KR20180024809A - Image geometric correction methods and apparatus for the same - Google Patents

Abstract

According to a variety of embodiments of the present invention, an image geometric correction method is disclosed. According to the present invention, the method comprises the following steps of: receiving posture information of a satellite and a first image from the satellite; generating grids at a predetermined interval in the first image and extracting a first pixel which is a pixel of the first image located at grid points on the grids; detecting a common feature point between the first image and a second image; generating reference coordinates of a map coordinate system based on the posture information and the feature point; and estimating a map coordinate point, which is a coordinate point of the map coordinate system corresponding to the first pixel, based on the reference coordinates and performing geometric correction on the first image based on the estimated map coordinate point. Therefore, the method can perform accurate image geometric correction.

Description

[0001] Image geometry correction methods and apparatus for the same [0002]

Embodiments of the present invention relate to a method and apparatus for image geometry correction of a satellite.

The satellite is mounted on the launch vehicle and is launched from the earth. After separating from the launch vehicle, the propulsion system of the satellite itself is operated until it enters the target track, and after entering the target track, Can continue to fly. The satellites include equatorial orbiting satellites, polar orbiting satellites, and tilted orbit satellites, and the purpose of the satellites varies depending on the orbit. A satellite can take a desired surface on the earth according to its mission such as communication, earth observation, weather, navigation and positioning, space science research, etc., and the shot image can be transmitted to earth's ground station.

Image data collected from satellites and transmitted to ground stations are geometrically distorted due to changes in position and attitude of the satellite, velocity change of the satellite, rotation speed change of the scan range and scan line, earth rotation, map projection, In other words, the image transmitted to the ground station has a geometric warp, which indicates the difference in the absolute position of the image. This geometric warp means the position variation of each point in the image, and superimposes the warped image with the existing topographic map existing on the plane It is necessary to transform the position of each point appearing in the satellite image to have the same size and projection value as the topographic map. Such a transformation process is called geometric correction, and it is only through this process that we can obtain a stable image of the type commonly seen through the map.

In the case of building a database, mapping metrics such as the area after image classification, calculating adjacent images, mosaic mapping, mapping, etc. It is necessary to correct this geometric distortion. The work of removing this distortion from the image is called geometric correction. In the case of low-earth orbit satellite, because the main purpose of the image is to provide precise image for a specific region, it is not necessary for the processing and providing time of the image to be strict real time. Therefore, the user can use the satellite's orbit, attitude information, And a landmark, it is possible to perform geometric correction such as rotating an image or adjusting a scale.

However, geostationary satellite images should be photographed 24 hours a day, 365 days a year, and must be processed in real time and processed within a short time by positioning and resampling regardless of clouds. In this case, it is difficult for the user to directly determine the position of the image using the ground control point (GCP) and the landmark of the surface, and the position of the image must be determined automatically. To do this, the geo-correction system of the ground station indirectly performs the positioning and resampling of the image using the state vector of the satellite. In this case, there is a problem in that the result of the image matching after the image resampling is not precise so that the boundary appears at the overlapping portion between the image and the image.

SUMMARY OF THE INVENTION The present invention provides a method and apparatus for geometric correction of an image of a satellite using feature points detected in overlapping portions between images taken by the satellite.

In accordance with an aspect of the present invention, there is provided a method of correcting an image geometry comprising: receiving satellite orientation information and a first image from a satellite; generating a grid at a predetermined interval in the first image; Extracting a first pixel that is a pixel of the first image located in the first image, detecting common feature points between the first image and the second image, generating reference coordinates of the map coordinate system based on the attitude information and the feature points Estimating a map coordinate point, which is a coordinate point of the map coordinate system corresponding to the first pixel, based on the reference coordinates, and geometrically correcting the first image based on the estimated map coordinate point .

According to an example of the image geometry correction method, the second image is an image of a ground surface adjacent to the ground surface taken by the first image.

According to another example of the image geometry correction method, the method further includes estimating a state vector of the first image based on the attitude information.

According to another example of the image geometry correction method, the step of generating the reference coordinates includes selecting at least one reference pixel of the first pixels and selecting the at least one reference pixel based on the state vector of the first image, Detecting a coordinate point of the map coordinate system corresponding to the detected map coordinate system, and generating a first reference coordinate including the coordinate point of the detected map coordinate system.

According to another example of the image geometry correction method, the step of extracting the feature points may include a step of detecting a redundant image which is an image part overlapping with the second image in the first image, And extracting a preset object.

According to another example of the image geometry correction method, the preset object includes a first marker, which is at least one of a terrain surface, a shoreline, and a structure.

According to another example of the image geometry correction method, the preset object further includes a second marker that is a cloud on the ground surface.

According to another example of the image geometry correction method, the step of generating the reference coordinates may include obtaining coordinate points of the map coordinate system corresponding to the first mark and the second mark from the second image, And generating a second reference coordinate including a coordinate point of the coordinate system, wherein the reference coordinate includes the first reference coordinate and the second reference coordinate.

According to another example of the image geometry correction method, the geometric correction step may include estimating a relational expression of the reference coordinates and a position of the corresponding pixel on the image, and calculating a map coordinate of the selected first pixel And estimating a point.

According to another example of the image geometry correction method, the second image is a coordinate-converted image of the map coordinate system.

According to an aspect of the present invention, there is provided an image geometry correcting apparatus comprising: a receiving unit that receives a first image and attitude information from a satellite; a generating unit that generates a grid at a predetermined interval in the first image, A feature point detecting unit for detecting a feature point common to the first and second images, a reference coordinate generating unit for generating reference coordinates of the map coordinate system based on the attitude information and the feature points And a geometric correction unit for geometrically correcting the first image by estimating a map coordinate point, which is a coordinate point of the map coordinate system corresponding to the first pixel, based on the reference coordinates.

According to an embodiment of the image geometry correcting apparatus, the second image is an image of a surface of the ground adjacent to the ground surface taken by the first image.

According to another example of the image geometry correcting apparatus, the image geometry correcting apparatus further includes a state vector estimating unit that estimates a state vector of the first image based on the attitude information.

According to another example of the image geometry correction apparatus, the reference coordinate generator selects at least one reference pixel among the first pixels, and based on the state vector of the first image, Detects a coordinate point of the coordinate system, and sets the detected coordinate point as a first reference coordinate point.

According to another example of the image geometry correcting apparatus, the feature point detecting unit detects a redundant image which is an image portion overlapping the second image in the first image, and detects a redundant image common to the second image in the redundant image .

According to another example of the image geometry correction device, the preset object includes a first marker, which is at least one of a terrain surface, a shoreline, and a structure.

According to another example of the image geometry correcting device, the preset object further comprises a second mark that is a cloud on the surface of the earth.

According to another example of the image geometry correcting apparatus, the reference coordinate generating unit obtains information on coordinate points of the map coordinate system corresponding to the first mark and the second mark from the second image, And the reference coordinates include the first reference coordinates and the second reference coordinates.

According to another example of the image geometry correcting apparatus, the image geometry correcting apparatus further includes a relational expression estimating unit for estimating a relational expression of the reference coordinates and the image position of the corresponding pixel.

According to another example of the image geometry correction apparatus, the geometric correction unit estimates the map coordinate point of the selected first pixel using the estimated relational expression.

According to another example of the image geometry correction device, the artificial satellite is a geostationary satellite.

According to another example of the image geometry correction apparatus, the second image is a coordinate-converted image of the map coordinate system.

According to various embodiments of the present invention, feature points, which are reference points of geometric correction, are additionally extracted from images captured by satellites and overlapped with other images, and the images are matched with each other so that the land surface is covered with clouds, It is possible to perform more accurate image geometry correction by correcting the state vector of the satellite estimated using the feature points.

1 is a schematic view of a geostationary orbiting satellite revolving around the earth according to one embodiment.
2 is a view schematically showing an image taken by a geometrically corrected artificial satellite.
3 is a block diagram briefly showing an internal configuration of an image geometry correcting apparatus according to an embodiment.
4 is a flowchart briefly illustrating an image geometry correction method according to an exemplary embodiment.
5 is a flowchart briefly illustrating a method of generating a first reference coordinate according to an embodiment.
6 is a flowchart briefly illustrating a method of generating second reference coordinates according to an embodiment.
7 is a flowchart schematically showing a method of geometric correction of a first image based on reference coordinates according to an embodiment.

The present invention is capable of various modifications and various embodiments, and specific embodiments are illustrated and described in the drawings. The effects and features of the present invention and methods of achieving them will be apparent with reference to the embodiments described in detail below with reference to the drawings. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Also, the singular expressions include plural expressions unless the context clearly dictates otherwise. Also, the terms include, including, etc. mean that there is a feature, or element, recited in the specification and does not preclude the possibility that one or more other features or components may be added.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings, wherein like reference numerals refer to like or corresponding components throughout the drawings, and a duplicate description thereof will be omitted .

1 is a schematic view of a geostationary orbiting satellite revolving around the earth according to one embodiment.

Referring to FIG. 1, the satellite 10 revolves along an orbital plane parallel to the equatorial plane of the earth 200

The orbital plane of the artificial satellite 10 is a constant path in which the artificial satellite 10 moves, and the satellite orbit has a circular orbit and an elliptical orbit. The larger the orbit, the longer it takes for the satellite to spin its orbit. The orbital plane of the satellite 10 is divided into an equatorial orbit, an oblique orbit, and a polar orbit, depending on the equator and the orbital angle to be made. The equatorial orbit is the case where the equator and the raceway are made at 0 degree angle, and the angle 90 degrees is called the pole orbit. And the orbit between the equatorial orbital and the polar orbital is called the oblique orbit. In particular, the satellite 10 that revolves along the equatorial plane is called a geostationary satellite 10.

The satellite 10 for photographing the ground image may be a geosynchronous orbit satellite. The orbiting speed of geostationary satellites has a speed corresponding to the Earth's rotating speed. The geostationary satellite is located continuously over a specific surface of the earth, and captures an image of the specific surface. In other words, it appears that the geostationary orbits of the geostationary satellites are at the same altitude as the earth's rotation cycle, so they are stationary at a point above the earth's surface. Specifically, the satellite is located at 35,786 km (geostationary orbit) above the equator (0 ° at the Earth's latitude). At this time, the satellite can fly along the earth's orbit 24 hours at the same speed as the earth, so you can look at a large area of the earth at a point above the equator. However, since it is always in the sky above the equator, it is difficult to observe the polar regions and observe the images as they go to higher latitudes. Geostationary satellites are used for communication, broadcasting, and weather observation. For convenience of explanation, the following description is limited to geostationary-satellite satellites for the purpose of observing the weather.

Geostationary satellites can observe the generation and disappearance of clouds affecting the meteorological phenomena of the earth, the movement of clouds over time, and the types of clouds. Geostationary orbiting satellites periodically photograph the surface of the earth and transmit the photographed image to earth's earth station. The geosynchronous orbiting satellite can also provide the attitude information of the geosynchronous orbit satellite so that the coordinate information of the photographed location can be known.

The geosynchronous orbit satellites sequentially capture the image of a predetermined ground surface by changing the angle of the photographing apparatus. That is, the geostationary satellite divides the image of the predetermined ground surface into a plurality of images and shoots them. The ground station may combine images sequentially captured by the geostationary orbit satellite to produce a mosaic-processed image of a predetermined surface.

The geosynchronous satellite captures an image so that an adjacent image overlaps with a predetermined portion. That is, in order to efficiently remove the space between the image and the image according to the angle of the photographing apparatus of the artificial satellite 10 and the error of the posture of the artificial satellite 10, And the angle of the photographing apparatus is adjusted so as to overlap with each other.

2 is a view schematically showing an image taken by a geometrically corrected artificial satellite.

Referring to FIG. 2, the ground station receiving the photographed image of the satellite 10 geometrically corrects and resamples the image.

Geometric correction is a correction that removes geometric distortion caused by satellites' attitude, curvature of the earth, direction of the satellite, difference of coordinate projection method, error of observation instrument, effect of earth rotation.

The low orbital satellite image is highly accurate because it performs coordinate determination and resampling on the map coordinate system of the image using the orbit and attitude information of the satellite and the ground control point (GCP) or landmark of the ground surface . At this time, the operator selects the point where the GCP is not covered by the cloud, and since the processing and the providing time of the image do not need to be real time, the operator can have a sufficient post-processing process.

On the other hand, the image of geosynchronous orbit satellite is taken 36 hours a day and 36 hours a day, and the coordinate of the coordinates of the pixels of the image is determined and resampled regardless of the cloud, . In this case, since it takes a lot of time for the operator to directly perform the geomorphometric correction of the detected landmark by detecting the landmark of the entire image, the image received from the geosynchronous satellite is estimated as an estimated state vector, Image positioning and resampling are automatically performed by the computer based on the image data. That is, a state vector is firstly estimated as a residual by landmark matching, which is not directly applied to the positioning and resampling of the landmark, and the coordinates of the image on the map are determined using only the estimated state vector And resampling. This indirect method is suitable for positioning and resampling images taken for 24 hours, but does not directly apply the landmark information of the image itself to the coordinate point determination and resampling of pixels of the image Matching results after resampling of the image are not precise.

In order to determine the state vector of the satellite 10, the ground station needs to detect the landmark of the ground surface photographed by the satellite 10, but when the photographed ground surface is covered by the cloud, It is difficult to estimate the state vector by the residual due to the landmark matching. In this case, the error rate with the coordinates of the actual map becomes higher even if the image is corrected for the ground surface covered by the cloud or the ocean. In this case, when the geometric correction error rate of the satellite 10 is increased, a phenomenon in which the boundary at the portion where the image meets the image may be conspicuous as shown in FIG. 2 may occur.

Furthermore, the Meteorological Agency receiving such an image increases the error in predicting the weather. In forecasting the weather, the Meteorological Agency analyzes information on each pixel of the received image. The Meteorological Administration can not obtain the accurate information on the boundary portion which can be seen and can predict the inaccurate weather information.

Accordingly, when there is no landmark in the image and the accuracy of the state vector greatly decreases, the image geometry correcting apparatus 100 receives the coordinate information from the image obtained by relatively accurately estimating the state vector, Can be supplemented. This will be described below with reference to Figs. 3 to 7.

3 is a block diagram briefly showing an internal configuration of an image geometry correcting apparatus according to an embodiment.

3, the image geometry correcting apparatus 100 includes a receiving unit 110, a state vector estimating unit 120, a pixel selecting unit 130, a feature point detecting unit 140, a reference coordinate generating unit 150, A geometry corrector 170, and a resampler 180. The resampler 180 may be implemented by a computer.

The receiving unit 110 receives the image taken by the satellite 10 from the satellite 10 and the attitude information of the satellite 10 at the time of shooting. The receiving unit 110 may receive a plurality of consecutive images. The receiving unit 110 also receives attitude information of the satellite 10 at the time of shooting corresponding to the plurality of images. The posture information of the artificial satellite 10 may be obtained from the attitude of the artificial satellite 10 measured through a sun sensor, a geomagnetism sensor, a star sensor or a gyro sensor, which is an inertial sensor included in the artificial satellite 10, And a photographing angle of the photographing apparatus for photographing the image.

The state vector estimator 120 estimates the state vector of the satellite 10 through the attitude information of the received satellite 10. The state vector estimating unit 120 may estimate a state vector of the satellite 10 by matching the landmark of the received image with a previously stored landmark and using the residual as a result of the matching. The state vector estimating unit 120 estimates a state vector for each received image. On the other hand, the attitude information of the satellite 10 can provide attitude information with high accuracy at the beginning of the image photographing. However, if the matching with the previously stored landmark is impossible continuously during the photographing of the image, Accuracy is low.

The pixel selection unit 130 can select an object to be converted into a map coordinate system coordinate among pixels constituting an image. The pixel selection unit 130 generates a grid at a predetermined interval in the received image and selects a pixel located at a lattice point where two lines of the grid lines meet. For example, in the case of an image having 500 by 500 pixels, the pixel selection unit 130 generates grid lines orthogonal to every 100 pixels. In this case, the pixel selection unit 130 may include (100, 100), 100, 200, 100, 300, 100, 400, 200, 100, 200, 200, 200, 300, 400, 400, 300, 100, 300, 200, 300, 300, 300, 400, 400, 100, 400, 200, Lt; / RTI > can be selected.

 The minutiae point detecting unit 140 can detect minutiae points of the overlapped portions of at least two images of the geo-stationary satellite. The satellite 10 photographs a plurality of images sequentially on a predetermined range of the ground surface so that predetermined portions overlap each other between images of adjacent ground surfaces. This is to ensure that there is no space between the image and the image when the entire surface of the predetermined surface is photographed by mosaic processing connecting the image and the image.

On the other hand, if the ground surface to be photographed by the satellite 10 is referred to as a first ground surface and the ground surface adjacent to the first ground surface is defined as a second ground surface, the photographed image of the second ground surface can be assumed as a second image have. The second image is an image captured before the first image is captured when the satellite 10 sequentially captures the image. The second image will be described below assuming that the geometric correction device 100 has already estimated the information of the map coordinate point for the second image in a state where the geometric correction is completed before the geometric correction of the first image.

The minutia detection unit 140 may detect a predetermined object capable of indicating an accurate map coordinate in the overlapped portion (hereinafter referred to as a redundant image portion). For example, the predetermined object is a landmark capable of accurately indicating a coordinate point of a specific area of the map coordinate system such as a terrain, a coastline, a structure (hereinafter referred to as a first cover) on the ground surface. Specifically, the feature point detection unit 140 may determine whether an object common to the second image is detected in the overlapping image portion of the first image, and may set the detected common object as the feature point. A map coordinate point (hereinafter referred to as a map coordinate point) of the map coordinate system with respect to the first mark may be acquired through a previously stored database or may be acquired from map coordinates estimated from the second image. On the other hand, the map coordinate system may be a GEOS coordinate system, but it may be a coordinate system that can be used to express the ocean, weather, and topography of the earth, but the present invention is not limited thereto.

In addition, the predetermined object may include a cloud. The minutiae point detecting unit 140 may extract an object corresponding to at least one cloud shape common to the redundant image portion of the first image and the second image. The minutiae point detection unit 140 may set the extracted cloud shape (second marker) as the second marker. The map coordinate point of the second marker can be obtained from the coordinate point of the map coordinate system estimated in the second image.

The reference coordinate generating unit 150 can generate reference coordinates that are map coordinate points corresponding to pixels to be displayed on a predetermined object. The reference coordinates are used to calculate a relational expression between the image coordinates of the pixels of the image and the corresponding map coordinate points. The reference coordinate generator 150 selects a reference pixel, which is a reference pixel, and calculates a first reference coordinate, which is a map coordinate point corresponding to the reference pixel, using the estimated state vector. Also, the reference coordinate generator 150 calculates a second reference coordinate which is a map coordinate point corresponding to the first mark and the second mark. On the other hand, the reference coordinate generator 150 may acquire the map coordinate points corresponding to the first mark and the second mark through map coordinate information previously estimated in the second image.

The relational expression estimating unit 160 estimates a relational expression between an image position (hereinafter referred to as an image coordinate) of a pixel of the first reference coordinates and the second reference coordinates and corresponding map coordinate points. The relational expression is a relational expression for Rational Polynomial Coefficients (RPC) or Rational Function Model (RFM). The image geometry correcting apparatus 100 may calculate the map coordinate points of the first pixels excluding the pixels located in the reference coordinates using the relational expression.

For example, when the coordinates of the first reference coordinates and the coordinates of the second reference coordinates are (x, y) and the image coordinates are (u, v), the relational expression estimator 160 concatenates the two coordinates Can be obtained. The transformation between these coordinates is expressed in the form of x = au + bv + c, y = du + ev + f. In this case, all six unknowns are determined to determine the values of the coefficients in the above conversion equation, and a relational expression between the image coordinates of the pixel and the corresponding map coordinate points can be calculated.

The geometric corrector 170 utilizes the relational expression estimated by the relational expression estimator 160 as a geometric correction formula that can connect the image coordinates of the pixels of the image with the map coordinate system. The geometric corrector 170 can obtain the map coordinate points corresponding to the image coordinates of the remaining first pixels (the first pixels excluding the pixels having the first reference coordinates and the second reference coordinates) have. The geometry correcting unit 170 converts the first image to match the map coordinate system based on the obtained map coordinate points. In this case, the geometric corrector 170 finds a north-to-north direction of the first image and performs 2D transformation on the first pixels of the first image so as to match the map coordinate system. For example, the geometric corrector 170 performs a translation of an image, a rotation of an image, and a scaling of an image in accordance with the north-to-north direction of the image, and outputs the first image to a map coordinate system Can be matched.

The resampling unit 180 may resamplify the geometrically corrected image. The resampling unit 180 may calculate the size of the post-correction image (position after conversion of the four corner points of the pre-correction image), and then obtain an inverse conversion formula that is a relational expression of the corrected image and the pre-correction image. The resampling unit 180 finds each pixel of the first pixel of the image subjected to the geometric correction using the obtained inverse transformation formula to find the position of the pre-geometry-corrected image, rearranges the remaining pixels except for the first pixel, And resampling can be performed.

4 is a flowchart briefly illustrating an image geometry correction method according to an exemplary embodiment.

The flowchart shown in FIG. 4 is composed of the steps that are processed in a time-series manner in the image geometry correcting apparatus 100 shown in FIG. Therefore, even if omitted from the following description, it can be understood that the description described above with respect to the configurations shown in FIG. 3 also applies to the flow chart shown in FIG.

Referring to FIG. 4, the image correcting apparatus 100 may receive posture information of the satellite 10 at the time of acquiring the image and the image from the satellite. The image geometry correction apparatus 100 may sequentially receive the images sequentially captured by the artificial satellite 10. When the images are received sequentially, the posture information of the artificial satellite 10 corresponding to each image is also included Can be received. On the other hand, the attitude information of the artificial satellite 10 can be obtained by measuring the attitude of the artificial satellite 10 measured through a sun sensor, a geomagnetic sensor, a star sensor included in the artificial satellite 10, a gyro sensor serving as an inertial sensor, 10), a shooting angle of a video apparatus for shooting an image, and the like (S101).

The image geometry correcting apparatus 100 estimates the state vector of the artificial satellite 10 based on the attitude information of the artificial satellite 10. The image geometry correcting apparatus 100 may calculate a map coordinate point corresponding to previously selected reference pixels based on the state vector of the estimated satellite 10 and generate the first reference coordinate. Also, the image geometry correcting apparatus 100 detects a feature point, which is a predetermined object, in an image of a portion overlapping between the first image and the second image described with reference to FIG. 3, and calculates a map coordinate point corresponding to the feature point Can be generated as second reference coordinates (S103).

The image geometry correcting apparatus 100 can determine the position of the first image by calculating the correlation between the image coordinates of the first image and the map coordinate system. Specifically, the image geometry correcting apparatus 100 may derive a relational expression between the image coordinates of the image pixel and the map coordinates using the first reference coordinates and the second reference coordinates (S105).

The image geometry correcting apparatus 100 may perform geometric correction of the image that converts the first pixels included in the image into the map coordinate system described in FIG. 3 using the derived relation (S107).

When the geometric correction is completed, the image correction device 100 may resamplify the geometrically corrected image. The image geometry correcting apparatus 100 may calculate the size of the post-correction image for rearrangement of the pixels included in the first image, and then obtain an inverse conversion formula that is a relational expression of the corrected image and the pre-correction image. The image geometry correcting apparatus 100 performs resampling to obtain a new image by rearranging the remaining pixels except for the first pixel to find where each pixel of the geometrically corrected image is located in the pre-geometrically corrected image, (S109).

On the other hand, rearrangement methods using rearranged pixel values of pre-correction images include nearest neighbor, bilinear interpolation, and cubic convolution.

5 is a flowchart briefly illustrating a method of generating a first reference coordinate according to an embodiment.

The flowchart shown in FIG. 5 consists of the steps that are processed in a time-series manner in the image geometry correcting apparatus 100 shown in FIG. Therefore, it is understood that the contents described above with respect to the configurations shown in FIG. 3 apply to the flowchart shown in FIG. 5, even if omitted from the following description.

Referring to FIG. 5, the image geometry correcting apparatus 100 estimates a state vector used to generate the first reference coordinates. The image geometry correcting apparatus 100 can estimate the state vector of the satellite based on the attitude information of the satellite 10. [ The image geometry correcting apparatus 100 can estimate which coordinate point the state vector is directed to in the map coordinate system described with reference to FIG. 3, and this is called a pointing model (S111).

The image geometry correcting apparatus 100 generates a grid at a predetermined interval in the first image and selects a first pixel which is a pixel located at a lattice point at which two lines of the grid are intersected. For example, in the case of an image having 500 by 500 pixels, the pixel selection unit 130 generates grid lines orthogonal to each 100 pixels. In this case, the image geometry correcting apparatus 100 may calculate the image coordinates (100, 100, 200, 100, 300, 100, 400, 200, 100, 200, 200, 400, 300, 100, 300, 200, 300, 300, 400, 400, 100, 400, 200, 400, 300, May be selected as the first pixel (S113).

The image geometry correcting apparatus 100 may select at least one reference pixel from among the first pixels for estimating a corresponding map coordinate point using the estimated state vector. For example, the image geometry correcting apparatus 100 may select four reference pixels spaced apart at regular intervals among the first pixels. The image geometry correcting apparatus 100 may acquire information on a map coordinate point corresponding to at least one reference pixel using the state vector. In this case, the image geometry correcting apparatus 100 may generate a first reference coordinate including coordinates on the image of the at least one reference pixel and corresponding map coordinate points (S115).

Meanwhile, the image geometry correcting apparatus 100 may determine the number of reference pixels in consideration of the number of unknowns included in the relational expression of the image coordinates of the first pixel and the map coordinate points. For example, when six unknowns are to be found in obtaining the relational expression, at least three correct map coordinate points are required for the image correcting apparatus 100, so that at least three reference pixels are selected .

6 is a flowchart briefly illustrating a method of generating second reference coordinates according to an embodiment.

The flowchart shown in FIG. 6 is composed of the steps that are processed in a time-series manner in the image geometry correcting apparatus 100 shown in FIG. Therefore, it is understood that the contents described above with respect to the configurations shown in FIG. 3 apply to the flow chart shown in FIG. 6 even if omitted from the following description.

Referring to FIG. 6, the image geometry correcting apparatus 100 may detect the overlapped image portion described with reference to FIG. As described above, the image correcting apparatus 100 detects feature points such as landmarks and clouds only for the redundant image portion, not the entire image, so that the feature points on the image, such as landmarks for geometric correction, have. Accordingly, the image geometry correcting apparatus 100 may compare the first image, which is a photographed image, with at least one second image that previously captured the ground surface, to detect the overlapped image portion (S121).

In the meantime, the redundant image portion is defined as a first image portion in the first image, and the redundant image portion is defined as a second image portion in the second image.

The image geometry correcting apparatus 100 may compare the first image portion and the second image portion to extract a common feature point, which is a predetermined object. The feature mark includes a first mark, which is a landmark that can accurately tell the map coordinates of a specific area such as a terrain, a coastline, and a structure, on the ground surface, and a second mark that is a shape of a cloud common to the overlapping image portion (S 123) .

For example, the image geometry correcting apparatus 100 extracts edges of the first image portion and the second image portion, and matches an edge extracted from the first image portion and an edge extracted from the second image portion So that the matching edge can be detected. The image geometry correcting apparatus 100 can detect the shape and the like of the detected edge and find the first marker and the second marker corresponding thereto. For example, the image geometry correcting apparatus 100 can detect, from images, boundaries or edge edges of objects whose colors or the like discontinuously change in the image. The image geometry correcting apparatus 100 may detect edges of the images to extract a plurality of objects. Specifically, when the values of a specific pixel and neighboring pixels differ from each other in the image by more than a reference value, the image correcting apparatus 100 groups the corresponding pixels and detects groups that are equal to or greater than the reference number of pixels as an edge have. On the other hand, the image geometry correcting apparatus 100 may detect edges corresponding to a plurality of objects, and may select only objects of interest among a plurality of objects corresponding to the edges.

The image geometry correcting apparatus 100 may generate a second reference coordinate including a map coordinate point corresponding to each of the first marker and the second marker based on the first marker and the second marker. Specifically, the image geometry correcting apparatus 100 may store map coordinate points corresponding to the first mark in advance, or may acquire map coordinate points calculated from the second image. In addition, the image geometry correcting apparatus 100 may acquire a map coordinate point corresponding to the second mark through a map coordinate point calculated in the second image. The image geometry correcting apparatus 100 obtains image coordinate information of the pixels corresponding to the first and second covers and map coordinate points corresponding to the first and second covers by the method described above, And the second reference coordinates including the image coordinates information of the second cover and the information of the map coordinate points (S125).

On the other hand, the position of the cloud changes between the time when the image geometry correcting apparatus 100 picks up the first image and when the second image is taken. Therefore, the map coordinate point in the pixel where the cloud is located in the second image and the coordinate point in the pixel where the corresponding cloud is located in the first image may be different. However, the satellite 10 photographs the first image within one minute after capturing the second image, and when the cloud speed is 15 m / s during the one minute, the cloud travel distance is only about 1 km. Therefore, since the position of the cloud is hardly changed for one minute in view of the resolution of the photographing apparatus, the image geometry correcting apparatus 100 calculates the map coordinate point of the common cloud shape from the map coordinates corresponding to the map coordinate point estimated from the second image Value, it is possible to perform highly accurate geometric correction.

2, when the surface to be photographed by the satellite 10 is covered with a cloud, the cloud is taken as a feature point, and the map coordinate point in the second image having the corresponding feature point is used So that the error of the state vector of the satellite 10 can be corrected. For example, when the ground surface to be photographed by the satellite 10 is covered with a cloud and can not find a suitable landmark, it is difficult for the ground station to estimate the state vector by residual due to landmark matching. In this case, The generated first reference coordinates have a high error rate with respect to the actual map coordinate point.

In this case, the image geometry correcting apparatus 100 according to the embodiment can detect a specific cloud on the surface of the earth as a feature point. The image geometry correcting apparatus 100 may obtain a map coordinate point for the minutiae from a second image of a portion adjacent to the ground surface. That is, the image geometry correcting apparatus 100 can acquire relatively accurate map coordinate information for a common cloud in the first image from the second image, which is a previous image that was not covered with the clouds by the clouds. That is, the image geometry correcting apparatus 100 can directly correct the error of the state vector in which the error rate is increased, by obscuring the cloud through the relatively accurate coordinate information obtained from the second image.

7 is a flowchart schematically showing a method of geometric correction of a first image based on reference coordinates according to an embodiment.

The flowchart shown in FIG. 7 is composed of the steps that are processed in a time-series manner in the image geometry correcting apparatus 100 shown in FIG. Therefore, even though omitted from the following description, it can be seen that the above description regarding the configurations shown in FIG. 3 also applies to the flow chart shown in FIG.

Referring to FIG. 7, the image geometry correcting apparatus 100 uses the first reference coordinates and the second reference coordinates described with reference to FIG. 6 to calculate the image coordinate of the pixel and the map coordinate point of the map coordinate system described with reference to FIG. Estimate the relationship between the two. The relational expression is a relational expression for Rational Polynomial Coefficients (RPC) or Rational Function Model (RFM) (S131).

The image geometry correcting apparatus 100 may calculate map coordinate points corresponding to the first pixels located at the lattice points described with reference to FIG. 5 using the relational expression. That is, the image geometry correcting apparatus 100 can calculate the output of the corresponding map coordinate point according to the geometrical correction by inputting the image coordinates of the first pixel according to the relational expression (S133).

 The image geometry correcting apparatus 100 may perform geometric correction to coordinate transform the first pixels to correspond to the calculated map coordinate point. Specifically, the image geometry correcting apparatus 100 may convert the first pixel of the first image into a map coordinate system by moving it by a two-dimensional transformation (movement, rotation, scaling, etc.) (S135).

 The image geometry correcting apparatus 100 can again obtain an inverse transformation formula which is a relational expression of the corrected image and the pre-corrected image. The image geometry correcting apparatus 100 performs resampling to obtain a new image by rearranging the remaining pixels except for the first pixel to find where each pixel of the geometrically corrected image is located in the pre-geometrically corrected image, (S137).

The embodiments of the present invention described above can be implemented in the form of program instructions that can be executed through various computer components and recorded in a computer-readable recording medium. The computer-readable recording medium may include program commands, data files, data structures, and the like, alone or in combination. The program instructions recorded on the computer-readable recording medium may be those specifically designed and configured for the present invention or may be those known and used by those skilled in the computer software arts. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape, optical recording media such as CD-ROM and DVD, magneto-optical media such as floptical disks, medium, and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code, such as those generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware device may be modified into one or more software modules for performing the processing according to the present invention, and vice versa.

The specific acts described in the present invention are, by way of example, not intended to limit the scope of the invention in any way. For brevity of description, descriptions of conventional electronic configurations, control systems, software, and other functional aspects of such systems may be omitted. Also, the connections or connecting members of the lines between the components shown in the figures are illustrative of functional connections and / or physical or circuit connections, which may be replaced or additionally provided by a variety of functional connections, physical Connection, or circuit connections. Also, unless explicitly mentioned, such as " essential ", " importantly ", etc., it may not be a necessary component for application of the present invention.

Accordingly, the spirit of the present invention should not be construed as being limited to the above-described embodiments, and all ranges that are equivalent to or equivalent to the claims of the present invention as well as the claims .

10: Satellite
100: image geometry correction device
110:
120: state vector estimator
130:
140:
150: reference coordinate generating unit
160: Relational expression estimating unit
170: Geometrical Correction
180: resampling unit
200: Earth
210: Equator

Claims (22)

An image geometry correction method for geometrically correcting an image photographed by a satellite with a map coordinate system for an earth surface of the earth,
Receiving attitude information and a first image of a satellite from a satellite;
Generating a grid at regular intervals in the first image and extracting a first pixel that is a pixel of the first image located at a lattice point on the grid;
Detecting common feature points between the first image and the second image;
Generating reference coordinates of a map coordinate system based on the attitude information and the minutiae; And
And a geometric correction step of estimating a map coordinate point, which is a coordinate point of the map coordinate system corresponding to the first pixel, based on the reference coordinates, and geometrically correcting the first image based on the estimated map coordinate point Image Geometric Correction Method. The method according to claim 1,
Wherein the second image is an image of a surface of the ground adjacent to the ground surface on which the first image is captured. The method according to claim 1,
And estimating a state vector of the first image based on the attitude information. The method of claim 3,
Wherein the step of generating the reference coordinates comprises:
Selecting at least one reference pixel among the first pixels; And
Detecting a coordinate point of the map coordinate system corresponding to the at least one reference pixel based on the state vector of the first image and generating a first reference coordinate including the coordinate point of the detected map coordinate system; Further comprising an image geometry correction method. 5. The method of claim 4,
Wherein the extracting of the feature points comprises:
Detecting a redundant image that is an image portion overlapping the second image in the first image;
And extracting a preset object common to the second image from the overlapped image. 6. The method of claim 5,
Wherein the predetermined object includes a first marker that is at least one of a terrain surface, a shoreline, and a structure. The method according to claim 6,
Wherein the preset object further comprises a second marker that is a cloud on the surface of the earth. 8. The method of claim 7,
Wherein the step of generating the reference coordinates comprises:
Acquiring a coordinate point of the map coordinate system corresponding to the first mark and the second mark from the second image and generating a second reference coordinate including the coordinate point of the acquired map coordinate system and,
Wherein the reference coordinates include the first reference coordinates and the second reference coordinates. 9. The method of claim 8,
Wherein the geometry correction step comprises:
Estimating a relational expression of the reference coordinates and an image position on the corresponding pixel; And
And estimating a map coordinate point of the selected first pixel using the estimated relational expression. The method according to claim 1,
Wherein the second image is a coordinate-converted image of the map coordinate system. 1. An image geometry correction apparatus for geometrically correcting an image photographed by a satellite with a map coordinate system for an earth surface of the earth,
A receiving unit receiving first image and attitude information from a satellite;
An estimation object selection unit which generates a grid at a predetermined interval in the first image and selects a first pixel which is a pixel of the first image located at a lattice point on the grid;
A feature point detector for detecting common feature points between the first image and the second image;
A reference coordinate generator for generating reference coordinates of a map coordinate system based on the attitude information and the minutiae points; And
And a geometric correcting unit for geometrically correcting the first image by estimating a map coordinate point, which is a coordinate point of the map coordinate system corresponding to the first pixel, based on the reference coordinates. 12. The method of claim 11,
Wherein the second image is an image of a surface of the ground adjacent to the ground surface on which the first image is captured. 13. The method of claim 12,
And a state vector estimating unit estimating a state vector of the first image based on the attitude information. 14. The method of claim 13,
Wherein the reference coordinate generator selects at least one reference pixel among the first pixels,
And detects a coordinate point of the map coordinate system corresponding to the at least one reference pixel based on the state vector of the first image and generates a first reference coordinate including the detected coordinate point. 15. The method of claim 14,
Wherein the feature point detection unit detects a redundant image that is an image portion overlapping the second image in the first image and extracts a predetermined object common to the second image in the redundant image. 16. The method of claim 15,
Wherein the predetermined object includes a first marker that is at least one of a terrain surface, a shoreline, and a structure. 17. The method of claim 16,
Wherein the predetermined object further comprises a second marker that is a cloud on an earth surface. 18. The method of claim 17,
Wherein the reference coordinates generator obtains information on coordinate points of the map coordinate system corresponding to the first mark and the second mark from the second image and generates second reference coordinates including coordinate points of the obtained map coordinate system / RTI >
Wherein the reference coordinates include the first reference coordinates and the second reference coordinates. 19. The method of claim 18,
And a relational expression estimating unit that estimates a relational expression between the reference coordinates and a position of an image on the corresponding pixel. 20. The method of claim 19,
And the geometric correction unit estimates a map coordinate point of the selected first pixel using the estimated relational expression. 12. The method of claim 11,
Wherein the second image is a coordinate-converted image of the map coordinate system. A computer-readable recording medium on which a program for executing the method according to any one of claims 1 to 10 is recorded.

Images