. 6 is the calibration result of embodiment;

Wherein: (a) piece image of still camera shooting; (b) be the image of taking with respect to the dynamic camera of (a); (c) be another width of cloth image that still camera is taken; (d) be the image of taking with respect to the still camera of (c).

Fig. 7 is an embodiment calibrated error rate synoptic diagram;

Wherein: (a) be the error rate synoptic diagram of embodiment directions X; (b) be the error rate synoptic diagram of embodiment Y direction.

Embodiment

Below in conjunction with accompanying drawing embodiments of the invention are elaborated: present embodiment has provided detailed embodiment and process being to implement under the prerequisite with the technical solution of the present invention, but protection scope of the present invention is not limited to following embodiment.

Embodiment

In the present embodiment certain an indoor still camera and dynamic camera that is installed on the different location is demarcated.The resolution of still camera and dynamic camera is 320*240.

Present embodiment comprises the steps:

Step 1, the movement locus that preestablishes dynamic camera is " it " font, dynamic camera rotates automatically along given movement locus, and every the time t width of cloth subimage of sampling, simultaneously by the time sequencing of gathering, to each width of cloth subimage label, the parameter that dynamic camera rotates when writing down each images acquired, that is: horizontally rotate angle and vertical rotation angle, and parameter and positional information that every width of cloth subimage and corresponding label thereof and corresponding dynamic camera rotate are stored in the lump.

The movement locus of dynamic camera is shown in arrow among Fig. 1 (a) in the present embodiment, and the width of cloth subimage of sampling at Fig. 2 orbicular spot place respectively obtains 64 width of cloth subimages altogether, and the plurality of sub image that obtains is shown in Fig. 1 (b).

Step 2, after the dynamic camera rotation finishes, the N width of cloth subimage that obtains is carried out arranged in groups by the camera site to be handled, obtain K group set of sub-images, adjacent two width of cloth subimages in every group of set of sub-images are carried out feature point extraction and matching treatment, it is right to obtain matching characteristic point, and then obtains the transform matrix M between two adjacent in same group of set of sub-images width of cloth subimages.

It is to be divided into 8 groups from small to large with 64 width of cloth subimages are horizontally rotated angle by video camera that described arranged in groups is handled, and every group comprises 8 width of cloth subimages, and with vertical rotation angle ascending sort of the subimage in every group according to the dynamic camera of correspondence.

Described feature point extraction and matching treatment are finished by the SURF method, are promptly extracted the unique point of adjacent two width of cloth subimages in every group of set of sub-images by quick Hessian extraction apparatus, obtain the SURF proper vector of unique point again.

Transform matrix M in the described same group of set of sub-images between adjacent two width of cloth subimages obtains by the RANSAC method, specifically:

$\left[\begin{array}{c}{x}^{\′}\\ y^{\′}\\ z^{\′}\end{array}\right]=\left[\begin{array}{ccc}{m}_0& m_1& m_2\\ m_3& m_4& m_5\\ m_6& m_7& m_8\end{array}\right]*\left[\begin{array}{c}x\\ y\\ z\end{array}\right]\mathrm{or}{u}^{\′}=\mathrm{Mu},$

Wherein: (x ', y ', z ') and (x, y z) are the coordinate of same unique point in two adjacent width of cloth subimage coordinate systems respectively, the subimage coordinate system with the image upper left corner be initial point, x axle horizontal to the right for just, the y axle is vertically downward for just setting up.

Step 3 according to the transformation matrix that obtains, is carried out image mosaic to every group of set of sub-images and is handled, thereby every group of set of sub-images is spliced into a width of cloth row stitching image, obtains 8 width of cloth row stitching images altogether.

Described image mosaic is handled, specifically: with the benchmark image of subimage middle in every group of set of sub-images (present embodiment is the 5th width of cloth subimage) for this group set of sub-images, other subimages in this group set of sub-images are converted in the benchmark image plane coordinate system according to the transformation matrix that obtains, the pixel in the coincidence zone of image is carried out the heavy gray scale splicing of cum rights to be handled, gray scale to the pixel in the non-coincidence zone of image remains unchanged, thereby every group of set of sub-images is spliced into a width of cloth row stitching image.

As shown in Figure 2, the gray scale splicing that described cum rights is heavy is handled, specifically:

$f_{\mathrm{res}}\left(P\right)=\frac{\underset{i=1}{\overset{8}{\mathrm{\Σ}}}{f}_i\left(P\right)d_i^n}{\underset{i=1}{\overset{8}{\mathrm{\Σ}}}{d}_i^n},$

Wherein: f _Res(P) be the pixel value that splicing back P is ordered, f _i(p) be the pixel value that P is ordered in the i width of cloth image, n is a constant, d _iBe the bee-line of P point to the i width of cloth image boundary that participates in splicing.

N is 3 in the present embodiment.

Step 4 as an image collection, is carried out 8 width of cloth row stitching images successively feature point extraction, matching treatment and image mosaic according to the method for step 2 and step 3 to it and is handled, thereby obtain a width of cloth mosaic stitching image.

The mosaic stitching image that present embodiment obtains as shown in Figure 3, white line bar wherein is the border of every width of cloth subimage.

Step 5, gather a width of cloth still image by still camera, this still image and mosaic stitching image are carried out feature point extraction and matching treatment, and utilize the further search characteristics point of polar curve geometrical principle, the matching characteristic point that obtains between still image and mosaic stitching image is right, and the unique point of coupling is evenly distributed on two width of cloth images as far as possible.

Described polar curve geometrical principle is: the corresponding point of the point of piece image on another width of cloth image are positioned on the corresponding polar curve.

Present embodiment is simple utilizes matching characteristic point that the SURF method obtains shown in Fig. 4 (a), further utilizes matching characteristic point that the polar curve geometrical principle obtains shown in Fig. 4 (b).

Step 6 is carried out the global calibration processing to still image and mosaic stitching image, thereby obtains the mapping relations between still image and mosaic stitching image, i.e. demarcation between principal and subordinate's video camera relation.

As shown in Figure 5, described global calibration is handled, and comprises that step is:

1) any point P in still image _s(x _s, y _s) in the adjacent domain, search still image SURF unique point, N _R(P _s) be range points P _sDistance is less than or equal to the set of all still image SURF unique points formations of R:

$N_R\left(P_s\right)=\{(P_s^1,r_1),(P_s^2,r_2),...\},$ Wherein: P _s ⁱIt is distance P _sBe r _iUnique point;

2) find N _R(P _s) each still image SURF unique point P in the set _s ⁱCorresponding mosaic image unique point P _d ⁱ, obtain comprising P _d ⁱSubimage, and then from comprising P _d ⁱSubimage in select P _d ⁱNear the subimage I at subimage center _r ⁱ

3) search subimage I _r ⁱLabel in all set of sub-images obtains dynamic camera and is taking subimage I _r ⁱThe time residing position S _i(α _i, β _i, Z _i) i=1,2,3...;

4) to S _i(α _i, β _i, Z _i) carry out interpolation processing, obtain any point P in the still image _s(x _s, y _s) with the corresponding relation of the location parameter of dynamic camera, specifically:

S＝S ₁*f ₁(r ₁，r ₂，...r _n)+S ₂*f ₂(r ₁，r ₂，...r _n)+...+S _n*f _n(r ₁，r ₂，...r _n)，

Wherein: f _iFor inserting the letter number.

Interpolating function is specially in the present embodiment: $f_i(r_1,r_2,...r_n)=r_i/\sqrt{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{r}_i^2}.$

Use still camera to take piece image shown in Fig. 6 (a), use the present embodiment method that the interesting target (people) among Fig. 6 (a) is carried out the demarcation of dynamic camera, thereby the image that the rotation dynamic camera obtains is shown in Fig. 6 (b); Similarly, still camera is taken another width of cloth image shown in Fig. 6 (c), and the image that obtains dynamic camera accordingly is shown in Fig. 6 (d).

Present embodiment is chosen any 8 points in the still image, calculate the anglec of rotation of dynamic camera correspondence by the relation of the demarcation between the principal and subordinate's video camera that obtains, behind the rotation dynamic camera, calculate the dynamic camera picture centre point corresponding at X with selected point, distance on the Y direction, thereby according to X, the ratio of the distance on the Y direction and dynamic camera resolution, obtain the error rate of principal and subordinate's camera calibration, the error rate of the directions X that present embodiment obtains is shown in Fig. 7 (a), the error rate of Y direction is shown in Fig. 7 (b), and the present embodiment method is simple and stated accuracy is high as shown in Figure 7.

Claims (9)

1. the full automatic calibration method of a master-slave camera chain is characterized in that, may further comprise the steps:

Step 1, preestablish the movement locus of dynamic camera, dynamic camera rotates automatically along this movement locus, and every the time t width of cloth subimage of sampling, simultaneously by the time sequencing that collects, to each width of cloth subimage label, the parameter that dynamic camera rotates during the every width of cloth subimage of record acquisition,, and parameter and positional information that every width of cloth subimage and corresponding label thereof and corresponding dynamic camera rotate stored;

Step 2, after the dynamic camera rotation finishes, the N width of cloth subimage that obtains is carried out arranged in groups by the camera site to be handled, obtain K group set of sub-images, two adjacent in every group of set of sub-images width of cloth subimages are carried out feature point extraction and matching treatment, it is right to obtain matching characteristic point, and then obtains the transform matrix M between two adjacent in same group of set of sub-images width of cloth subimages;

Step 3 according to the transformation matrix that obtains, is carried out image mosaic to every group of set of sub-images and is handled, thereby every group of set of sub-images is spliced into a width of cloth row stitching image, obtains K width of cloth row stitching image altogether;

Step 4 as an image collection, is carried out K width of cloth row stitching image successively feature point extraction, matching treatment and image mosaic to it and is handled, thereby obtain a width of cloth mosaic stitching image;

Step 5, gather a width of cloth still image by still camera, this still image and mosaic stitching image are carried out feature point extraction and matching treatment, and utilize the further search characteristics point of polar curve geometrical principle, the matching characteristic point that obtains between still image and mosaic stitching image is right, and the unique point of coupling is evenly distributed on still image and the mosaic stitching image as far as possible;

Step 6 is carried out the global calibration processing to still image and mosaic stitching image, thereby obtains the mapping relations between still image and mosaic stitching image, i.e. demarcation between principal and subordinate's video camera relation.

2. the full automatic calibration method of master-slave camera chain according to claim 1 is characterized in that, the parameter that the dynamic camera described in the step 1 rotates is meant: horizontally rotate angle and vertical rotation angle.

3. the full automatic calibration method of master-slave camera chain according to claim 1, it is characterized in that, it is that dynamic camera with the subimage correspondence horizontally rotates differential seat angle and is divided into one group of set of sub-images less than the subimage of threshold value that arranged in groups described in the step 2 is handled, thereby obtain K group set of sub-images, and the subimage in every group is arranged according to the vertical rotation angular dimension of the dynamic camera of correspondence.

4. the full automatic calibration method of master-slave camera chain according to claim 1, it is characterized in that, feature point extraction described in the step 2 and matching treatment are to be extracted the unique point of adjacent two width of cloth subimages in every group of set of sub-images by quick Hessian extraction apparatus, obtain the SURF proper vector of unique point again.

5. the full automatic calibration method of master-slave camera chain according to claim 1 is characterized in that, the transform matrix M in the same group of set of sub-images described in the step 2 between adjacent two width of cloth subimages is:

$\left[\begin{array}{c}{x}^{\′}\\ y^{\′}\\ z^{\′}\end{array}\right]=\left[\begin{array}{ccc}{m}_0& m_1& m_2\\ m_3& m_4& m_5\\ m_6& m_7& m_8\end{array}\right]*\left[\begin{array}{c}x\\ y\\ z\end{array}\right]\mathrm{or}{u}^{\′}=\mathrm{Mu},$

Wherein: (x ', y ', z ') and (x, y z) are the coordinate of same unique point in two adjacent width of cloth subimage coordinate systems respectively, the subimage coordinate system with the image upper left corner be initial point, x axle horizontal to the right for just, the y axle is being vertically downward for just to set up.

6. the full automatic calibration method of master-slave camera chain according to claim 1, it is characterized in that, image mosaic described in the step 3 is handled: with the benchmark image of subimage middle in every group of set of sub-images for this group set of sub-images, other subimages in this group set of sub-images are converted in the benchmark image plane coordinate system according to the transformation matrix that obtains, the pixel in the coincidence zone of image is carried out the heavy gray scale splicing of cum rights to be handled, gray scale to the pixel in the non-coincidence zone of image remains unchanged, thereby every group of set of sub-images is spliced into a width of cloth row stitching image.

7. the full automatic calibration method of master-slave camera chain according to claim 6 is characterized in that, the gray scale splicing that described cum rights is heavy is handled and is:

$f_{\mathrm{res}}\left(P\right)=\frac{\underset{i=1}{\overset{W}{\mathrm{\Σ}}}{f}_i\left(P\right)d_i^n}{\underset{i=1}{\overset{W}{\mathrm{\Σ}}}{d}_i^n},$

Wherein: f _Res(P) be the pixel value that splicing back P is ordered, f _i(p) be the pixel value that P is ordered in the i width of cloth image, W is for participating in spliced image quantity, and n is a constant, d _iBe the bee-line of P point to the i width of cloth image boundary that participates in splicing.

8. the full automatic calibration method of master-slave camera chain according to claim 1, it is characterized in that, further search characteristics point described in the step 5 is that the unique point of still image is positioned on the corresponding polar curve in the corresponding point on the mosaic stitching image, thereby increases the quantity of unique point.

9. the full automatic calibration method of master-slave camera chain according to claim 1 is characterized in that, the global calibration described in the step 6 is handled, and comprises that step is:

1) any point P in still image _s(x _s, y _s) in the adjacent domain, search still image unique point, N _R(P _s) be range points P _sDistance is less than or equal to the set of all still image unique points formations of R, that is:

Wherein: P _s ⁱIt is distance P _sBe r _iUnique point;

2) find N _R(P _s) each still image unique point P in the set _s ⁱCorresponding mosaic image unique point P _d ⁱ, obtain comprising P _d ⁱSubimage, and then from comprising P _d ⁱSubimage in select P _d ⁱNear the width of cloth subimage I at its subimage center, place _r ⁱ

3) search subimage I _r ⁱLabel in all set of sub-images obtains dynamic camera and is taking subimage I _r ⁱThe time residing location parameter S _i

4) to the location parameter S of dynamic camera _iCarry out interpolation processing, obtain any point P in the still image _s(x _s, y _s) with the corresponding relation of the location parameter of dynamic camera be:

S＝S ₁*f ₁(r ₁，r ₂，...r _n)+S ₂*f ₂(r ₁，r ₂，...r _n)+...+S _n*f _n(r ₁，r ₂，...r _n)，

Wherein: f _iBe interpolating function.

Images