KR100467466B1 - Method for obtaining the geometry coordinate of a target point - Google Patents

Abstract

The present invention relates to a method of obtaining geographic coordinates of an object point from an image coordinate, and to a computer-readable information recording medium having recorded thereon a program for implementing the acquisition method. A technique for obtaining three-dimensional geographic coordinates from coordinates. According to the present invention, there are advantages in solving the infinite repetition and the point convergence problem in obtaining three-dimensional geographic coordinates from two-dimensional image coordinates.

Description

Method for obtaining the geometry coordinate of a target point}

The present invention relates to a method of obtaining geographic coordinates of an object point from an image coordinate, and to a computer-readable information recording medium having recorded thereon a program for implementing the acquisition method. A technique for obtaining three-dimensional geographic coordinates from coordinates.

In the satellite image of the earth, the process of establishing the relationship between the coordinates (image coordinates) and the actual three-dimensional geographical coordinates (geographic coordinates) on the two-dimensional satellite image is called camera modeling. It is represented by the focus of the sensor to be photographed, a target point on the ground surface, and a ray connecting the focus and the target point. This ray is generally modeled as a straight line without considering complex physics such as atmospheric refraction.

In camera modeling, the process of finding geographic coordinates of an object point from image coordinates is called direct modeling, and the process of finding image coordinates from geographic coordinates is called indirect modeling. Unfortunately, due to the nonlinear / time-varying nature of the observation system and the irregularities of the terrain surface, neither direct modeling nor indirect modeling is clearly solved, and thus converged by iterative computation according to the general problem solving theory of nonlinear systems. get a converged solution.

On the other hand, one of the important objectives of camera modeling is to obtain ortho-images of observation objects using satellite images and surface elevation models. An orthographic image is described with reference to FIG. 1. As shown in FIG. 1, the focal point 110 of the satellite observation system acquires an image by scanning the ground surface 120. The satellite observation system assumes that the ground surface 120 is positioned at a predetermined altitude reference 170. Since the scanning image is generated, a predetermined error 160 determined by the altitude 150 and the incident angle θ between the geographic coordinates of the actual objective point 130 and the image coordinate represented by the virtual objective point 140 in FIG. Will be present. This error exists for each pixel of the satellite image, and these errors are combined to form the image distortion as a whole. The orthographic projection image is an image from which such distortion is removed and refers to an image in which the destination point 130 appears at the original position.

It is known that the above-described direct modeling method, that is, the method of finding the geographic coordinates of the target point from the image coordinates, has two disadvantages in obtaining the orthographic images. The first is that iterative operation is required and the second is image. The mapping to brightness values is incorrect. In particular, the second disadvantage has a negative effect on the quality of the product, which is particularly noticeable when the position of the actual target point projected for each image pixel is irregularly placed on the geographical coordinates.

Since geographical information can be obtained in advance by other methods without scanning by satellite, it is possible to generate orthoimages and surface elevation models in advance for a predetermined region. Raw satellite images can be generated by using ortho-images and surface elevation models, which are generated. In general, the generated image data is intended to develop an application algorithm of image data before actual launch of the earth observation system. Is provided.

In order to obtain the raw satellite image, it is necessary to determine the brightness value on the raw image coordinate system from the surface elevation information and the orthographic image. Therefore, in the case of the mapping problem of the above-described image brightness value, in this case, indirect modeling, that is, the image coordinate from the geographical coordinate Applies to course That is, in the process of finding the image coordinates from the geographic coordinates, there are problems in that iterative operation is required and mapping of the image brightness values is incorrect. In order to be closer to, it is more preferable to use a direct modeling technique, that is, a method of finding the geographical coordinates of the target point from the image coordinates.

In the direct modeling method, a ray tracing algorithm has been conventionally used to determine the three-dimensional geographic coordinates from two-dimensional image coordinates. This ray tracing algorithm has a disadvantage that an error may occur in a terrain where the altitude changes rapidly, which will be described with reference to FIG.

Fig. 2 is a diagram showing a ray tracing algorithm of the prior art. According to the ray tracing algorithm, as shown in Fig. 2 (a), the ray is expected to converge by accurately tracking the desired point on the ground surface. However, as shown in Fig. 2 (b), convergence does not occur when the inclination angle of the ground surface is larger than the incident angle of the light beam. Furthermore, in this case, the ray tracing algorithm does not end through a predetermined divergence condition but proceeds indefinitely. do. Also, as shown in Fig. 2 (c), when the ray intersects the ground surface at a plurality of points, the ray tracing algorithm converges to the wrong point. That is, the original destination point A is misinterpreted as another point B, converges to the wrong point B, and as a result, an incorrect brightness value is assigned to the simulated image.

These problems (Fig. 2 [b] and Fig. 2 [c]) with the prior art ray tracing algorithms arise for terrains with rapidly varying altitudes, which is not the case with high resolution digital surface models (DSM). As described above, there are many terrains in which the altitude changes rapidly. Therefore, the ray tracing algorithm of the prior art is difficult to apply to the simulation of high resolution satellite images.

Accordingly, an object of the present invention is to provide a method for obtaining three-dimensional object point geographic coordinates from two-dimensional image coordinates using a so-called ray following algorithm for simulation of a high resolution satellite image.

1 is a view for explaining the concept of an orthographic image.

2 illustrates a ray tracing algorithm of the prior art.

Fig. 3 is a diagram showing an example in which the ray-tracking algorithm of the present invention is well applied to a terrain having a sharp elevation change.

4 is a diagram for obtaining a vector equation of a ray in a method of obtaining geographic coordinates according to the present invention.

5 is a flowchart showing an example of a method of obtaining geographic coordinates according to the present invention.

6 is a diagram showing triangular geometry for obtaining a preferable r _min value in a method of obtaining geographic coordinates according to the present invention.

<Explanation of symbols for the main parts of the drawings>

310: focus of the satellite

320: surface

330: destination point

340: rays

350: increase

360, 370: additional intersection

Hereinafter, with reference to the accompanying drawings will be described in detail the present invention.

3 is a diagram showing an example in which the ray tracing algorithm of the present invention is well applied to a terrain 320 having a sharp elevation change. The terrain 320 of FIG. 3 is a terrain having a sharp change in altitude, and has a plurality of intersections 330, 360, and 370 as shown. Converging to the wrong point 370.

According to the ray tracing algorithm of the present invention, as shown in Fig. 3, the topographic altitude and the light altitude are compared while following the light beam from a predetermined altitude, and the algorithm is terminated when the topographic altitude is higher than the light altitude. This end point can be seen as the result of direct modeling. By using such a ray tracing technique, the infinite repetition problem (Fig. 2 [b]) and the point convergence problem (Fig. 2 [c]) which occur in the ray tracing algorithm of the prior art can be completely solved for the high resolution numerical indicator model. have.

In FIG. 3, the light ray 340 connecting the focal point 310 and the destination point 330 starts at a predetermined altitude (point [0]) and follows the light ray 340 and is lowered by a predetermined increment 350. Go to. ([0], [1], [2], ...) Comparing light elevation with topographic elevation while following a ray, at point ([0]) to point ([3]), the light elevation is higher than the topographic elevation. Although high, the terrain elevation is higher than the light elevation at the next point [4] after passing through the destination point 330. The ray tracing algorithm of the present invention detects this condition and uses it for detection of the target point 330. As shown in FIG. 3, the ray 340 and the terrain 320 have a plurality of intersections 330, 360, and 370, but according to the ray tracking algorithm of the present invention, the coordinates of the correct target point 330 can be obtained. .

An example of mathematically modeling a ray follower algorithm according to the present invention will be described with reference to FIGS. 4 to 6. As in the conventional photogrammetry, if the earth is assumed to be a plane and the intersection point between the ray and the surface is determined by adjusting the altitude variable (Z), modeling is easy but there is a big error. Therefore, in order to obtain more accurate results, the earth should be assumed to be elliptical, and in this specification, an Earth Centered of Rotation reference frame is used in modeling all nonlinear characteristics of the observation system and the ellipsoid of the earth. In this geocentric coordinate system, the origin is the Earth's center, the Z axis is the North Pole, the X axis is 0 longitude, the direction of the Greenwich Observatory in England, and the Y axis is 90 degrees longitude by the right hand rule.

First, referring to FIG. 4, a light ray according to the present invention is represented by a vector equation. As shown, assuming that the three-dimensional vectors (x _s , y _s , z _s ) and (x _d , y _d , z _d ) are the vectors representing the position of the focus and the normalized direction vectors of the rays, respectively, The vector equation is expressed as in Equation 1.

Here, the real number r determines the size of the light beam, that is, the distance from the satellite (focus) to the target point, so that the light ray can be followed by appropriately controlling the variable r.

According to the ray tracing algorithm of the present invention, a topographic altitude must be obtained for each point following the ray. In general, since the numerical indicator model is managed by the latitude (φ) and the longitude (λ), the geocentric coordinates (x, y, z) of the beam point are the latitude (φ), longitude (λ), Converted to a value for altitude h.

Here, a, b, and e are the semi-major, semi-minor, and eccentricity of the earth ellipsoid, respectively, and θ is defined as in Equation 3. In Equation 2, the height h corresponds to the above-described light altitude.

In Equations 2 and 3, if the data used in the numerical indicator model are based on geodetic criteria other than WGS-84, a process of converting geodetic criteria is required. In addition, if the altitude level of the numerical model is based on the average sea level and not the ellipsoid surface, a geoid model should be added.

Referring to Fig. 5, the ray tracking algorithm used in the present invention is summarized as follows. First, in step 510, the initial conditions for the ray tracing algorithm of the present invention are set. As one of the initial conditions, the r _min value corresponding to the starting point on the ray is preferably determined. This value is preferably determined as shown in equation (4) by the triangular geometry shown in FIG.

Here, H _focus is the altitude value of the focus, H _max is the maximum value, that is, the maximum altitude value of the altitude of the numerical indicator model, and θ _i is the angle of incidence of the light ray on the earth ellipsoid. The determination of the H _max value can be obtained by using prior knowledge or by comparing each grid value (height value) of the numerical indicator model. If the incidence angle θ _i of Equation 4 is 90 degrees, since the latitude, longitude, and altitude of the initial projection point are directly obtained from the numerical indicator model, it is no longer necessary to proceed with the ray tracking algorithm.

Further, in step 510, as one of the initial conditions, Δr 350, which is preferably an increment of r in ray tracing, is appropriately determined. In general, the smaller the value of the incremental value (Δr), the more the light-tracking algorithm of the present invention requires a large amount of calculation, while the larger the value of the incremental value (Δr), there is a risk of missing the sudden height discontinuity region of the numerical indicator model. do. In this sense, the value of the increment Δr is preferably determined from Equation 5 from the resolution of the numerical indicator model lattice. However, depending on the implementation, the incremental value Δr may be determined by prior knowledge or experience.

Here, D _DSM is the resolution of the numerical indicator model lattice and the incident angle θ _i is as defined in FIG. 6.

Next, in step 520, the value of the variable r to be used is determined. In the first case, the value of r is preferably determined as r _min , and in other cases, the value of r is increased by a predetermined increment Δr as described later with reference to Equation 6.

Subsequently, in step 530, the current geocentric coordinates (x, y, z) are determined from the value of the variable r in the light beam, and the values are substituted into Equation (2) for latitude / longitude / altitude (φ, λ, h). ) As described above, the altitude value h calculated in step 530 corresponds to the light altitude.

In operation 540, the terrain elevation H corresponding to the latitude / longitude φ and λ obtained in operation 530 is obtained from the numerical indicator model. In this case, the surface altitude is generally a value based on the ellipsoid plane, and if necessary, a transformation using a geoid is used.

In step 550, the topographical altitude H obtained in step 540 is compared with the light altitude altitude h obtained in step 540, and branches accordingly. First, if the terrain altitude H is less than the light altitude h, go to step 520 described above to reset the variable r and perform step 530 ?? 560 again, which indicates that the ray is still at ground surface. Because we judged that we are higher. On the other hand, when the terrain altitude H is greater than or equal to the light altitude h, the ray tracing according to the present algorithm is actually terminated because it has detected that the ray intersects the ground for the first time.

As an optional predetermined post-processing step 560, the final results φ, λ, h of direct modeling can be adjusted to have more accurate values by appropriately using the results at the previous point and the results at the current point on the ray. . In general, interpolation is properly performed from an altitude value between both points.

In the above-described step 520, the value of the variable r to be used is determined at this time. Each time the ray tracking process is performed, the position of the target point on the ray is as shown in Equation 6 by the predetermined increment determined in the step 510. Moving down, the ray tracking algorithm according to the present invention then performs step 530 ?? 560 described above again.

According to the method for obtaining geographic coordinates of the present invention, two problems that occur in the acquisition method according to the conventional ray tracing algorithm in acquiring three-dimensional geographic coordinates from two-dimensional image coordinates (direct modeling), namely There is an advantage that can solve the infinite repetition problem (Fig. 1 [b]) and the point convergence problem (Fig. 1 [c]).

Claims (10)

In the method for obtaining the three-dimensional geographic coordinates of the observation point from the two-dimensional image coordinates through the light beam connecting the observation focus point and the observation point, A first step of setting an initial condition of geographic coordinate search; A second step of repeatedly searching for a point at which the ray intersects the observation target point to obtain the intersection point; And a third step of calculating a geographical coordinate of the observation target point from the intersection point. The second step may include: selecting a specific tracking point on the ray and determining 3D coordinates of the selected tracking point; Determining a light ray elevation h for the light ray corresponding to the coordinate of the tracking point; Determining a topographic elevation (H) for the tracking point corresponding to the coordinate of the tracking point; And a second step of comparing the light elevation with the topographical altitude if the topographical elevation is equal to or greater than the light altitude, and processing the acquired intersection point; otherwise, proceeding to step 2a to perform the repetitive search. Method of obtaining geographic coordinates of the observation point, characterized in that the configuration. delete The method of claim 1, wherein the light ray is a vector equation [Where r is a variable for determining a tracking point on said ray; (x _d , y _d , z _d ) are the direction vectors of the rays; (x _s , y _s , z _s ) is the position of the observation focal point], and the setting of the first stage initial condition is based on the initial value (r ₀ ) and the increment value (Δr) of the variable r. r ₀ = F ₁ (H _focus , H _max , θ _i ) Δr = F ₂ (D _DSM , θ _i ) [Wherein H _focus is the altitude of the observation focal point; H _max is the maximum value of the altitude of the numerical indicator model for the observation point; D _DSM is the resolution of the numerical model grid for the observation point; [theta] _i is set according to the incident angle of the light to the ellipsoid of the earth]. The method of claim 3, wherein the initial value and the increment value of the variable r are respectively Method of obtaining geographic coordinates of the observation point, characterized in that set according to. delete delete 4. The method of claim 3, wherein the step 2a adds the increment value to the previous value of the variable r to determine the current value of the variable r and substitutes the current value of the variable r into the vector equation of the ray to follow the earth of the following point. The center coordinates (x, y, z) are determined and the step 2b is performed by calculating the geocentric coordinates of the following point. [Where a, b, and e are the major, minor, and eccentric ratios of the ellipsoids of the earth, respectively; θ is ] To calculate the latitude (φ), the longitude (λ) and the light altitude (h) of the following point, and the step 2c is a numerical index model for the observed point. Geographical coordinates acquisition method of the target point of observation, characterized in that performed by determining the terrain altitude (H) for the tracking point by substituting to. 4. The method of claim 3, wherein the third step substitutes the current value of the variable r into the vector equation of the ray to calculate the geocentric coordinates of the following point, and calculates the geocentric coordinates of the observed point. Method of obtaining geographic coordinates of an observation point, characterized in that determined by the coordinates. 4. The method of claim 3, wherein the third step substitutes the current value and the immediately previous value of the variable r into the vector equation of the ray, respectively, and calculates the geocentric coordinate values for each, and interpolates the two calculated results. And obtaining a geographical coordinate of the observation target point. A computer-readable information recording medium having recorded thereon a program that implements a method for acquiring a geographical coordinate of an objective point for observation according to any one of claims 1, 3, 4, and 7-9.