CN103363960A - Large-breadth digital aerial surveying camera image splicing method based on DLT (Direct Linear Transformation) coefficients - Google Patents

Abstract

The invention discloses a large-breadth digital aerial surveying camera image splicing method based on DLT (Direct Linear Transformation) coefficients. The method is mainly used for splicing sub-images of a large-breadth digital aerial surveying camera into a virtual image and comprises the following steps: exposing at the same moment to obtain four sub-images; carrying out connection point matching on the four sub-images in a cross-shaped overlapping region; calculating to obtain an outer orientation element of each sub-image according to connection points among the sub-images and an initial platform checking parameter of each sub-image; determining a relative position relation of four cameras according to the outer orientation elements of the four sub-images which are exposed at the same moment; calculating a DLT platform checking parameter according to the relative position relation; generating a spliced virtual image by the sub-images according to the DLT platform checking parameter. According to the large-breadth digital aerial surveying camera image splicing method based on the DLT coefficients, disclosed by the invention, the complicated data operation processing is avoided, so that the splicing method is simple and is good for the image splicing method to rapidly process.

Description

A kind of based on DLT coefficient large format digital air photo instrument image splicing method

Technical field

The present invention relates to the photogrammetric measurement technical field, relate in particular to a kind of based on DLT coefficient large format digital air photo instrument image splicing method.

Background technology

Practical digital air photo instrument requires to possess simultaneously high resolving power and large format covers.Yet in the certain situation of film size, the spatial resolution of image and ground are inverse relation between covering, and the effective way that solves the two contradiction is to increase the film size size.The digital aerial surveying camera film size of the statuette width of cloth is little, and data volume is large, has increased the workload of aftertreatment, and base-height ratio is little, has reduced vertical accuracy.Multiaspect battle array digital air photo instrument is compared with the single face battle array, take the photograph (flying height, degree of overlapping is identical) under the condition in same boat, it is large that film size becomes, baseline elongates, base-height ratio increases, and the vertical accuracy large format improves, and can improve the work unit work efficiency, reduce workload, shorten working hours, saved huge financial resources, material resources and manpower, played huge impetus for photogrammetric development.

In the prior art, in June, 2007, the Research Team headed by the woods academician of Liu elder generation has been released the digital aerial surveying camera SWDC of the large film size series of products of China's independent development first.The digital aerial surveying camera of the large film size of SWDC adopts directly with traditional platform calibration parameter by being spliced to form virtual image as the plane transformation method.Existing platform calibration parameter is by need to calculating rotation matrix as plane transformation, so operand is large, is unfavorable for that data splicing on a large scale processes.

Summary of the invention

The object of the present invention is to provide a kind of based on DLT coefficient large format digital air photo instrument image splicing method, to address the above problem.

In order to achieve the above object, technical scheme of the present invention is achieved in that

A kind of based on DLT coefficient large format digital air photo instrument image splicing method, comprise the steps:

Exposure obtains four sub-images to synchronization;

In the cruciform overlay region, carry out the tie point coupling of four sub-images, according to the tie point between sub-image, and the platform calibration initial value of each sub-image, calculate the elements of exterior orientation of every sub-image;

According to the elements of exterior orientation of four sub-images of synchronization exposure, determine the relative position relation of four cameras, calculate DLT platform calibration parameter according to described relative position relation;

Generate spliced virtual image according to described DLT platform calibration parameter from sub-image.Compared with prior art, the advantage of the embodiment of the invention is:

Provided by the invention a kind of based on DLT coefficient large format digital air photo instrument image splicing method, comprise the steps:

At first, exposure obtains four sub-images to synchronization, and this is the element task that digitized video is obtained in realization, and above-mentioned corresponding sub-image is preserved.

Then, in the cruciform overlay region, carry out the tie point coupling of a plurality of sub-images, according to the tie point between sub-image, and the platform calibration initial value of each sub-image, calculate the elements of exterior orientation of every sub-image.Particularly, utilize above-mentioned steps can realize calculating the data such as elements of exterior orientation, elements of interior orientation, can provide the data basis for the follow-up calibration platform model of setting up like this.

Then, according to the elements of exterior orientation of four sub-images of synchronization exposure, determine the relative position relation of four cameras, calculate DLT(direct linear transformation solution according to described relative position relation) platform calibration parameter; Here step its essence is the foundation of carrying out mathematical model and finds the solution.Set up the mathematical model of calibration, employing is based on the mathematical model of space resection, take collinearity equation (being that object point, picture point and photo centre are positioned on the straight line) as the basis, solution is tried to achieve out the DLT coefficient (be each of single camera as the plane to the DLT conversion coefficient on virtual representation plane) of every camera.

At last, generate spliced virtual image according to described DLT platform calibration parameter from sub-image.

Provided by the invention based on DLT coefficient large format digital air photo instrument image splicing method, introducing by the DLT coefficient, platform calibration parameter can be converted to the DLT coefficient, thereby set up one camera as the relation of plane between the virtual representation plane, only need to calculate simple multiplication and division plus and minus calculation, namely one camera can be converted to the virtual representation planimetric coordinates as planimetric coordinates, finish image joint, avoided complicated data operation to process, therefore this joining method is simpler, more is conducive to the image splicing method fast processing.

Description of drawings

The schematic flow sheet based on DLT coefficient large format digital air photo instrument image splicing method that Fig. 1 provides for the embodiment of the invention.

Embodiment

Also by reference to the accompanying drawings the present invention is described in further detail below by specific embodiment.

Referring to Fig. 1, it is a kind of based on DLT coefficient large format digital air photo instrument image splicing method that the embodiment of the invention provides, and comprises the steps:

Step S100, synchronization exposure is obtained four sub-images;

Step S200, in the cruciform overlay region, carry out the tie point coupling of four sub-images, according to the tie point between sub-image, and the platform calibration initial value of each sub-image, calculate the elements of exterior orientation of every sub-image;

Step S300, according to the elements of exterior orientation of four sub-images of synchronization exposure, determine the relative position relation of four cameras, calculate DLT platform calibration parameter according to described relative position relation;

Step S400, generate spliced virtual image according to described DLT platform calibration parameter from sub-image.

In embodiments of the present invention, need to prove that the enforcement of above-mentioned each step is finished by application system, this application system comprises calibration platform, the digital picture splicing processor (comprising software section) that four cameras consist of.This processor can resolve elements of relative orientation automatically, sets up the area adjustment model and carries out resolving of aerotriangulation, analytical photogrammetry.

For large format multiaspect battle array digital camera, obtain higher calibration precision, should adopt the method for controlling filed calibration, can reach the purpose of obtaining sub-image elements of exterior orientation and other distortion factors by calibration.Can determine the error (namely comprising optical parallax, machine error and electricity error etc.) of digital camera by calibration, thereby carry out follow-up processing.

The digital air photo instrument image splicing method that the embodiment of the invention provides, at first, exposure obtains four sub-images to (same subject) synchronization, and this is the element task that digitized video is obtained in realization, and above-mentioned corresponding sub-image is preserved.

Then, in the cruciform overlay region, carry out the tie point coupling of four sub-images, according to the tie point between sub-image, and the platform calibration initial value of each sub-image, calculate the elements of exterior orientation of every sub-image.Particularly, utilize above-mentioned steps can realize calculating the data such as elements of exterior orientation, elements of interior orientation, can provide the data basis for the follow-up calibration platform model of setting up like this.

Then, elements of exterior orientation according to four sub-images of synchronization exposure, determine the relative position relation of four cameras, calculate DLT((direct linear transformation, DLT according to described relative position relation) direct linear transformation's solution) platform calibration parameter; Here step its essence is the foundation of carrying out mathematical model and finds the solution.Set up the mathematical model of calibration, employing is based on the mathematical model of space resection, take collinearity equation (being that object point, picture point and photo centre are positioned on the straight line) as the basis, solution is tried to achieve out the DLT coefficient (be each of single camera as the plane to the DLT conversion coefficient on virtual representation plane) of every camera.At last, generate spliced virtual image according to described DLT platform calibration parameter from sub-image.

The embodiment of the invention provide based on DLT coefficient large format digital air photo instrument image splicing method, introducing by the DLT coefficient, platform calibration parameter can be converted to the DLT coefficient, thereby set up one camera as the relation of plane between the virtual representation plane, only need to calculate simple multiplication and division plus and minus calculation, namely one camera can be converted to the virtual representation planimetric coordinates as planimetric coordinates, finish image joint.Avoided complex calculations processing in the prior art, so this joining method is simpler, is conducive to the image splicing method fast processing.

The below is elaborated to above-mentioned each step based on DLT coefficient large format digital air photo instrument image splicing method that the embodiment of the invention provides:

Particularly, in step S300, the elements of exterior orientation of described four sub-images according to synchronization exposure is determined the relative position relation of four cameras to calculate DLT platform calibration parameter according to described relative position relation, comprises the steps:

Step S310, calculate every camera in four cameras respectively each as the DLT conversion coefficient of plane to the virtual representation plane;

Step S320, according to the DLT conversion coefficient of four cameras, determine the relative position relation of four cameras, calculate DLT platform calibration parameter according to the relative position relation of four cameras.

Preferably, in step S310, described each that calculate respectively every camera in four cameras comprises the steps: as the DLT conversion coefficient of plane to the virtual representation plane

Step S311, calculate single camera to the perspective transform of surface level;

If the elements of exterior orientation of camera is X _S1, Y _S1, Z _S1, ω ₁, κ ₁, elements of interior orientation is x ₀₁, y ₀₁, f ₁, be x as coordinate ₁, y ₁, corresponding surface level is (X, Y, H), rotation matrix is R1, obtains formula one according to collinearity equation:

$\left(\begin{array}{c}{x}_1-x_{01}\\ y_1-y_{01}\\ -f_1\end{array}\right)={\mathrm{\λ}}_1{R}_1\left(\begin{array}{c}X-X_{s1}\\ Y-Y_{s1}\\ H-Z_{s1}\end{array}\right);$

Rotation matrix: $R_1=\left[\begin{array}{ccc}{a}_1& a_2& a_3\\ b_1& b_2& b_3\\ c_1& c_2& c_3\end{array}\right];$

During coordinates computed, establish k ₀₁=a ₃/ (H-Z _S1), k ₀₂=b ₃/ (H-Z _S1),

k ₀₃=-(a ₃x ₀₁+b ₃y ₀₁+c ₃f ₁)/(H-Z _s1)；

λ ₁=(a ₃(x ₁-x ₀₁)+b ₃(y ₁-y ₀₁)-c ₃f ₁)/(H-Z _s1)=k ₀₁x+k ₀₂y+k ₀₃；

Wherein, need to prove k ₀₁, k ₀₂, k ₀₃Coefficient required hypothesis numerical value when calculating, do not have any practical significance, in follow-up computation process, will be used.

The perspective transform that step S312, calculated level face arrive the virtual representation plane;

If the elements of exterior orientation of virtual camera is X _S0, Y _S0, Z _S0,

ω ₀, κ ₀, focal length is f, is x, y as coordinate, and corresponding surface level is (X, Y, 0), and rotation matrix is R' ₁

Wherein, described elements of exterior orientation all is the numerical value after processing through centralization:

X _s0=(X _s1+X _s2)/2

Y _s0=(Y _s1+Y _s2)/2

Z _s0=(Z _s1+Z _s2)/2

ω _s0=(ω _s1+ω _s2)/2

κ _s0=(κ _s1+κ _s2)/2

Obtain formula two according to collinearity equation:

$\left(\begin{array}{c}x\\ y\\ -f\end{array}\right)={\mathrm{\λ}}_1^{\′}{R}_1^{\′}\left(\begin{array}{c}X-X_{s0}\\ Y-Y_{s0}\\ H-Z_{s0}\end{array}\right);$

Wherein, rotation matrix is ${R_1}^{\′}=\left[\begin{array}{ccc}{c}_{00}& c_{01}& c_{02}\\ c_{10}& c_{11}& c_{12}\\ c_{20}& c_{21}& c_{22}\end{array}\right];$

Obtain thus:

$\left(\begin{array}{c}X-X_{s0}\\ Y-Y_{s0}\\ H-Z_{s0}\end{array}\right)=R_1^{{\′}^{-1}}/{\text{\λ}}_1^{\′}\left(\begin{array}{c}x\\ y\\ -f\end{array}\right);$

Obtain perspective transform coefficient: λ ' ₁=(C ₂₀X+C ₂₁Y-C ₂₂F)/(H-Z _S0);

Calculating ${\mathrm{\λ}}_1{R}_1\left(\begin{array}{c}{X}_{s1}-X_{s0}\\ Y_{s1}-Y_{s0}\\ Z_{s1}-Z_{s0}\end{array}\right)$ The time, establish it and be ${\mathrm{\λ}}_1\left(\begin{array}{c}{k}_{11}\\ k_{12}\\ k_{13}\end{array}\right),$ Then calculate:

k ₁₁=C ₀₀(X _s1-X _s0)+C ₀₁(Y _s1-Y _s0)+C ₀₂(Z _s1-Z _s0)；

k ₁₂=C ₁₀(X _s1-X _s0)+C ₁₁(Y _s1-Y _s0)+C ₁₂(Z _s1-Z _s0)；

k ₁₃=C ₂₀(X _s1-X _s0)+C ₂₁(Y _s1-Y _s0)+C ₂₂(Z _s1-Z _s0)

；

Step S313, according to single camera to the perspective transform that the perspective transform coefficient of surface level and surface level arrive the virtual representation plane, ask for single camera to the DLT conversion coefficient of the Relative Transformation of virtual camera;

Can be obtained by formula one:

Formula three $\left(\begin{array}{c}X-X_{s1}\\ Y-Y_{s1}\\ H-Z_{s1}\end{array}\right)={R_1}^{-1}/{\mathrm{\λ}}_1\left(\begin{array}{c}{x}_1-x_{01}\\ y_1-y_{01}\\ -f_1\end{array}\right);$

Obtain thus formula four:

$\left(\begin{array}{c}X\\ Y\\ H\end{array}\right)={R_1}^{-1}/{\mathrm{\λ}}_1\left(\begin{array}{c}{x}_1-x_{01}\\ y_1-y_{01}\\ {-f}_1\end{array}\right)+\left(\begin{array}{c}{X}_{s1}\\ Y_{s1}\\ Z_{s1}\end{array}\right);$

Formula four substitution formula two are obtained formula five:

Formula five is:

$\left(\begin{array}{c}x\\ y\\ -f\end{array}\right)={\mathrm{\λ}}_1^{\′}/{\mathrm{\λ}}_1{R}_1^{\′}{R_1}^{-1}(\left(\begin{array}{c}{x}_1-x_{01}\\ y_1-y_{01}\\ {-f}_1\end{array}\right)+{\mathrm{\λ}}_1{R}_1\left(\begin{array}{c}{X}_{s1}-X_{s0}\\ Y_{s1}-Y_{s0}\\ Z_{s1}-Z_{s0}\end{array}\right));$

Wherein: ${R_1}^{\′}{R}_1^{-1}=\left[\begin{array}{ccc}{M}_{00}& M_{01}& M_{02}\\ M_{10}& M_{11}& M_{12}\\ M_{20}& M_{21}& M_{22}\end{array}\right];$

Calculate $\left(\begin{array}{c}x\\ y\\ -f\end{array}\right)$ The time, establish it and be $\left(\begin{array}{c}{k}_{20}x+k_{21}y+k_{22}\\ k_{23}x+k_{24}y+k_{25}\\ k_{26}+k_{27}y+k_{28}\end{array}\right),$ Then calculate nine coefficients:

k ₂₀=M ₀₀(1+k ₁₁k ₀₁)+M ₀₁k ₀₁k ₁₂+M ₀₂k ₀₁k ₁₃；

k ₂₁=M ₀₁(1+k ₁₂k ₀₂)+M ₀₀k ₁₁k ₀₂+M ₀₂k ₀₂k ₁₃；

k ₂₂=M ₀₂(-f+k ₁₃k ₀₃)+M ₀₁k ₀₃k ₁₂+M ₀₀k ₀₃k ₁₁-M ₀₀x ₁-M ₀₁y ₁；

k ₂₃=M ₁₀(1+k ₁₁k ₀₁)+M ₁₁k ₀₁k ₁₂+M ₁₂k ₀₁k ₁₃；

k ₂₄=M ₁₁(1+k ₁₂k ₀₂)+M ₁₀k ₁₁k ₀₂+M ₁₂k ₀₂k ₁₃；

k ₂₅=M ₁₂(-f+k ₁₃k ₀₃)+M ₁₁k ₀₃k ₁₂+M ₁₀k ₀₃k ₁₁-M ₁₀x ₁-M ₁₁y ₁；

k ₂₆=M ₂₀(1+k ₁₁k ₀₁)+M ₂₁k ₀₁k ₁₂+M ₂₂k ₀₁k ₁₃；

k ₂₇=M ₂₁(1+k ₁₂k ₀₂)+M ₂₀k ₁₁k ₀₂+M ₂₂k ₀₂k ₁₃；

k ₂₈=M ₂₂(-f+k ₁₃k ₀₃)+M ₂₁k ₀₃k ₁₂+M ₂₀k ₀₃k ₁₁-M ₂₀x ₁-M ₂₁y ₁；

Step S314, with the first eight coefficient k ₂₀, k ₂₁, k ₂₂, k ₂₃, k ₂₄, k _25,k ₂₆, k ₂₇Respectively to the 9th coefficient k ₂₈Carry out division arithmetic and obtain one group of totally eight DLT conversion coefficient;

Described eight DLT conversion coefficients arrive the DLT conversion coefficient on virtual representation plane as each of single camera as the plane.

Preferably, in step S400, describedly generate spliced virtual image according to described DLT platform calibration parameter from sub-image, comprise the steps:

Arrive the DLT conversion coefficient on virtual representation plane as the plane according to each, calculate respectively each one camera and be converted to the virtual representation planimetric coordinates as planimetric coordinates, determine the overlapping region of four sub-images, the virtual representation planimetric coordinates of per two sub-images is weighted the coordinate of the new virtual image of average conduct, finishes the concatenation of virtual image according to the coordinate of described new virtual representation.

Further, after generating spliced virtual image according to described DLT platform calibration parameter from sub-image, also comprise the steps:

Step S500, the virtual image that will splice reconstruct carry out smoothing processing, obtain the central projection virtual image.

Wherein, spliced virtual image also needs to carry out the operations such as dodging, and to obtain stitching image, the embodiment of the invention repeats no more this.

Need to prove, can also obtain formula six in the joining method that the embodiment of the invention provides:

$\left(\begin{array}{c}{x}_1-x_{01}\\ y_1-y_{01}\\ {-f}_1\end{array}\right)={\mathrm{\λ}}_1/{\mathrm{\λ}}_1^{\′}{R}_1{R}_1^{{\′}^{-1}}(\left(\begin{array}{c}x\\ y\\ -f\end{array}\right)+{\mathrm{\λ}}_1^{\′}{R}_1^{\′}\left(\begin{array}{c}{X}_{s0}-X_{s1}\\ Y_{s0}-Y_{s1}\\ Z_{s0}-Z_{s1}\end{array}\right));$

Utilize formula six, equally also can obtain each virtual representation plane to the DLT conversion coefficient on picture plane by calculating.The joining method that the embodiment of the invention provides is given unnecessary details no longer one by one to this.

Those skilled in the art are to be understood that, the embodiment of the invention provide based on DLT coefficient large format digital air photo instrument image splicing method, introducing by the DLT coefficient, platform calibration parameter can be converted to the DLT coefficient, thereby set up one camera as the relation of plane between the virtual representation plane, only need to calculate simple multiplication and division plus and minus calculation, namely one camera can be converted to the virtual representation planimetric coordinates as planimetric coordinates, finish image joint.Avoided complex calculations processing in the prior art, so this joining method is simpler, is conducive to the image splicing method fast processing.

The above is the preferred embodiments of the present invention only, is not limited to the present invention, and for a person skilled in the art, the present invention can have various modifications and variations.Within the spirit and principles in the present invention all, any modification of doing, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (3)

1. one kind based on DLT coefficient large format digital air photo instrument image splicing method, it is characterized in that, comprises the steps:

Exposure obtains four sub-images to synchronization;

In the cruciform overlay region, carry out the tie point coupling of four sub-images, according to the tie point between sub-image, and the platform calibration initial value of each sub-image, calculate the elements of exterior orientation of every sub-image;

According to the elements of exterior orientation of four sub-images of synchronization exposure, determine the relative position relation of four cameras, calculate DLT platform calibration parameter according to described relative position relation;

Generate spliced virtual image according to described DLT platform calibration parameter from sub-image.

2. as claimed in claim 1ly it is characterized in that based on DLT coefficient large format digital air photo instrument image splicing method,

The elements of exterior orientation of described four sub-images according to synchronization exposure is determined the relative position relation of four cameras to calculate DLT platform calibration parameter according to described relative position relation, comprises the steps:

Each that calculate respectively every camera in four cameras is as the DLT conversion coefficient of plane to the virtual representation plane;

According to the DLT conversion coefficient of four cameras, determine the relative position relation of four cameras, according to the relative position relation calculating DLT platform calibration parameter of four cameras.

3. as claimed in claim 1ly it is characterized in that based on DLT coefficient large format digital air photo instrument image splicing method,

Described each that calculate respectively every camera in four cameras comprises the steps: as the DLT conversion coefficient of plane to the virtual representation plane

Calculate single camera to the perspective transform of surface level;

If the elements of exterior orientation of camera is X _S1, Y _S1, Z _S1,

ω ₁, κ ₁, elements of interior orientation is x ₀₁, y ₀₁, f ₁, be x as coordinate ₁, y ₁, corresponding surface level is (X, Y, H), rotation matrix is R1, obtains formula one according to collinearity equation:

$\left(\begin{array}{c}{x}_1-x_{01}\\ y_1-y_{01}\\ -f_1\end{array}\right)={\mathrm{\λ}}_1{R}_1\left(\begin{array}{c}X-X_{s1}\\ Y-Y_{s1}\\ H-Z_{s1}\end{array}\right);$

Rotation matrix: $R_1=\left[\begin{array}{ccc}{a}_1& a_2& a_3\\ b_1& b_2& b_3\\ c_1& c_2& c_3\end{array}\right];$

During coordinates computed, establish k ₀₁=a ₃/ (H-Z _S1), k ₀₂=b ₃/ (H-Z _S1), k ₀₃=-(a ₃x ₀₁+ b ₃y ₀₁+ c ₃f ₁)/(H-Z _S1);

λ ₁=(a ₃(x ₁-x ₀₁)+b ₃(y ₁-y ₀₁)-c ₃f ₁)/(H-Z _s1)=k ₀₁x+k ₀₂y+k ₀₃；

The calculated level face is to the perspective transform on virtual representation plane;

If the elements of exterior orientation of virtual camera is X _S0, Y _S0, Z _S0,

ω ₀, κ ₀, focal length is f, is x, y as coordinate, and corresponding surface level is (X, Y, 0), and rotation matrix is R' ₁:

X _s0=(X _s1+X _s2)/2

Y _s0=(Y _s1+Y _s2)/2

Z _s0=(Z _s1+Z _s2)/2

ω _s0x(ω _s1+ω _s2)/2

κ _s0=(κ _s1+κ _s2)/2

Obtain formula two according to collinearity equation:

$\left(\begin{array}{c}x\\ y\\ -f\end{array}\right)={\mathrm{\λ}}_1^{\′}{R}_1^{\′}\left(\begin{array}{c}X-X_{s0}\\ Y-Y_{s0}\\ H-Z_{s0}\end{array}\right);$

Wherein, rotation matrix is ${R_1}^{\′}=\left[\begin{array}{ccc}{c}_{00}& c_{01}& c_{02}\\ c_{10}& c_{11}& c_{12}\\ c_{20}& c_{21}& c_{22}\end{array}\right];$

Obtain thus:

$\left(\begin{array}{c}X-X_{s0}\\ Y-Y_{s0}\\ H-Z_{s0}\end{array}\right)=R_1^{{\′}^{-1}}/{\text{\λ}}_1^{\′}\left(\begin{array}{c}x\\ y\\ -f\end{array}\right);$

Obtain perspective transform coefficient: λ ' ₁=(C ₂₀X+C ₂₁Y-C ₂₂F)/(H-Z _S0);

Calculating ${\mathrm{\λ}}_1{R}_1\left(\begin{array}{c}{X}_{s1}-X_{s0}\\ Y_{s1}-Y_{s0}\\ Z_{s1}-Z_{s0}\end{array}\right)$ The time, establish it and be ${\mathrm{\λ}}_1\left(\begin{array}{c}{k}_{11}\\ k_{12}\\ k_{13}\end{array}\right),$ Then calculate:

k ₁₁=C ₀₀(X _s1-X _s0)+C ₀₁(Y _s1-Y _s0)+C ₀₂(Z _s1-Z _s0)；

k ₁₂=C ₁₀(X _s1-X _s0)+C ₁₁(Y _s1-Y _s0)+C ₁₂(Z _s1-Z _s0)；

k ₁₃=C ₂₀(X _s1-X _s0)+C ₂₁(Y _s1-Y _s0)+C ₂₂(Z _s1-Z _s0)；

Arrive the perspective transform coefficient of surface level and the perspective transform coefficient that surface level arrives the virtual representation plane according to single camera, ask for single camera to the DLT conversion coefficient of the Relative Transformation of virtual camera;

Can be obtained by formula one:

Formula three $\left(\begin{array}{c}X-X_{s1}\\ Y-Y_{s1}\\ H-Z_{s1}\end{array}\right)={R_1}^{-1}/{\mathrm{\λ}}_1\left(\begin{array}{c}{x}_1-x_{01}\\ y_1-y_{01}\\ -f_1\end{array}\right);$

Obtain thus formula four:

$\left(\begin{array}{c}X\\ Y\\ H\end{array}\right)={R_1}^{-1}/{\mathrm{\λ}}_1\left(\begin{array}{c}{x}_1-x_{01}\\ y_1-y_{01}\\ {-f}_1\end{array}\right)+\left(\begin{array}{c}{X}_{s1}\\ Y_{s1}\\ Z_{s1}\end{array}\right);$

Formula four substitution formula two are obtained formula five:

Formula five is:

$\left(\begin{array}{c}x\\ y\\ -f\end{array}\right)={\mathrm{\λ}}_1^{\′}/{\mathrm{\λ}}_1{R}_1^{\′}{R_1}^{-1}(\left(\begin{array}{c}{x}_1-x_{01}\\ y_1-y_{01}\\ {-f}_1\end{array}\right)+{\mathrm{\λ}}_1{R}_1\left(\begin{array}{c}{X}_{s1}-X_{s0}\\ Y_{s1}-Y_{s0}\\ Z_{s1}-Z_{s0}\end{array}\right));$

Wherein: ${R_1}^{\′}{R}_1^{-1}=\left[\begin{array}{ccc}{M}_{00}& M_{01}& M_{02}\\ M_{10}& M_{11}& M_{12}\\ M_{20}& M_{21}& M_{22}\end{array}\right];$

Calculate $\left(\begin{array}{c}x\\ y\\ -f\end{array}\right)$ The time, establish it and be $\left(\begin{array}{c}{k}_{20}x+k_{21}y+k_{22}\\ k_{23}x+k_{24}y+k_{25}\\ k_{26}+k_{27}y+k_{28}\end{array}\right),$ Then calculate nine coefficients:

k ₂₀=M ₀₀(1+k ₁₁k ₀₁)+M ₀₁k ₀₁k ₁₂+M ₀₂k ₀₁k ₁₃；

k ₂₁=M ₀₁(1+k ₁₂k ₀₂)+M ₀₀k ₁₁k ₀₂+M ₀₂k ₀₂k ₁₃；

k ₂₂=M ₀₂(-f+k ₁₃k ₀₃)+M ₀₁k ₀₃k ₁₂+M ₀₀k ₀₃k ₁₁-M ₀₀x ₁-M ₀₁y ₁；

k ₂₃=M ₁₀(1+k ₁₁k ₀₁)+M ₁₁k ₀₁k ₁₂+M ₁₂k ₀₁k ₁₃；

k ₂₄=M ₁₁(1+k ₁₂k ₀₂)+M ₁₀k ₁₁k ₀₂+M ₁₂k ₀₂k ₁₃；

k ₂₅=M ₁₂(-f+k ₁₃k ₀₃)+M ₁₁k ₀₃k ₁₂+M ₁₀k ₀₃k ₁₁-M ₁₀x ₁-M ₁₁y ₁；

k ₂₆=M ₂₀(1+k ₁₁k ₀₁)+M ₂₁k ₀₁k ₁₂+M ₂₂k ₀₁k ₁₃；

k ₂₇=M ₂₁(1+k ₁₂k ₀₂)+M ₂₀k ₁₁k ₀₂+M ₂₂k ₀₂k ₁₃；

k ₂₈=M ₂₂(-f+k ₁₃k ₀₃)+M ₂₁k ₀₃k ₁₂+M ₂₀k ₀₃k ₁₁-M ₂₀x ₁-M ₂₁y ₁；

With the first eight coefficient k ₂₀, k ₂₁, k ₂₂, k ₂₃, k ₂₄, k ₂₅, k ₂₆, k ₂₇Respectively to the 9th coefficient k ₂₈Carry out division arithmetic and obtain one group of totally eight DLT conversion coefficient;

Described eight DLT conversion coefficients arrive the DLT conversion coefficient on virtual representation plane as each of single camera as the plane.

As claimed in claim 3ly it is characterized in that based on DLT coefficient large format digital air photo instrument image splicing method,

Generate spliced virtual image according to described DLT platform calibration parameter from sub-image, comprise the steps:

Arrive the DLT conversion coefficient on virtual representation plane as the plane according to each, calculate respectively each one camera and be converted to the virtual representation planimetric coordinates as planimetric coordinates, determine the overlapping region of four sub-images, the virtual representation planimetric coordinates of per two sub-images is weighted the coordinate of the new virtual image of average conduct, finishes the concatenation of virtual image according to the coordinate of described new virtual image.

As claimed in claim 4ly it is characterized in that based on DLT coefficient large format digital air photo instrument image splicing method,

, after generating spliced virtual image, sub-image also comprises the steps: according to described DLT platform calibration parameter

The virtual image of splicing reconstruct is carried out smoothing processing, obtain smooth and seamless central projection image.