JP2006148205A - Method of calibrating optical sensor mounted on satellite, and image pickup device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain a calibration method for acquiring a high accuracy with the small number of pixels by changing direction of a linear CCD in executing calibration of a gain and the offset correcting coefficient of an optical line sensor to be equipped on an artificial satellite. <P>SOLUTION: When the satellite equipped with the optical sensor 2 of a line sensor system executes the calibration of the gain and offset correction coefficient, images are picked up in a state where the direction of arrangement of pixels 5 of a footprint 7 in an image pickup target of each of the pixels 5 in the optical sensor 2 may almost match the direction of movement of the artificial satellite 1. Thus, the calibration of the gain and offset correction coefficient is executed. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a calibration method for gain, offset, and the like of an optical sensor mounted on an artificial satellite, and an imaging apparatus using the same.

A line sensor is known as an optical sensor system that is mounted on a satellite and acquires an image. The line sensor has a linear detector composed of a large number of pixels, and is arranged so that the arrangement of pixels projected onto the observation object is substantially orthogonal to the satellite moving direction.
In the line sensor, the observation times of all the pixels included in one scanning line coincide with each other, and two-dimensional image data is acquired by using the movement of the satellite.

  Since the gain and offset of each pixel differ from pixel to pixel, striped noise appears in the acquired image data in the scanning direction (movement direction). Usually, a gain and offset database for each pixel is created using a uniform light source such as an integrating sphere at the time of development of an optical sensor, and correction processing for removing the stripe noise is performed using the database during image acquisition.

  However, the gain and offset may change due to environmental conditions of the sensor, changes over time, etc., and sufficient correction cannot be performed with a database acquired in advance, and stripe noise may occur. In this case, it is disclosed that an output average value, standard deviation, cumulative histogram, etc. of each pixel are obtained using acquired image data, and output correction coefficients such as gain and offset are calibrated so that these statistical values are equal. (See, for example, Non-Patent Document 1).

“Image Analysis Handbook”, supervised by Mikio Takagi and Yoshihisa Shimoda, University of Tokyo Press, November 10, 1995, 6th edition, (p448-p450)

In the conventional calibration method, it is a precondition that the luminance frequency distribution in the imaging target area is the same for each pixel.
In the line sensor, since each pixel scans a different position of the object, the luminance frequency distribution of each pixel generally differs with a small number of data such as one scene.
However, since it can be assumed that the frequency distribution of luminance is the same for every pixel if an infinite number of image data is used, the above-described calibration is performed for the line sensor, but the number of images that can actually be used is limited. There was a problem that high calibration accuracy could not be obtained.

  The present invention has been made to solve such a problem, and an object of the present invention is to obtain a calibration method for obtaining high accuracy with a small number of images by changing the direction of a linear CCD.

  The satellite-mounted optical sensor calibration method according to the present invention is such that, when a satellite equipped with a line sensor optical sensor performs calibration of gain and offset correction coefficients in each pixel of the line sensor, The corresponding imaging target direction and the movement direction of the artificial satellite are substantially matched, and the gain and offset are calibrated based on the output level acquisition data of each pixel detected by the line sensor at different times.

  An imaging apparatus according to the present invention includes an optical sensor that is mounted on an artificial satellite and captures an image of the surface of the earth, and a gimbal mechanism having a drive unit that controls the posture of the optical sensor so that the optical axis is always directed to the earth. A rotation mechanism that rotates the attitude of the optical sensor around an axis in the direction of the earth at the time of calibration and a satellite equipped with the optical sensor perform calibration of gain and offset correction coefficients in each pixel of the line sensor. The image capturing direction corresponding to the scanning line of the line sensor and the moving direction of the artificial satellite are generally matched, and the gain and offset are calibrated based on the acquired data of the output level of each pixel detected by the line sensor at different times. And a data analyzing device.

  At the time of calibration of the line sensor, by changing the direction of the line sensor at the time of imaging by 90 degrees, there is an effect that the sensitivity and offset of the line sensor can be calibrated with high accuracy based on data of a small number of images.

Embodiment 1 FIG.
FIG. 1 is a diagram for explaining an imaging state according to Embodiment 1 of the present invention. 1 is an artificial satellite, 2 is an optical sensor, 3 is an imaging target, 4 is an optical system, 5 is a pixel, 6 is a line sensor, Reference numeral 7 denotes a footprint representing an imaging target of each pixel of the line sensor.

The optical sensor 2 mounted on the artificial satellite 1 has an optical system 4 that generates an image of the imaging target 3. A line sensor 6 having a plurality of pixels 5 is installed on the image plane of the optical system 4.
The moving direction of the footprint 7 accompanying the movement of the artificial satellite 1 is indicated by an arrow.
The output of the line sensor 6 is sequentially read according to the imaging cycle, transmitted to the ground via the artificial satellite 1, and subjected to image processing. (Details will be described later with reference to FIG. 6).

  When normal observation is performed, the pixel arrangement direction of the footprint 7 of the line sensor 6 in the imaging target is set to a direction substantially orthogonal to the moving direction of the artificial satellite 1.

  FIG. 2 is a diagram for explaining a calibration method according to Embodiment 1 of the present invention, and 4 to 7 are the same as those in FIG. The perspective view shows the relationship between the pixel alignment direction of the footprint 7 at the time of calibration and its moving direction.

  When the calibration method according to the first embodiment is performed, the orientation of the artificial satellite 1 is rotationally controlled or the optical sensor 2 is rotated, and as shown in FIG. The moving direction of the satellite 1 is matched.

The arrow indicates the movement direction of the footprint 7 with the movement of the artificial satellite 1.
As is clear from FIG. 2, each pixel 5 observes the same part of the imaging target as the footprint 7 moves, and therefore the luminance frequency distribution of the imaging target is the same for each pixel.
Therefore, high correction accuracy can be achieved with a small amount of image data.

It should be noted that the direction of alignment corresponding to the pixels 5 of the footprint 7 and the moving direction of the artificial satellite 1 may not completely match. If an error occurs, each pixel 5 cannot observe the same part, but can observe a neighboring part.
Since natural scenes to be observed have high spatial redundancy and generally neighboring luminance values are close to each other, each pixel 5 can obtain the same luminance frequency distribution with less image data than the conventional example.

  FIG. 3 is a diagram showing the state of the ground image captured by the optical sensor 2 during imaging according to Embodiment 1 of the present invention.

  When the artificial satellite 1 moves in the direction of the arrow, the footprint 7 simultaneously imaged by the line sensor 6 represents the locus and the direction of the footprint 7 that can be generated when the artificial satellite 1 moves over time. Is.

  Since the moving direction of the artificial satellite 1 and the footprint 7 of the line sensor 6 are perpendicular, a two-dimensional image can be easily obtained as the artificial satellite 1 moves.

  FIG. 4 is a diagram showing the state of the ground image taken by the optical sensor 2 during calibration according to Embodiment 1 of the present invention.

  When the artificial satellite 1 moves in the direction of the arrow, the footprint 7 simultaneously imaged by the line sensor 6 represents the locus and the direction of the footprint 7 that can be generated when the artificial satellite 1 moves over time. Is.

Since the movement direction of the artificial satellite 1 and the footprint 7 of the line sensor 6 are in the same direction, if one pixel of the footprint 7 is noticed as the artificial satellite 1 moves, the position of the same pixel is It can be easily seen that the satellite passes after the movement.
It is obvious that the output correction coefficient for each pixel of the line sensor 6 can be easily calibrated using this characteristic.

  FIG. 5 is a diagram for explaining the relationship between pixels used for calibration. L1 to L4 represent footprints 7 in each cycle of the line sensor 6. Actually, images are taken on the same line as shown in FIG. However, for the sake of easy understanding, the description is shifted downward as time passes.

  Here, n represents the number of pixels corresponding to the amount of movement of the artificial satellite 1 during each period. In addition, i represents an arbitrary pixel 5 intended for an arbitrary location, and the one pixel 5 is i−1, which indicates that i-2 and i-3 are sequentially continued. Yes.

  In order to calibrate the gain and offset correction coefficients, data of the video level value for each pixel 5 of the line sensor 6 is collected, and after a time lapse of an integral multiple of the cycle of the line sensor 6, the artificial satellite 1 The data of the video level value of the other pixels 5 of the line sensor 6 when moving and imaging the same place can be collected, and the calibration coefficient of each pixel 5 of the line sensor 6 can be calculated.

  Assuming that the input L (luminance) to one pixel 5 and the output X are linear, the output X can be expressed by the relationship as shown in Expression (1) at the input L.

each value in the i-th pixel 5 _{X i,} shall be expressed by _{a i, b i, L i} .
The relationship between the output X and the input L when the i th place is imaged in each of the lines L1 to L4 in FIG. 5 can be expressed by equation (2), and the i−1 th line in each of the lines L1 to L3. The time when the place is imaged is Expression (3), and similarly, Expression (4) and Expression (5) are established.

In addition, since the i-th to i-3th places are very close places, assuming that they are almost equal, Expression (6) is established.
Here, it is possible to calculate the offset a _i and the gain b _{i of} each pixel 5 by calculating the equations (2) to (6).

  In order to calibrate the gain and offset correction coefficients, data of video level values of a plurality of pixels 5 of the line sensor 6 are collected, and the image of the same place is imaged over time. Data of average values of video levels of other plural pixels 5 can be collected, and the calibration coefficient of each pixel 5 of the line sensor can be calculated.

  It is also possible to obtain the output average value, standard deviation, cumulative histogram, etc. of each pixel 5 using the acquired image data and calibrate output correction coefficients such as gain and offset so that these statistical values are equal. is there.

  Even if the line sensor 6 is a TDI (Time Delay & Integration) type sensor, the same effect can be obtained.

Embodiment 2. FIG.
FIG. 6 is a diagram for explaining a block diagram according to Embodiment 2 of the present invention. 8 is a gimbal mechanism, 9 is a rotation mechanism, 10 is a data processing device, 11 is a data recording device, and 12 is a data analysis device. 2 is the same as the description of FIG.

  The optical sensor 2 is a camera based on the line sensor 6, converts ground image data obtained by photographing an image of the ground surface into digital data, and transfers the digital data to the data processing device 10.

  By installing and integrating the gimbal mechanism 8 in the same housing as the optical sensor 2, the attitude of the optical sensor 2 at the time of shooting is tilted in three axes (ω, φ, u: roll angle, pitch angle, yaw angle. The rotation mechanism 9 is rotated 90 degrees in the yaw angle direction when the optical sensor 2 is calibrated.

  In the present invention, the gimbal 8 is installed in the body of the artificial satellite 1 by GPS (Global Positioning System) installed in the body, and the position of the optical sensor 2 at the time of photographing is obtained as position information in the three-dimensional space.

The data processing device 10 controls the gimbal mechanism 8 and the rotation mechanism 9 by sending drive signals, and sends the image data from the optical sensor 2 to the data recording device 11.
The data recording device 11 transfers image data to a data analysis device 12 which is a ground facility, and performs detailed analysis there.

  7 is a diagram for explaining the control mechanism of the image pickup apparatus according to the second embodiment of the present invention. 2 is the same as the description of FIG.

  When carrying out calibration according to the present invention, the attitude of the artificial satellite 1 is controlled to rotate, and instead of controlling the attitude of the artificial satellite, the image of the object to be imaged by the optical sensor 2 is rotated using the rotation mechanism 9. is there.

  The optical sensor 2 and the gimbal mechanism 8 are integrally attached, and the angle sensor measures based on attitude data (roll angle, pitch angle, yaw angle) of the optical sensor 2 with respect to the gimbal 8 measured by the angle sensor of the gimbal mechanism 8. The sensor 2 provided in the gimbal 8 is rotationally driven so that the roll angle, pitch angle, and yaw angle coincide with the target angle. A voltage for controlling the motors (M1, M2, M3) of the three-axis attitude control mechanism is output, and control is performed so that the direction of the optical sensor 2 matches the earth direction.

  When the artificial satellite 1 equipped with the optical sensor 2 flies in an ideal course (constant altitude / constant direction), the optical sensor 2 always faces the earth, and a ground image free from data jumping and distortion can be obtained.

  However, since it is impossible for the artificial satellite 1 to fly on an ideal course, the optical sensor 2 always faces the earth direction at the time of photographing so that the optical sensor 2 is orthogonal to the moving direction of the artificial satellite 1. In addition, it is necessary to control the attitude of the optical sensor 2 in real time by controlling in three axis directions.

  If the attitude data at this time is recorded together with the image data without performing this control, the satellite 1 will deviate from the target ground area when the aircraft 1 is tilted so that the satellite 1 flies over a certain course. In other words, the image cannot be corrected later in the computer program in the area that could not be photographed.

  As described above, the optical sensor 2 detects the line sensor at different times by substantially matching the imaging target direction corresponding to the scanning line of the line sensor and the moving direction of the artificial satellite as in the first embodiment. Based on the acquired output level data of each pixel, there is an effect that after performing calibration, imaging is performed, and stable imaging can be obtained using the control mechanism of the imaging apparatus.

It is a figure for demonstrating the imaging state by Embodiment 1 of this invention. It is a figure for demonstrating the calibration method by Embodiment 1 of this invention. It is a figure which shows the state of the surface image which the optical sensor at the time of imaging image | photographs. It is a figure which shows the state of the surface image which the optical sensor at the time of calibration image | photographs. It is a figure for demonstrating the relationship of the pixel used for calibration. It is a figure for demonstrating the block diagram of the imaging device by Embodiment 2 of this invention. It is a figure for demonstrating the control mechanism of the imaging device by Embodiment 2 of this invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 Artificial satellite, 2 Optical sensor, 3 Imaging object, 4 Optical system, 5 Pixel, 6 Line sensor, 7 Line sensor footprint, 8 Gimbal mechanism, 9 Rotation mechanism, 10 Data processing device, 11 Data recording device, 12 Data Analysis device.

Claims (7)

When a satellite equipped with a line sensor optical sensor calibrates the gain and offset correction coefficients in each pixel of the line sensor,
The imaging target direction corresponding to the scanning line of the line sensor and the moving direction of the artificial satellite are substantially matched,
A calibration method for an optical sensor mounted on a satellite, which performs calibration of gain and offset based on acquired data of output levels of pixels detected by a line sensor at different times. In the calibration of the gain and offset correction coefficients, each pixel of the line sensor is obtained from the level value in units of pixels when the same place is imaged over time.
2. The satellite-borne optical sensor calibration method according to claim 1, wherein a calibration coefficient of each pixel of the line sensor is calculated. In the calibration of the correction coefficient for the gain and offset, from the average level value of any plurality of pixels when the plurality of pixels of the line sensor image the same area over time,
2. The satellite-borne optical sensor calibration method according to claim 1, wherein a calibration coefficient of each pixel of the line sensor is calculated. In order to make the pixel alignment direction of the footprint substantially coincide with the direction of movement of the artificial satellite, the attitude of the artificial satellite is rotated by approximately 90 degrees,
The method for calibrating a satellite-mounted optical sensor according to any one of claims 1 to 3, wherein the flight is performed. In order to make the pixel arrangement direction of the footprint approximately coincide with the satellite movement direction, the optical sensor is rotated approximately 90 degrees,
The method for calibrating a satellite-mounted optical sensor according to any one of claims 1 to 3, wherein the flight is performed. 6. The satellite-mounted optical sensor calibration method according to claim 1, wherein the optical sensor uses a TDI (Time Delay & Integration) type sensor. An optical sensor mounted on an artificial satellite to take images of the ground surface;
A gimbal mechanism having a drive unit for controlling the posture so that the optical axis of the optical sensor always faces the earth,
A rotation mechanism that rotates the attitude of the optical sensor around an axis in the direction of the earth during calibration;
When a satellite equipped with the optical sensor performs calibration of gain and offset correction coefficients in each pixel of the line sensor,
The imaging target direction corresponding to the scanning line of the line sensor and the moving direction of the artificial satellite are substantially matched,
A data analysis device that calibrates the gain and offset based on the acquired data of the output level of each pixel detected by the line sensor at different times; and
An imaging apparatus comprising:

Images