KR19990047500A - Semi-automatic online geocorrection method - Google Patents

Abstract

The present invention relates to a method for correcting geometric distortion of an image collected by a satellite and an aircraft. The present invention relates to a method of initializing a digitizer device by matching a coordinate system between a digitizer capable of reading coordinates by attaching a reference map and a reference map. Step 1, the second step of obtaining a correction conversion equation using four ground control points (GCP) corresponding to the image to be corrected and the reference map, selects a random point on the map and the above The new GCP obtained in the third and third stages of calculating the corresponding point of the image by applying the inverse correction transformation formula obtained in step 2, zooming-up and displaying the portion of the image on the screen, and selecting and adding a new GCP. The fourth step of repeating the third step until the correction conversion formula for the error is within the allowable range of the RMS error, and the RMS error is satisfied in the fourth step In this case, the image is corrected using the correction conversion equation at that time, and the fifth and fourth steps of inputting the upper, lower, and lower geographic coordinates into the corrected header of the corrected image include the corrected coordinates. The function is integrated into a convenient user interface using X / Motif, so that accurate and efficient GCP setting and continuous calibration can be performed without depending on the user's skill in using the device.

Description

Semi-automatic online geocorrection method

The present invention relates to a method for correcting a geometrical distortion that an image collected by a satellite or an aircraft has with a reference map using a GCP, and particularly, a method for implementing a convenient user interface and a semi-automated GCP setting task. It is about.

In general, the remote sensing data observed through satellites or aircrafts are subject to a lot of errors and distortions depending on the state of the device, atmospheric conditions, direction and attitude of the platform, and the projection of the map to be used. It is included.

In the process of processing most of the remote sensing data, the following steps are first performed after restoring or correcting the error and distortion. This process of correcting the distortion of the image itself before performing the full-scale processing is called preprocessing.

The preprocessing process may include: an atmospheric correction process for restoring an image by removing an air influence, a noise removing process for removing periodic or non-periodic noise generated during image collection, and If there is a difference between the geometric shape of the collected image and the reference map, a geometric correction or the like is mainly performed.

As described above, the geometric correction is a process of correcting the geometric distortion that the image originally has. This correction includes system correction processing to correct the cause of errors and distortions, and ground control point (GCP) correction using ground control points (GCP). Processing is used.

The system correction analyzes the causes of all distortions such as elevation change and shaking of the observation platform, optical characteristics of the observing device, the ups and downs of the terrain, the rotation of the earth, and the projection of the earth on the reference map. It is a method of correcting distortion by obtaining an inverse transformation system that restores a distorted image to its original state using the analysis result.

The system correction has an advantage that the distortion of all images included in the same system can be easily corrected only by accurately analyzing all the causes of the distortion and obtaining an inverse transform system. However, it is not easy to analyze the cause of all the distortions, and when the terrain is undulating or the resolution of the image is high, there is a disadvantage that precise correction is difficult.

The ground reference point correction only analyzes the degree of distortion without considering the cause of the distortion. Using the analysis result, a correction equation for connecting the collected image and the reference map is obtained, and the distortion is corrected by using the correction equation.

Here, the ground reference point is a coordinate from which a point having the same shape is extracted from the image and the map in order to fit the collected image with the reference map. The ground reference point means a control point used to obtain a correction equation.

The distortion correction method using the ground reference point may be used even when the cause of the distortion is unknown or difficult to analyze, and if only the correct ground reference point is selected, the image correction may be more accurate than the system correction. However, there is a disadvantage that the correction result greatly varies depending on the selection of the ground reference point. That is, the correction using the ground reference point is highly dependent on the skill of the operator selecting it. Therefore, the result may vary depending on the individual, and even if the observation date changes even in the same region, it is cumbersome to obtain it again. For this reason, many studies have been conducted to automatically obtain the coordinates of the ground reference point.

Conventionally known methods of automatically obtaining the coordinates of the ground reference point, there is a method of automatically extracting the ground reference point or a geometrical feature using a GCP chip (Chip) to match. The use of GCP chips is particularly common. GCP chip is a kind of image that records the ground coordinates of this point and the alternative coordinates in the scene by extracting a certain region that can be clearly distinguished from other regions in the image and whose position or brightness does not change according to time change. It is a database.

Such a GCP chip is used to find an area where characteristics match when a new image is input by an automatic matching method. Using this method, it is known that the coordinates can be extracted with sub-pixel precision.

However, this method has the disadvantage that it cannot be used if it is not a pre-database area.

Accordingly, the present invention is to provide an apparatus capable of semi-automating the ground reference point setting of the geometric correction required in the satellite and aerial image that will include the large-capacity and wide area to be spread in the future.

According to the present invention, if only the first four ground reference points are set, the following ground reference points can be generated, and accurate correction can be performed by using the following ground reference points. It is also intended to facilitate the operation by a convenient user interface.

The present invention focuses on the fact that (N + 2) * (N + 1) / 2 is required when the order of the conversion equation is N as the minimum number of ground reference points required to calculate the conversion equation. When setting the next GCP in the reference map on the digitizer, which is relatively easy to set, the GCP of the corresponding image is calculated by using a previously calculated conversion formula including an error, and drawn on the image screen as expected coordinates. give.

You can add a new GCP by selecting around or just this coordinate. In addition, by applying the conversion equation to the image to be corrected in real time, it is possible to confirm the corrected result and to read the precision. The whole process uses a convenient user interface implemented in X / Motif, most of which is easy to use with buttons and mouse.

Means for achieving the present invention, the first step of matching the coordinate system of the digitizer and the reference map for reading the coordinates of the map by attaching the reference map, and the ground reference point corresponding to the image to be corrected and the reference map (GCP: Ground Control Point) is selected in the second step to obtain a correction conversion equation, and an arbitrary point of the reference map is selected, and using the inverse correction conversion equation of the correction conversion equation obtained in the second step to the point A third step of obtaining an expected point of the corresponding image and outputting an area including the expected point to the screen, and further selecting a ground reference point by selecting a point of the image corresponding to the arbitrary point on the screen; Obtain the calibration conversion equation based on the ground reference point selected in step 3 and check whether it is within the allowed RMS error range, In this case, the fourth step of repeating the third step of adding the new ground reference point, and if the RMS error falls within the allowable range, correct the image by the correction conversion equation and specify and rearrange the output coordinates of the corrected image A semi-automatic on-line geographic correction method is constructed, including a fifth step of converting options to output a corrected image.

All processes are integrated using X / Motif using a convenient user interface, allowing accurate and efficient GCP setup and continuous calibration without relying on the user's skill in using the device.

1 is an illustration of a semi-automatic online geographic correction system for applying the present invention.

2 is an overall flowchart of a semi-automatic online geographic correction method according to the present invention.

Figure 3 is a digitizer and map matching flow chart according to the present invention.

4 is an explanatory diagram of a GCP setting of an image to be corrected and a digitizer according to the present invention;

5 is a flowchart illustrating a new GCP addition setting according to the present invention.

6 is a flowchart illustrating a rearrangement option setting according to an embodiment of the present invention.

<Description of the code | symbol about the principal part of drawing>

100: Unix Environment Workstation 110: Digitizer

120: keyboard 130: mouse

140: CD-ROM drive 150: color monitor

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

FIG. 1 is a view illustrating a connection of a digitizer having a Calcomp 24000 × 36000 resolution connected to a sun workstation which is a desktop of the present invention. The keyboard 120, the mouse 130, the CD-ROM drive 140, the color monitor 150, and the like are equipped with a workstation 100 in a Unix environment equipped with a peripheral device and a reference map. It is configured to include a digitalizer 110 for reading the coordinates of. Currently, a fixed drive device is constructed by connecting to a serial port in a Unix environment and fixed in a digitizing receiving mode only in point mode. The pointer device used is a wireless cursor and has 16 input buttons.

FIG. 2 is a flowchart of a semi-automatic online geographic correction method according to the present invention. As shown in FIG. 1, the first step is to determine a map coordinate system option (S1) and match a coordinate system between a digitizer and a reference map ( S2) In the second step, the image to be corrected is selected and loaded (S3), and the point of the image corresponding to the point on the map is selected by the GCP (S5). It is checked whether the GCP number is four (S4), and a correction conversion equation is calculated (S6). The third step is to apply the inverse correction conversion formula of the correction conversion formula to the arbitrary point on the map to find the point of the image and output the part, and to find the point corresponding to the arbitrary point on the map in the output image, the new GCP Select further (S8). The fourth step checks whether the RMS error range is satisfied by calculating a correction conversion formula including the new GCP (S6) (S7). Repeat the process of selecting the new GCP to obtain the calibration conversion formula until the RMS error range is satisfied. In the fifth step, when the RMS error range is satisfied, the image is corrected using the correction conversion equation (S9), and the image is output by changing the rearrangement option for the corrected image (S10).

3 illustrates a first step of matching a reference map to be attached with a coordinate system of a digitizer in order to use a digitizer for reading a coordinate system of a reference map according to the present invention.

The user straightens and anchors the reference map on the digitizer. The map coordinate system selection is determined, the earth type and the main coordinate system of the map are selected, and the zone axis zone number of the map area and the zone transverse zone number of the map area are input (S31).

Subsequently, a matching GCP is selected for matching the digitizer and the map (S32), and it is checked whether the number of matching GCPs is four (S33). If the number of matching GCPs is not four, the coordinates of four points on the map and four points of the corresponding digitizer are selected as the matching GCPs.

In order to select the matching GCP, one corner point on the map is selected (S34). This selects a point, or corner, on the map attached to the digitizer using the digitizer's cursor. At this time, the position of the selected point is displayed in the virtual digitizer screen displayed on the screen (S35). The coordinates of this selection point are read from a map and input as map coordinates (S36). A map coordinate, a digitizer coordinate value and a digitizer coordinate value of the selection point are selected as a junction GCP value (S37).

It is determined whether the selected matching GCP satisfies the allowable range of the matching RMS error (S38). If not, the new selection selects the matching GCP, and if it is satisfied, the matching result is determined (S39).

4 is a flowchart illustrating a GCP selection process for map matching with an image to be corrected according to the present invention, which represents a second step.

In the second step, GCP settings of the image to be corrected and the digitizer are made. Select one point on the map of the matched digitizer and one point on the image to be corrected and set it as the ground reference point. The ground reference point selects (N + 2) * (N + 1) / 2 ground reference points when the conversion order is N. For example, in the first case, four ground control points are needed. It is difficult to trust the result of the ground control point setting of the first four or less because the automatic conversion is impossible because there is no correction conversion formula calculated previously. When the first four ground control points are set, a correction equation involving the RMS error is calculated.

The GCP is selected to match the image to be corrected with the map to check whether the number of GCPs is 4 (S42). In the method of selecting the GCP, a point on the map attached to the digitizer, the intersection point or the end of the leg, which is well seen in the corresponding image, is selected using the cursor of the digitizer (S43). The position of the selection point is displayed on the virtual digitizer screen (S44). A point corresponding to the selection point is found in the image (S45). The selection point of the corresponding image is clicked (CLICK) and selected (S46). The coordinate value on the map and the image file coordinate value are selected by GCP (S47), and a correction conversion equation is obtained (S48). In this way, a correction conversion equation is obtained for the four GCPs, and an inverse correction conversion equation is obtained (S49) to prepare for the next step.

5 shows a third step and a fourth step of the present invention. In the third step, a good point to select on the map is selected (S51). The user selects an arbitrary point of a road intersection, a river, a dam, etc. that is relatively easy to select on a map on a reference map with a mouse. The point of the image corresponding to the selection point is automatically calculated by applying the inverse correction conversion equation calculated above to the selected point (S52). A rectangle of hidden lines is displayed around the point including the calculated point on the image and output to the screen (S53). The portion to be displayed on the screen zooms up the region of the image and the reference map, that is, the rectangle represented by the hidden line and the periphery thereof to the center portion of the screen. The user finally selects the corresponding point of the image corresponding to the reference map by examining only the point or the surrounding of the output image, and adds it as another ground reference point (S54).

When the ground reference point addition setting of the image and the reference map is completed, the fourth step checks whether the correction conversion equation for the newly added GCP satisfies the allowable range of the RMS error (S55). If the allowable range is not satisfied, a new GCP is additionally selected to repeat the third step from the new point selection (S51) on the map until the allowable range of the RMS error is satisfied.

In this way, when the GCP is added to satisfy the RMS error, the rearrangement button is pressed to correct the image to be corrected by the correction conversion equation (S56).

And, the fifth step is as shown in FIG.

A variety of rearrangement options are available to create the corrected image, allowing the user to create a result file with different selection options. Optionally, the portion corresponding to the desired area of the reference map may be selected, the resolution ratio between the reference map and the image to be corrected, and the size of the image to be output may be selected. First, a corrected image is output by applying a correction conversion formula to a corrected image (S61). The corrected image can be obtained by applying the entire image or only a desired portion. At this time, the coordinate value is extracted by specifying the upper left and lower right values as coordinates on a map of a desired region (S62). In addition, a rearrangement method of an image to be corrected is selected (S63). Accordingly, the rearrangement size, which is the resolution of the image to be output, is automatically calculated based on the size of the image to be corrected and the coordinate value on the map corresponding thereto (S64). In addition, when the 'OK' button, which is an execution button, is pressed to output the corrected image, the corrected image is output (S65).

As described in detail above, the present invention has an effect of semi-automating the ground reference point setting operation of the geometric adjustment operation required in the image collected using the satellite and the aircraft. In addition, by applying the conversion formula to the image to be adjusted in real time, it is possible to check the corrected result and to read the density, and the entire process is easily performed with the button and mouse using the convenient user interface implemented with X / Motif. There is an effect available.

Claims (4)

A first step of matching a coordinate system between a digitizer and a reference map for reading the coordinates of the map by attaching the reference map; A second step of selecting four ground control points (GCP) corresponding to the image to be corrected and the reference map to obtain a correction conversion equation; Selects an arbitrary point of the reference map, obtains an expected point of an image corresponding to the point by using an inverse correction conversion equation of the correction conversion equation obtained in the second step, and outputs an area including the expected point on the screen; A third step of further selecting a ground reference point by selecting a point of an image corresponding to the arbitrary point on the screen; In the third step, if the correction conversion equation based on the ground reference point additionally selected is obtained, and it is within the allowable RMS error range, the third step of adding the new ground reference point is repeated. The fourth step, And a fifth step of correcting the image by a correction conversion equation, converting output coordinate designation and rearrangement options of the corrected image, and outputting the corrected image if the RMS error falls within the allowable range. Geographic correction method. The method of claim 1, wherein the first step, Selecting the earth type and the main coordinate system of the map, inputting the zone axis zone number of the map area and the zone transverse zone number of the map area, A matching CGP selection step of straightening the reference map on the digitizer to fix it, selecting four points on the digitizer and inputting the coordinates of the corresponding map coordinate system; Determining whether the coordinates of the four points and the corresponding map coordinate system are within an allowed RMS range, Repeating the process of determining the RMS error range by selecting four points on the digitizer and inputting the corresponding map coordinate system until it is within the allowed RMS range, and terminating the initialization when it falls within the allowed RMS range. Semi-automatic online geographic correction method, characterized in that. The method of claim 1, wherein the second step, Select one point on the reference map and one point on the image to be corrected to set the ground reference point (GCP), If the ground reference point is of order N of the correction conversion equation, (N + 2) * (N + 1) / 2 Semi-automatic online geocorrection methods characterized by selecting two ground reference points. The method of claim 1, wherein the third step, A step in which the user selects an arbitrary point such as an intersection of a road, a river, a dam, etc., which is easily distinguished from surrounding points on a reference map on the digitizer, Generating a ground reference point of the image corresponding to the selection point by applying the correction conversion equation calculated in the second step to the selected point in reverse; Zoom-up and outputting the area including the newly selected ground control point to the center of the screen; Semi-automatic online geographic correction method characterized in that the user finally selects the point of the image corresponding to the arbitrary point of the reference map selected in the first step in the screen, and adds it as another ground reference point .