CN104021535A - Method for splicing stepping and framing CCD image - Google Patents

Abstract

The invention provides a stepping and framing CCD image splicing method. The method includes: step A: in the overlapping region of two adjacent images, using a scale-invariant feature transformation operator to extract the two feature points with the same name that are farthest away from the image-the feature point with the same name P1 and the feature point with the same name P2; step B: According to the coordinates of the two extracted feature points of the same name, calculate four parameters of the elastic transformation model between two adjacent images; step C: according to the elastic transformation model, carry out coordinate transformation processing on each point of the current image; and step D: Combine the transformed current image and the previous image into one image according to the coordinate position, and obtain the spliced image. The present invention only needs to quickly extract two feature points with the same name which are the farthest in the overlapping area, and better adapts to the feature of low overlap between stepping and framing images.

Description

Method of stepping and framing CCD image mosaic

technical field

The invention relates to the technical field of image processing, in particular to a method for mosaicing CCD images by stepping and framing.

Background technique

The aerial stepping and framing camera exchanges time for space, takes pictures and images at multiple positions perpendicular to the flight course, and performs image stitching processing on the basis of ensuring a certain overlap rate, which can realize large field of view observation on the ground. It has the advantages of small size and light weight, and can be easily applied to platforms such as drones. Its basic working principle is that the camera controls the swing mechanism to move the field of view to the starting point for exposure imaging to obtain an image. During the process of image data transfer and storage, the swing mechanism moves the field of view step by step to the second position. The position is still and static; after the data transfer of the previous image is completed, the field of view of the second position is imaged. Under the premise that the camera controls the sweeping mechanism, the field of view will be moved to the first field of view position in the next round, and a new round of sweeping and stepping shooting process will start.

At present, foreign typical representatives include cameras such as CA-260, CA-261, CA-265 and CA-270, as well as A3 ultra-wide-format aerial digital cameras. -14, SR-71, P-3 and RQ-4 Global Hawk and other aircraft equipment. The framing camera takes several whole images from different points in space, and can obtain a single image with high geometric fidelity, but it puts forward higher requirements for image stitching technology. Because the image acquires an image sequence with a certain overlap rate at different times, affected by the uncertainty of the aircraft attitude and flight speed, the overlap between the acquired regional images is also constantly changing, which increases the complexity and difficulty of regional image processing. , so the image stitching algorithm has higher requirements.

At present, there are very few reports on the research results of stepping and framing CCD image mosaic processing at home and abroad, only similar rocking image mosaic, such as Gao Yinghui (Gao Yinghui, Shen Zhenkang, registration mosaic algorithm for aerial rocking image group, volume 38 No. 1 Period, Infrared and Laser Engineering 2009) studied the swing model of the aerial swing camera and adopted the correction model of rotation and perspective deformation, and proposed a registration and stitching method for the swing image group, mainly by defining the energy function to search for the optimal match region, followed by smooth stitching of groups of rocking images.

At present, there are more studies on the splicing of CCD area array images acquired by drones or other aerial area array images. These splicing processing methods focus on how to extract the feature points of the same name in the overlapping area through the matching algorithm, and then calculate the spatial transformation relationship of adjacent area array images, and complete the image splicing on the basis of the transformation. For example, Yu Huan et al. (Yu Huan, Kong Bo, Research on Automatic Seamless Stitching Technology of UAV Remote Sensing Images, Vol. 27, No. 3, June 2012, Remote Sensing Technology and Application) proposed the control point registration using features The algorithm and the contrast modulation method of image grayscale fusion realize the automatic seamless splicing technology of multiple UAV images. The eight-parameter transformation model is adopted, and at least 10-20 feature points with the same name are passed between adjacent images. Wang Bo et al. (Wang Bo, Gong Zhihui, Jin Keqiang, etc., An Improved UAV Image Stitching Method, Computer Engineering and Application, 2011, Volume 47, Issue 35) Aiming at the slow processing speed of traditional SIFT-based image stitching, by analyzing The homography matrix (perspective transformation model) relationship between the original image and the reduced image is proposed to improve the stitching speed.

The existing aerial remote sensing image mosaic technology, especially the aerial image acquired by UAV, is aimed at the mosaic problem of multiple flight belts and multiple images. Generally, a high overlap rate is required, and more parameters are often used in the transformation model. The eight-parameter transformation model of the homography matrix, the perspective transformation of the homography matrix, or the affine transformation that requires six parameters, which requires extracting more feature points with the same name in the overlapping area for calculation. If there are few feature points with the same name or the distribution is unreasonable, it will be difficult. Calculate the parameters. However, the images that need to be spliced for step-by-step framing images are images that are acquired sequentially on the same airway, and have a low overlap rate between images. A complex transformation model with many parameters cannot be used, but a very small amount must be extracted based on the low overlap rate. The feature points with the same name construct a transformation model with fewer parameters.

Contents of the invention

(1) Technical problems to be solved

In view of the above-mentioned technical problems, the present invention provides a step-by-frame CCD image stitching method to stitch multiple images accurately and quickly.

(2) Technical solution

The method for stepping and framing CCD image mosaic of the present invention includes: Step A: In the overlapping region of two adjacent images, use the scale-invariant feature transformation operator to extract the two feature points with the same name at the farthest distance in the image-the feature point with the same name P1 and the feature point P2 with the same name; Step B: According to the coordinates of the extracted two feature points with the same name, calculate the four parameters of the elastic transformation model between two adjacent images, the calculation formula is:

$\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}\\ {\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}\end{array}\right]<span\; class="notranslate">\; =</span>{\left[\begin{array}{cccc}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& {\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\\ {\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\end{array}\right]$

Among them: k ₁ , k ₂ , d _x , d _y are the four parameters of the elastic transformation model, (x ₁ , y ₁ ) and (u ₁ , v ₁ ) are the feature point P1 with the same name in the current image and the previous image respectively. The image coordinates in the image, (x ₂ , y ₂ ) and (u ₂ , v ₂ ) are respectively the image coordinates of the feature point P2 with the same name in the current image and the previous image, indicating the feature point with the same name; step C: according to The elastic transformation model performs coordinate transformation processing on each point of the current image, and the transformation formula is:

$\left\{\begin{array}{c}\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}\\ \mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}\end{array}\right.$

Wherein, (x, y) are the original image coordinates of the current image, and (u, v) are the image coordinates after the transformation of the current image; and step D: the transformed current image and the previous image according to the coordinate position Synthesize into an image to obtain a spliced image.

(3) Beneficial effects

As can be seen from the foregoing technical solutions, the step-by-step framing CCD image splicing method of the present invention has the following beneficial effects:

(1) According to the characteristics of low overlap rate between adjacent images, it only needs to quickly extract the two feature points with the same name at the farthest distance in the overlapping area, which better adapts to the characteristics of low overlap between stepping and framing images. Moreover, since the extracted feature points with the same name have the farthest distance, the accurate solution of the model parameters is ensured.

(2) The transformation model between adjacent images considers the four-parameter elastic transformation model of translation, scale and rotation, which has the characteristics of less parameter models and strong practicability, and better solves the geometric transformation problem between stepping and framing images .

Description of drawings

Fig. 1 is the flow chart of the method for splicing CCD images stepping and framing according to an embodiment of the present invention;

Fig. 2A is 16 step-by-step framing images obtained before splicing;

FIG. 2B is an effect diagram after splicing the stepping and framing images shown in FIG. 2A by using the method in FIG. 1 .

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be described in further detail below in conjunction with specific embodiments and with reference to the accompanying drawings. It should be noted that, in the drawings or descriptions of the specification, similar or identical parts all use the same figure numbers. Implementations not shown or described in the accompanying drawings are forms known to those of ordinary skill in the art. Additionally, while illustrations of parameters including particular values may be provided herein, it should be understood that the parameters need not be exactly equal to the corresponding values, but rather may approximate the corresponding values within acceptable error margins or design constraints. The directional terms mentioned in the embodiments, such as "upper", "lower", "front", "rear", "left", "right", etc., are only referring to the directions of the drawings. Therefore, the directional terms used are for illustration and not for limiting the protection scope of the present invention.

The method for splicing and stepping and framing CCD images of the present invention uses two feature points with the same name that are farthest apart in space to calculate the four parameters of the elastic transformation model between two adjacent images, and realizes the space between two adjacent images through coordinate transformation processing. Unification of coordinates; Adjacent images are synthesized together on the basis of unified spatial coordinates, thus ensuring the splicing of multiple images.

In an exemplary embodiment of the present invention, a method for fast stitching of CCD images with stepping and framing is provided. FIG. 1 is a flow chart of a step-and-frame CCD image mosaic method according to an embodiment of the present invention. As shown in Figure 1, the method for stepping and framing CCD image mosaic of the present embodiment comprises:

Step A: In the overlapping area of two adjacent images, use the Scale Invariant Feature Transform (SIFT) operator to extract the two feature points with the same name that are farthest away in the image - feature point P1 and feature point P2;

In this step, the method of using the scale-invariant feature transformation operator to extract the feature points of the same name in two adjacent images is a common method in the field. Generally speaking, it mainly includes:

Sub-step A1, using the difference Gaussian operator to detect the points of interest in the image scale space for the current image and the previous image respectively;

Sub-step A2, on the basis of the detected interest point, apply the gradient information in the neighborhood window of the interest point to calculate the main direction of the gradient, and construct the 128-dimensional feature vector corresponding to the interest point according to the gradient histogram;

Sub-step A3, calculate the Euclidean distance between the feature vector of the point of interest detected in the current image and the feature vector of the point of interest detected in the previous frame of image one by one, and calculate the ratio of the closest distance to the next closest distance, and set the ratio less than 0.5 Those interest points are candidate feature points with the same name;

Sub-step A4, using these candidate feature points with the same name, calculate the spatial distance between any one of the feature points and the remaining feature points one by one according to the coordinates of the feature points of the same name, and select the distance corresponding to the maximum value from the calculated space The two points of , as the final feature point P1 and feature point P2.

Step B: Calculate the four parameters of the elastic transformation model between two adjacent images according to the extracted coordinates of the two feature points with the same name, the calculation formula is:

$\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}\\ {\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}\end{array}\right]<span\; class="notranslate">\; =</span>{\left[\begin{array}{cccc}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& {\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\\ {\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; \&CenterDot;</span>\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\end{array}\right]<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Among them: k ₁ , k ₂ , d _x , d _y are the four parameters to be calculated, (x ₁ , y ₁ ) represents the image coordinates of the feature point P1 with the same name in the current image, (u ₁ , v ₁ ) represents The image coordinates of the feature point P1 with the same name in the previous image, (x ₂ , y ₂ ) means the image coordinates of the feature point P2 with the same name in the current image (u ₂ , v ₂ ) means the feature point P2 with the same name in the previous image The image coordinates in the image, [] is the expression form of a linear algebra matrix, []·[] means multiplying two matrices, and [] ^-1 means inverting the matrix.

Step C: Transform the coordinates of the current image according to the calculated four parameters. The transformation formula is:

$\left\{\begin{array}{c}\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}\\ \mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}\end{array}\right.$

Among them: k ₁ , k ₂ , d _x , d _y are four transformation parameters, these four values are calculated by step B, (x, y) represents the original image coordinates of the current image, (u, v) represents the current The transformed image coordinates.

Step D: Combining the transformed current image and the previous image into one image according to the coordinate positions, and then the spliced image can be obtained.

This embodiment only provides a method for realizing splicing of two stepping images. For multiple two stepping images, it is only necessary to perform the above similar processing on each image and the spliced image, that is, the spliced image is used as the previous image for each processing, and The image to be spliced is used as the current image, and multiple spliced stepping images are obtained after multiple splicing processes.

In order to verify the effect of this embodiment, FIG. 2A shows 16 stepping and framing images obtained before splicing; FIG. 2B is an effect diagram after splicing processing. From Figure 2A and Figure 2B, it can be seen that the spliced images have no gaps in the adjacent joint parts, and the transition of ground objects is natural, accurate and smooth, and the accurate splicing of adjacent images is achieved.

It should be noted that, since Figure 2A is a real image obtained by an aerial stepping and framing camera, it carries part of the longitude and latitude signals, and Figure 2B also has part of the longitude and latitude signals. The relationship between this part of the signal and the present invention is not very great, rather vague, and can be ignored.

So far, the present embodiment has been described in detail with reference to the drawings. Based on the above description, those skilled in the art should have a clear understanding of the method for stitching aerial multiple simultaneous images of the present invention.

In addition, the above-mentioned definitions of each element and method are not limited to the various specific structures, shapes and methods mentioned in the embodiments, and those skilled in the art can easily and well-known replace them.

In summary, in the method for splicing stepping images of the present invention, by only extracting two feature points with the same name that are farthest apart in space in the overlapping area of adjacent images, it better adapts to the characteristics of low overlap between stepping and framing images . However, using two feature points with the same name to calculate the four-parameter transformation model between adjacent images overcomes the shortcomings of the conventional camera transformation model with more parameters, has the characteristics of less parameter models and strong practicability, and better solves the problem of stepping and framing. The problem of geometric transformation between images.

The specific embodiments described above have further described the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above descriptions are only specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (3)

1. a method for stepping framing ccd image splicing, is characterized in that, comprising:

Steps A: in adjacent two width doubling of the image regions, utilize yardstick invariant features transformation operator to extract image middle distance two unique point of the same name-unique point P1 of the same name and unique point P2 of the same name farthest;

Step B: according to the coordinate of two unique points of the same name extracting, calculate four parameters of the elastic registration model between adjacent two width images, computing formula is:

$\left[\begin{array}{c}{k}_1\\ k_2\\ d_x\\ d_y\end{array}\right]={\left[\begin{array}{cccc}{x}_1& y_1& 1& 0\\ -y_1& x_1& 0& 1\\ x_2& y_2& 1& 0\\ -y_2& x_2& 0& 1\end{array}\right]}^{-1}\&CenterDot;\left[\begin{array}{c}{u}_1\\ v_1\\ u_2\\ v_2\end{array}\right]$

Wherein: k ₁, k ₂, d _x, d _yfor four parameters of elastic registration model, (x ₁, y ₁) and (u ₁, v ₁) be respectively the image coordinate of unique point P1 of the same name in current width image and upper piece image, (x ₂, y ₂) and (u ₂, v ₂) be respectively the image coordinate of unique point P2 of the same name in current width image and upper piece image and represent unique point of the same name;

Step C: according to elastic registration model, each point of current width image is carried out to coordinate transform processing, the formula of conversion is:

$\left\{\begin{array}{c}u=k_{1}x+k_{2}y+d_x\\ v=-k_{2}x+k_{1}y+d_y\end{array}\right.$

Wherein, (x, y) is the original image coordinate of current width image, and (u, v) is the image coordinate after current width image conversion; And

Step D: current width image and upper piece image after conversion are synthesized to piece image according to coordinate position, obtain spliced image.

2. method according to claim 1, is characterized in that, if while still having image to be spliced, after described step D, also comprises:

Using spliced image as upper piece image, using image to be spliced as current width image, continue execution step A～D.

3. method according to claim 1 and 2, is characterized in that, described steps A comprises:

Sub-step A1, utilizes the point of interest in difference Gauss operator detected image metric space for current width image and upper piece image respectively;

Sub-step A2 applies the gradient information compute gradient principal direction in this point of interest neighborhood window, and according to histogram of gradients, builds the proper vector of 128 dimensions that this point of interest is corresponding on the point of interest basis of detecting;

Sub-step A3, calculate one by one the Euclidean distance between the point of interest proper vector detecting in the point of interest proper vector that detects in current width image and previous frame image, and calculate the ratio of minimum distance and time minimum distance, ratio is less than to those points of interest of 0.5 as candidate's unique point of the same name; And

Sub-step A4, these candidates unique point of the same name that utilization obtains, according to the coordinate of unique point of the same name, calculate one by one the wherein space length between any one unique point and remaining other unique points, corresponding two points of chosen distance maximal value from the space calculating, as final unique point P1 and unique point P2.