EP4185532A1 - Method and system for detecting one or more targets in space based on images taken from a spacecraft such as a visual inspection satellite, and satellite equipped with such a system - Google Patents

Abstract

Disclosed is a method and system for detecting a target in space, based on images taken from a spacecraft such as a visual inspection satellite. The invention relates to a method for remotely detecting one or more targets (1) in space in low orbit around the Earth based on an imager installed on-board a spacecraft (2) that is in a sun-synchronous orbit and having to approach one or more targets located in a near orbital plane, comprising: acquiring sequentially taken images of a region of space potentially containing the one or more targets; realigning the images thus acquired by means of reference points represented by stars; detecting, within the sequentially taken and realigned images, groups of virtually stationary pixels each representing a target; measuring a time-lapse necessary for these groups of pixels thus detected to be shifted by a predetermined number of pixels; and processing this time measurement in order to deliver information on the position of the one or more targets (1) relative to the imager.

Description

DESCRIPTION

TITLE: Method and system for detecting one or more targets in space from images taken from a spacecraft such as a visual inspection satellite, and satellite equipped with such a system

FIELD OF INVENTION

The present invention relates to a method for detecting one or more targets in space from images taken from a spacecraft such as a visual inspection satellite. It also relates to a system for implementing this process, as well as a satellite equipped with such a system. The field of the invention is more particularly that of the visual inspection of space objects.

STATE OF THE ART

Satellite operators are faced with maintenance issues for their equipment, which involves being able to carry out a visual inspection in space by taking images of the target using an imager from a satellite dedicated to these operations. . The onboard imager can also be used to detect the position of the target so that the satellite approaches at the right distance.

The imagers on board the satellites essentially serve three functions: to take images of the Earth. The viewing angle of the imager is then low to have good resolution; take pictures of their solar panels to ensure they are deployed. The angle of view is then very wide. allow a "hard docking" type docking operation, when an aircraft clings to another [2] There, too, the viewing angle is wide.

Usually, the approach of a target is carried out using a radar which allows significant detection distances.

There are already methods for detecting objects in space, such as a method for detecting space objects in video satellite images using motion information [3] This detection method uses video images, c that is to say with a reduced time between images. The stars are almost in place from one image to another. A non-stellar object is then easily detected because it is not fixed. This process only works if the target has an angular speed that is really different from that of the satellite, which is not our case because the satellite and the target are in very close orbits and therefore with very close angular speeds.

There are detection methods by comparison with a catalog of stars [4], implementing the shooting of a single image and a search for the stars visible on the image in the on-board catalog. We can then distinguish a non-stellar object on the image. This method requires onboarding a catalog of stars specific to the viewing angle of the onboard imager. Star Trackers (or star chasers), with this kind of catalog, use optics with a wider angle of view.

US 5512743 discloses a satellite system for the detection of near-Earth objects. This system periodically captures two images of the celestial sphere around the earth using a pair of linear arrays of optical detectors oriented so that the image lines of the detectors on the celestial sphere are separated by the same angle. for each pixel in the azimuth. The position of the linear detectors is adjusted for each orbit of the earth by the satellite so that the entire celestial sphere can be imaged.

The document EPI 168003 Al discloses a device for measuring space pollution, intended to be embarked on a satellite, comprising: at least one passive imager capable of providing images of space; means for detecting the space debris present on the images of space, determining the trace of said debris; means for locating the distance of the detected debris, determining the distances of the detected debris with respect to the satellite; means for classifying the localized debris, determining for each localized debris its class which is the product of its average albedo by its apparent surface (surf).

The object of the present invention is therefore to propose a method and system for detecting one or more targets in space from images taken from a visual inspection satellite, which can provide detections of targets at several kilometers away from the satellite and to determine their trajectory. DISCLOSURE OF THE INVENTION

This objective is achieved with a method for remotely detecting one or more targets in space in low orbit around the Earth from an imager on board a spacecraft in a heliosynchronous orbit and having to approach one or more targets located on a close orbital plane, comprising the following steps: an acquisition of images taken sequentially of an area of space potentially containing said one or more targets, said images being spaced apart by a predetermined period of time, timestamped and associated with the corresponding inertial position of the spacecraft, a realignment of said images thus acquired by means of reference points represented by stars,

(i) detection, within the images taken sequentially and realigned, of groups of almost fixed pixels each representing a target,

(ii) a measurement of a lapse of time necessary for these groups of pixels thus detected to be shifted by a predetermined number of pixels, and

(iii) a processing of this time measurement to deliver position information of said one or more targets (1) with respect to said imager.

In an advantageous implementation of the detection method according to the invention corresponding to the detection of a single target, the latter comprises the following steps: acquiring two images, spaced apart by a predetermined period of time, to generate a first and a second image of an area of space potentially containing the target, said first and second images being time-stamped and associated with the corresponding inertial position of the spacecraft, realigning the two images by means of reference points represented by stars , performing a difference of the two realigned images, then processing this difference to produce a differential image, processing the differential image to detect therein a target in the form of a group of nearly fixed pixels, following the target thus detected, by acquisition of images of the target thus detected and positioned, until the image thus acquired is shifted by a predetermined number of pixels with respect to a reference image previously chosen from among the images already acquired, said images thus acquired being timestamped and associated with the corresponding inertial position of the spacecraft, measuring a lapse of time between the shooting of the reference image and the shooting of the last image acquired, and processing of this lapse of time thus measured to deliver position information of the target.

At least one of the acquisition, realignment, or difference phases can be carried out several times, to confirm the location of the target and overcome any erroneous detections due to stars or errors of acquisition.

The realignment step can also comprise an intermediate calculation to correct a deformation of the images according to the focal length of the lens of the imager.

According to another aspect of the invention, there is proposed a system for remotely detecting one or more targets in space in low orbit around the Earth from an imager on board a spacecraft in a heliosynchronous orbit and in front of approaching one or more targets situated on a close orbital plane, implementing the detection method according to the invention, this system comprising: means for acquiring images taken sequentially of a zone of space potentially containing said one or more targets, said images being spaced apart by a predetermined period of time, time-stamped and associated with the corresponding inertial position of the spacecraft, means for realigning said images thus acquired by means of reference points represented by stars,

(i) means for detecting, within the images taken sequentially and realigned, groups of almost fixed pixels each representing a target,

(ii) means for measuring a period of time necessary for these groups of pixels thus detected to be shifted by a predetermined number of pixels, and

(iii) means for processing this time measurement to provide position information of said one or more targets relative to said imager.

The detection system according to the invention can advantageously comprise; means for acquiring two images, separated by a predetermined period of time, to generate a first and a second image of an area of space potentially containing one or more targets, said first and second images being timestamped and associated with the corresponding inertial position of the spacecraft, means for realigning the two images by means of reference points represented by stars, means producing a difference between the two realigned images, then processing this difference to produce a differential image, means to process the differential image to detect a target therein in the form of a group of quasi-fixed pixels, means for following the target thus detected, by acquiring images of the target thus detected and positioned, until the image thus acquired is shifted by a predetermined number of pixels with respect to a reference image previously chosen from among the images already acquired, said images are thus acquired being timestamped and associated with the corresponding inertial position of the spacecraft, and means for measuring a lapse of time between the shooting of the reference image and the shooting of the last image acquired, and processing of this lapse of time thus measured to deliver information on the position of the target.

The image acquisition means can advantageously comprise an imager on board a visual inspection satellite.

The acquisition, realignment or image difference means can be implemented several times, to confirm the location of the target and overcome any erroneous detections due to stars or acquisition errors .

According to yet another aspect of the invention, a visual inspection satellite is proposed, comprising a system for the remote detection of a target in space in low orbit according to the invention.

This visual inspection satellite may further comprise means for controlling the orientation of an imager integrated in the detection system according to the invention towards the target. We use the characteristic of parallax, namely the displacement of the apparent position due to the change of position of the observer. By taking two images shifted in time and space and then analyzing them, we can detect fixed or mobile objects and calculate their characteristics such as their position, their speed because the fixed objects will not be displayed in the same place on the two images according to their distance.

The principle of the method according to the invention is thus to detect, in a first step, almost fixed pixels and then, in a second step, to measure the duration necessary for a shift of these pixels. This method is effective even with a target represented by a few pixels, that is to say when the target is several kilometers away.

DEFINITIONS

Altitude: height of a satellite above the Earth's surface Orbital speed: speed at which a satellite orbits the Earth Orbital period: time taken by a satellite to complete one complete revolution around the Earth

Pixel: base unit used to measure the definition of a digital raster image

DESCRIPTION OF FIGURES

Other advantages and particularities of the invention will appear on reading the detailed description of implementations and non-limiting embodiments, and of the following appended drawings:

• Figure 1 illustrates sun-synchronous orbits, with the sun in the observer's position (drawing not to scale);

• Figure 2 illustrates offset shots taken in the detection method according to the invention;

• Figure 3 illustrates, in inverted color mode, two images representing the same scene but with a lateral offset due to the movement of the satellite in its orbit; • Figure 4 shows, in inverted color mode, the two images of Figure 3, realigned and differentiated; and

• Figure 5 is a block diagram of an embodiment of a detection system/method according to the invention.

DETAILED DESCRIPTION

If we consider, with reference to FIGS. 1 and 2, the example of a satellite 2 in a heliosynchronous orbit (which makes it possible to have a substantially constant lateral illumination of the target) which must approach a target 1 situated on a plane close orbital, with reference to FIG. 1, this satellite carrying the imager is located between the Earth and the target, thus neither the Earth nor the Sun are in the field of vision.

The detection method according to the invention comprises, with reference to FIG. 5, five phases:

- shooting (PI)

- realignment (P2)

- difference (P3)

- positioning (P4)

- tracking (P5)

The detection, positioning and target-imager relative distance data are then transmitted from the satellite 2 to a terrestrial reception station 3 connected to a server 4 for processing and supplying spatial data.

It is also possible to provide that these detection, positioning and target-imager relative distance data are not transmitted to the ground but are used locally in a feedback loop during an approach phase of the satellite 2.

Description of the phases

In the first phase called PI shooting, a first image is taken (Image 1), then, after a predetermined period of time, a second image is taken (Image 2). Figure 2 shows the angles of view.

Two images are obtained (figure 3) representing the same scene but with a lateral offset due to the movement of the satellite in its orbit. Note: the colors have been reversed in order to facilitate the interpretation of this figure and allow its reproduction. These images have special characteristics: the target appears in the same place on both images, while the stars are shifted.

It should be noted that in this figure, the target is indicated just for information because in this shooting phase, we do not yet know its position.

In the second phase called realignment P2, the two images must be realigned with each other thanks to the stars located on the two images. The gap between the two images is theoretically known since the time lapse between the shooting of the 2 images is fixed. It is enough to adjust within a few pixels in the four directions to have the best possible precision. An intermediate calculation may be necessary to correct the deformation of the images according to the focal length of the lens. The advantage of this realignment is that it also makes it possible to correct changes in the attitude of the satellite.

In the third so-called P3 difference phase, the difference between the two realigned images is calculated. As the luminosity of the stars is substantially the same on the two images, the groups of pixels representing the stars are removed, and it is easy to detect the pixels which have not changed place between Image 1 and Image 2 because they are shifted further to realignment, as shown in Figure 4, the colors of which have been reversed to allow reproduction.

It should be noted that the shooting, realignment and difference phases can be carried out several times to confirm the location of the target and overcome bad detections due to stars and acquisition errors.

In the fourth phase called positioning P4, we calculate the spatial position of the target in 2D when we know the position of the target on the images, the field covered by the imager, the definition of the sensor, and the position of the satellite and its attitude (orientation) when taking each image.

In the fifth phase called P5 tracking, implemented once the target is located on an image, we will focus on this area of the image, which makes it possible to reduce the analysis time.

We keep Image 1 as a reference and we will take images continuously until the target is shifted by 1 pixel on the new image (Image 2) compared to Image 1.

Indeed, the difference in angular velocity between the satellite and the target causes a slow lag. The time measurement necessary for this offset makes it possible to calculate the relative angular velocity of the target, therefore its altitude and therefore consequently the distance between the satellite and the target.

Calculation formula examples and numerical applications

Imager Data

The imager has an angle of view (Field of View) side-to-side equal to FoV= 7.1° The definition of the sensor (Height and Width) is:

ResH = 2048 pixels and ResL = 2048 pixels The angular resolution in width is:

FoV

ResAL = 0.0034667969°/pixel

ResL

The angular resolution in Height is:

FnV

ResAH = — — = 0.0034667969°/pixel R6SH

Satellite Data

When the satellite is at an altitude AltSat = 550 km, its orbital period is, according to Kepler's Laws [1]:

5738.8226 seconds with

RTearth = 6378 km (average radius of the Earth) m = 398600.4418 km ³ . s ² (Standard gravitational parameter for Earth)

Its angular velocity is

360

VASat=-=0.062730637°/second

PSat Target Data

When the target is at an altitude AltCib = 555 km, its orbital period is according to Kepler's laws: 5745.0363 seconds (1) with

RTearth = 6378 km (mean radius of the Earth) m = 398600.4418 km ³ .s ² (standard gravitational parameter for the Earth)

Its angular velocity is:

360

VACib = - - = 0.062662789 °/ second

PCib'

Calculations to find the optimal time between two images:

The relative angular velocity of the satellite relative to the target is:

VARel = VASat - VACib = 6.7848 x 10 ⁵ °/second (2)

The maximum time between Image 1 and Image 2 must be chosen to have a target offset of less than one pixel between the two images, i.e. it is linked to the relative angular velocity and the angular resolution of the sensor :

ResAL

TimeMaxl = NPixel x - - = 51 seconds VARel with NPixel = 1

But for the realignment on the stars to be done correctly, the offset of the stars must be less than or equal to 10% of the width of the sensor:

ResAL

MaxTime2 = 0.1 x ResL x = 11 seconds

VASat

It is then necessary to take the minimum value between TimeMaxl and TimeMax2.

The minimum time between the shooting of Image 1 and that of Image 2 must be chosen to have a star shift greater than or equal to a few pixels three pixels between the two images, i.e. it is linked to the Satellite angular velocity and sensor angular resolution: He

ResAL

MinTime = N Pixel x = 166 ms

VASat with NPixel = 3

The value of the star offset can be expressed as a function of the time between the shooting of image 1 and that of image 2:

VASat

DecStars = time x

ResAL

The value of the target shift is, for her:

VAREL

DecTarget = time x

ResAL

Conclusion: for the shooting phase, the time lapse between Image 1 and Image 2 must be between 166 ms and 11 seconds. For example, with a time lapse of 1 second between Image 1 and Image 2, we have a star offset of 18 pixels, and a target offset of 0.02 pixels.

Calculations to estimate the distance between the satellite and the target:

During the tracking phase, the relative angular velocity of the target can be calculated in relation to the time required (TimeDec) to have a one-pixel shift between Image 1 and Image 2.

It should be noted that the longer the time required, the more this implies that the distance between the target and the satellite is low because the relative angular speed is low.

ResAL

VARel = NPix x -

TimeDec with

VARel = value in °/s

NPix = target offset value in pixels

TempsDec = time required to have the shift of NPix

The aforementioned equation (2) makes it possible to find the angular velocity of the target since we know the angular velocity of the satellite:

VACib = VASat - VARel or

ResAL

VACib = VASat - NPix x -

TimeDec

The estimated period of a target orbit is the time to travel 360° at this speed angular:

360 Where

360

PCB =

ResAL

VASat — NPix x DecTime

The altitude is deduced according to (1):

RTearth = 6378 km (average radius of the Earth) m = 398600.4418 km ³ . s ² (Standard gravitational parameter for Earth)

The distance between the satellite and the target:

With for example a measured time of 60 seconds to have a shift of 1 pixel of the target between Image 1 and Image 2, this gives a satellite-target distance of 4.257 km.

Precision on distance calculation

If the time required is no longer 60 seconds but 59 seconds, the Satellite-Target distance is 4.329 km.

So we have a resolution of 72 meters when we shoot and analyze every second. But the accuracy of the measurement can be increased by reducing the time lapse between two images: since in the tracking phase, we focus on a very small area of the image, the analysis time is greatly reduced .

With a capture and its analysis every 100 ms, if the time required is 59.9 seconds, the Satellite-Target distance is 4.265 km, which gives a resolution of 7.2 meters. Calculations over a shorter distance

When the target is represented on an image by a group with a larger number of pixels, the principle remains the same, just take the center of this group or its edge and detect the shift from one image to another .

If the time required to detect a 1-pixel offset from the target is 500 seconds, we obtain a distance of 0.510 km

At this distance, the distance resolution is 1 meter with one shot and one scan every second. Distance accuracy increases as distance decreases.

Image resolution

At a given distance, the image taken represents an area of: which makes it possible to give the resolution of the image according to the distance:

At 5 km distance, 1 pixel represents 30.29 cm At 500 m distance, 1 pixel represents 3.29 cm

If we consider that the target must represent at least a square of 2x2 pixels on each image, we can detect: a target of 60 cm side at a distance of 5 km. a target of 100 cm side at a distance of 8 km.

It should be noted that these pixels represent areas illuminated by the sun, the moon, or the

Earth. Unlit areas of the target will not be detected.

Of course, the invention is not limited to the examples which have just been described and many other embodiments can be envisaged without departing from the scope of the present invention. In particular, it can be envisaged that the differential detection system according to the invention can be embarked in all kinds of spacecraft operating in low orbit. REFERENCES

[1] "Kepler's 3 laws" https://fr.wikipedia.0rg/wiki/Satellite_artificiel#Les_trois_lois_de_Kepler

[2] "Docking with precision" (NASA, 22 Nov 2007) https://www.nasa.gov/missions/shuttle/f_docking.html

[3] Xueyang Zhang, Junhua Xiang and Yulin Zhang "Space Object Detection in Video Satellite Images Using Motion Information" (17 Oct 2017)

[4] J. Cheng, W. Zhang, M. Y. Cong, and H. B. Pan, “Research of detecting algorithm for space object based on star map recognition” (May 2010)

Claims

1. Method for remotely detecting one or more targets (1) in space in low orbit around the Earth from an imager on board a spacecraft (2) in a heliosynchronous orbit and having to approach a or several targets located on a close orbital plane, comprising the following steps: an acquisition of images taken sequentially of an area of space potentially containing said one or several targets, said images being spaced apart by a predetermined period of time, timestamped and associated with the corresponding inertial position of the spacecraft (2), a realignment of said images thus acquired by means of reference points represented by stars, (i) a detection, within the images taken sequentially and realigned, of groups of almost fixed pixels each representing a target, - (ii) a measurement of a lapse of time necessary for these groups of pixels thus detected to be shifted of a predetermined number of pixels, and (iii) a processing of this time measurement to deliver position information of said one or more targets (1) with respect to said imager. 2. Method according to claim 1, implemented for the detection of a single target, characterized in that it comprises the following steps: acquiring two images, spaced apart by a predetermined period of time, to generate a first and a second image of an area of space potentially containing said target, said first and second images being timestamped and associated with the corresponding inertial position of the spacecraft (2), realigning the two images by means of reference points represented by stars, create a difference between the two realigned images, then process this difference to produce a differential image, processing the differential image to detect therein a target (1) in the form of a group of almost fixed pixels, following the target (1) thus detected, by acquiring images of the target (1) thus detected and positioned, until the image thus acquired is shifted by a predetermined number of pixels with respect to a reference image previously chosen from among the images already acquired, said images thus acquired being timestamped and associated with the corresponding inertial position of the spacecraft (2), measuring a lapse of time between the shooting of the reference image and the shooting of the last acquired image, and processing of this lapse of time thus measured to deliver information on the position of the target (1). 3. Method according to claim 2, characterized in that at least one of the phases of acquisition, realignment, or difference are carried out several times, to confirm the location of the target (1) and to overcome possible erroneous detections due to stars or acquisition errors. 4. Method according to one of Claims 2 or 3, characterized in that the realignment step further comprises an intermediate calculation to correct a distortion of the images as a function of the focal length of the lens of the imager. 5. System for remotely detecting one or more targets (1) in space in low orbit around the Earth from an imager on board a spacecraft (2) in a heliosynchronous orbit and having to approach a or several targets situated on a close orbital plane, implementing the detection method according to any one of the preceding claims, this system comprising: means for acquiring images taken sequentially of a zone of space potentially containing the said one or several targets, said images being spaced apart by a predetermined period of time, timestamped and associated with the corresponding inertial position of the spacecraft (2), means for realigning said images thus acquired by means of reference points represented by stars, (i) means for detecting, within the images taken sequentially and realigned, groups of almost fixed pixels each representing a target, (ii) means for measuring a period of time necessary for these groups of pixels thus detected to be shifted by a predetermined number of pixels, and (iii) means for processing this time measurement to deliver position information of said one or more targets (1) relative to said imager. 6. System according to claim 5, implemented for the detection of a single target, characterized in that it comprises: means for acquiring two images, spaced apart by a predetermined period of time, to generate a first and a second image of an area of space potentially containing the target (1), said first and second images being timestamped and associated with the corresponding inertial position of the spacecraft (2), means for realigning the two images by means reference points represented by stars, means for making a difference between the two realigned images, then processing this difference to produce a differential image, means for processing the differential image in order to detect a target (1) therein in the form of 'a group of almost fixed pixels, means for following the target (1) thus detected, by acquiring images of the target (1) thus detected and positioned, until the last image thus acquired is shifted by a predetermined number of pixels with respect to a reference image chosen beforehand from among the images already acquired, the said images thus acquired being timestamped and associated with the corresponding inertial position of the spacecraft (2), and means for measuring a lapse of time between the shooting of the reference image and the shooting of the last image acquired, and processing of this lapse of time thus measured to deliver information on the position of the target (1). 7. Detection system according to claim 6, characterized in that the means image acquisition systems include an on-board imager in a visual inspection satellite. 8. Detection system according to one of claims 6 or 7, characterized in that the means of acquisition, realignment or image difference are implemented several times, to confirm the location of the target (1 ) and overcome any erroneous detections due to stars or acquisition errors. 9. Visual inspection satellite, comprising a system for the remote detection of a target (1) in space in low orbit according to any one of claims 5 to 8. 10. Visual inspection satellite according to claim 9, characterized in that it further comprises means for controlling the orientation of an imager integrated in the detection system according to any one of claims 5 to 8, towards target (1) ·