• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The invention relates to a method for acquiring images of a scene, from a carrier (2) in displacement and equipped with a sensor, which comprises a step of acquiring images (1) of the scene during the movement of the carrier, and a step of servocontrolling the angular direction of line of sight of the sensor. The acquisition is performed: ○ for a first position of the carrier, with a servo enabling a forward scan of a field band (4a) of the scene, combined with a step-and-stare scan with a micro-motion "stare" bi-axis, a first banner of images thus acquired, and in that at least one other banner of images is acquired on the same area of land by repeating these scanning steps for at least another position (P2) of the wearer, each image of another strip being acquired with a predetermined recovery rate with the images of the first banner.
    公开号:FR3044101A1
    申请号:FR1502431
    申请日:2015-11-20
    公开日:2017-05-26
    发明作者:Raphael Horak
    申请人:Thales SA；
    IPC主号:
    专利说明:

    METHOD FOR ACQUIRING IMAGES OF A SCENE FROM A SENSOR ABOARD A MOVING CARRIER WITH THE SERVING OF THE VISE LINE
    The field of the invention is that of the acquisition of images by servocontrolling the line of sight of a sensor coupled to image processing, for example for surveillance or recognition applications.
    Line-of-sight (LdV) scanning patterns of optical sensors or enslaved radars implemented in PODs, aircraft or land vehicles are conventional figures of the type: - constant angle lateral scanning, also called strip-map mode in the radar domain, - spotlight scanning on fixed point (= tracking scan on fixed point), - tracking tracking on moving target, - circular scanning for terrestrial applications.
    The first scan sweeps a strip of land but with a constant viewpoint between the aircraft and the ground point being scanned.
    The second allows to see an object on the ground according to different points of view but covers a very limited area on the ground.
    The third is a variant of the second and does not guarantee the multiplicity of points of view.
    The fourth is similar to the first and has a constant azimuth angle.
    These four scanning classes do not make it possible to visualize with a good angular resolution a large area of ground according to different points of view.
    The object of the invention is to overcome these disadvantages.
    The solution provided consists in applying a specific servo-control mode to the line of sight of the sensor by several series of ground-to-ground sweeps, each forward sweep being generally followed by a rapid-return ground sweep, in which micromovements of step-and-stare "(translation and pointing according to a summary translation) which make it possible to obtain sufficiently stable and enlightened images in spite of the fast movements of the sweeps. An image processing makes it possible to control this servocontrol to gain precision.
    More specifically, the subject of the invention is a method for acquiring images of a predetermined ground scene, from a carrier traveling along a trajectory, and equipped with a sensor having a line of sight, which comprises a step of acquisition by the sensor of successive images of the scene during the movement of the carrier, and a step of servocontrolling the angular direction of the line of sight by a processing unit connected to the sensor. It is mainly characterized in that the acquisition is carried out: for a first position of the wearer on his trajectory, with a servo-control of the line of sight making it possible to scan a predetermined strip of terrain of the scene, said forward scan, combined with a step-and-stare scan with so-called "step" micro-motions which have an amplitude controlled by the processing unit, which comprises: o at least one "step" to main component perpendicular to the forward and side scan, combined with at least one main component step parallel to the forward scan and said longitudinal step, a two-axis stare micro-motion to compensate for a translational motion of the line of sight during the acquisition of each image, a first strip of images thus acquired, and in that at least one other banner of images is acquired on the same area of ter than the first strip by repeating these scanning steps for at least one other position of the wearer on his trajectory, each image of another strip being acquired for a portion of the strip of land with a recovery rate with the (or the) image (s) of the first strip for the same portion of the strip of land, greater than a predetermined rate of recovery, the overlapping images from one banner to another from the same portion of land being thus respectively acquired during the these iterations in different directions of the line of sight.
    This solution with scanning movements adapted to the line of sight, allows both to cover a wide area possibly in the form of different contiguous or non-contiguous land strips and to visualize all of these points scanned at different angles. In particular, it makes it possible to provide strips of images covering a large area of the ground at different times and points of view.
    It respects, thanks to a specific bi-directional "step-and-stare" movement, synchronous with the main servocontrol of the line of sight, a sufficiently long illumination time for each image and thus makes it possible to guarantee the images a sufficient quality.
    The solution also makes it possible, after processing the images of the strips, these containing the ground points which are seen according to different viewing angles, to build geometrically compliant image bands, that is to say superimposable on an ortho -photography or a geographical map, which does not allow the other methods when the movement of the carrier is important compared to the relief of the zone.
    The solution proposed uses the following new concepts: Sweeping and counter-scanning repeated frontal of the line of sight, followed by acquisitions, on one or more strips on the ground, the wearer equipped with the sensor being in motion, so as to return several times on the same point of the ground with different points of view and this for all the points of a strip of land of large longitudinal extent (or if necessary for all the points of different strips of ground traversed in parallel),
    A two-axis step-and-stare that combines, in addition to conventional step-and-stare continuous micromovements, lateral steps and stare that can be large in amplitude, all combined with image processing to correctly assemble the images of each banner. While in the state of the art the micromotion is mono-axis, another peculiarity of this step-and-stare is that it introduces a micro-motion stare bi-axis to eliminate the blur and improve the quality of the image during its integration time. It also leaves the possibility of adding a rotational component of the image during the stare, thus ensuring the most perfect possible assembly of images in the image banner generated.
    Simultaneous production of various bands of images, each strip of which corresponds to a strip of land having a form controllable in shape both in width and length.
    Production for the same band of land chosen in the scene of different bands superimposed at different regular moments and taken according to different points of view.
    Ability to generalize the bands to scenes not limited in the direction of the length (bands taking place according to the advance of the carrier).
    The method is realizable in real time as the carrier progresses without constraint imposed on the trajectory of the latter: the servo compensates for the movements of the wearer during the scanning of the bands on the ground.
    The first position and the other successive positions of the carrier serving as basis for the scan may be either unique during the scan, or multiple and decomposed in as many positions as there are images during scanning when the wearer moves during the scan. scanning.
    Each image of another strip is advantageously acquired for a portion of the strip of land, with a precision of alignment with the image (s) of the first strip for the same portion of the strip of land, greater than a precision predetermined alignment.
    Preferably, within the same image strip, the adjacent images overlap partially at a recovery rate below a threshold, so as to obtain a continuous banner of images. The slaving is advantageously performed by image processing so that in the same strip, adjacent images are aligned with an alignment quality higher than a predefined quality.
    The bi-axis "stare" micro-movement can also be associated with a counter-rotating movement of the line of sight determined by the processing unit, to compensate for a rotational movement of the line of sight during the acquisition of 'picture.
    According to one characteristic of the invention, at least one other predetermined terrain band is associated with the scene and the step-and-stare mode further comprises at least one lateral band-change movement, to pass a band from land to another strip of land.
    This lateral band change movement can be combined with a front band change motion.
    The forward scan is usually done in the direction of the trajectory. Generally, a counter-scan reverse to the forward scan, said counter-scan or reverse scan, is performed following the forward scan and before reiteration.
    This counter-scan can be direct or itself combined with a step-and-stare scan. Indeed, images are possibly acquired during the counter-scanning which is combined with a "step-and-stare" scanning with so-called "step" micro-movements which have an amplitude controlled by the processing unit so that the successively acquired images overlap partially, which comprises: 1. at least one "step" perpendicular to the counter-scanning and said lateral, to scan the zone laterally, combined with 2. at least one "step" parallel to the counter scanning and said longitudinal step to scan the area longitudinally, 3. a bi-axis "stare" micro-movement to compensate for a translational movement of the line of sight during the acquisition of each image.
    A method is thus obtained which makes it possible: to enslave the line of sight of the sensor of a moving carrier, so that it sweeps a large stage on the ground while ensuring that each point of the ground is seen several times. times with different point-of-view angles covering the widest range of possible angles; to simultaneously reproduce one or more strips of contiguous (or adjacent) and overlapping images, corresponding to the scene freely chosen by the operator; that any point contained in these bands is reconstituted temporally according to different points of view, covering a wide angular range, making possible, for example, 3D rendering processes of the objects contained in these bands; - that the images used for the reconstruction of the banners are corrected in terms of conformity by exploiting the 3D information resulting from the multiplicity of points of view over a fairly wide angular domain.
    It also respects the servo constraints related to the sensor: speed of rotation and acceleration of the line of sight. Other advantages of the method can be cited as: allowing the operator to see in real time, according to different points of view, objects of the scene designated by himself, make possible a 3D mapping on all the bands , make it possible to continuously strip rolling strip modes (not spatially limited in the direction of the movement of the wearer), make it possible to concatenate the strips in a wider strip at any moment of the trajectory (notion of programmable monitoring zone in real time), allow a trajectory of the freely chosen carrier, and movements of roll, pitch, yaw of the wearer, within the limit imposed by the system (line of sight which must continue to be able to scan the banners). Other characteristics and advantages of the invention will become apparent on reading the detailed description which follows, given by way of non-limiting example and with reference to the appended drawings, in which: FIG. 1 schematically illustrates from above an example of acquisition of images from an onboard sensor on board an aircraft and for a scene with 3 disjointed strips of land not closed on themselves (fig 1a), with 3 strips of land of which 2 closed on themselves (FIG. 1b), FIG. 1c illustrating in more detail consecutive scans conducted in parallel on two distinct land strips, FIG. 2 schematically illustrating from above an example of image acquisition in step-and-stare mode. According to the state of the art (FIG. 2a) and according to the invention (FIG. 2b), FIG. 3 schematically illustrates an example of movement of the line of sight of a sensor according to the method of FIG. invention, the elevation and azimuth being expressed in degrees. From one figure to another, the same elements are identified by the same references.
    The invention will be described by considering that the scene whose image is to be acquired is predetermined and divided into several possibly disjointed and possibly parallel terrain strips, but not necessarily. Strip means a terrain area of average longitudinal dimension larger than its mean lateral dimension. These bands correspond to the terrestrial space under the carrier, which can present infrastructures, but can be generalized to any other type of surface of the scene visible from the carrier, for example the facades of buildings seen from a vehicle earthly.
    Since the shape of the scene to be scanned may be arbitrary, the corresponding bands may also be arbitrary, as well as their number and their orientation. The shape / width of the strip traveled can take for example a shape of ellipse or S, the width of the strip can also widen or decrease on the contrary during the scan. The orientation of the strip on the floor is chosen freely.
    Examples of sweeping strip strips will be shown in connection with FIGS.
    FIG. 1a illustrates a first initial scan of three terrain strips 4a, 4b and 4c from a first position P1 of the carrier 2 (which itself moves during this first scan), the lines of sight being indicated in a solid line, and a second scan of the three bands from a second position P2 distinct from the wearer, the lines of sight being indicated in dashed lines. It is understood that several cases may appear: the case where the initial scan is strictly from the same position with a second scan (and possibly other scans) made from a (or of) single point (s) different from the first / previous (this would be the case for a moving carrier making time jumps to change position, such as a helicopter mixing stationary positions to sweep the ground and movements from one point to another to change your point of view).
    The case where the initial scanning and the following are done during a continuous movement of the carrier (the points (P1) corresponding to each image of the scan are then different successive points close to each other). The proximity of these points depends on the scan speed assumed to be large enough that a second scan (or other scans) can be done before starting another scan at a point (Pi) far enough away from the previous one, but not too much also remote so that the different sweeps (at least two) can be done before the land strips traversed by the line of sight of the sensor out of the field of view of the sensor scanning.
    The third case is the case mixing the two previous modes: this is the case of a carrier scanning the scene with its sensor throughout its movement on a portion of its trajectory, and performing scans for different spatial positions different stationary on the rest of the trajectory (if the wearer allows it).
    The scanning according to the invention covers these different cases, as long as we consider in all cases a beginning of next scan at different points (P1), (P2), (Pi). Figures 1 and 2 show only case 1 or case 2 with a very fast sweeping movement (the sighting lines start from each point (P1) almost coincidentally). Case 2 with a sweeping motion a little slower (but still quite fast so that at the end of the scan, the wearer does not have time to reach the next position at which the next scan begins) is not shown in the figures.
    In all cases, an image corresponds to a terrain zone (this is the footprint of the image projected on the terrain along the line of sight and the sensor field) and at a point in the trajectory of the carrier unique. We can also say that a point of the field swept twice or more, is from points of trajectory (Pi) inevitably different.
    The different scans are made so that most of the images 1 from the first scan of the terrain strips are covered by images from the second scan (or subsequent scans if they exist).
    FIG. 1b illustrates different directions of the line of sight during a first scan of two land strips 4a and 4b closed on themselves and in this case surrounding the carrier 2, and a third land strip 4c. . A single direction of the line of sight 3pi, ..., 3p5, is indicated for each of the 5 positions P1, ..., P5 of the wearer shown in the figure, and from the 5th direction 3P5 images are acquired which are superimposed on certain images acquired from the position P1 of the wearer. The case presented is a case study that would correspond to a very fast movement of the wearer and slower the line of sight, because we did not want to overload the figure. In reality, the 5 line of sight directions can be interpreted as an example of lines of sight corresponding to 5 full sweeps of the line of sight for 5 positions of the wearer.
    The direction of the forward scan is generally related to the direction of the trajectory of the wearer, however for a band closed on itself, (as shown in Figure 1b with the bands 4a and 4b) or for a band strongly inclined with respect to the trajectory of the carrier (see band 4c of Figure 1b), the direction of scanning "before" is determined by the processing unit so as to minimize the constraints of movement of the line of sight during the movement of the wearer and increase in fine the number of round trips on each band.
    According to the invention described with reference to FIGS. 1 and 2, the images 1 are acquired by a sensor on board a carrier 2 in motion along a trajectory 20.
    We first consider the acquisition of a first banner of images also designated banner initial. It is carried out for a first position P1 of the carrier 2 on its trajectory, with a servo-control of the line of sight 3 making it possible to perform: a scan of a first predetermined terrain strip 4 of the scene, starting from a initial position on the ground of the line of sight, said forward scan, combined with a step-and-stare scan with so-called "step" micro-movements, which comprises: at least one perpendicular step in the forward and lateral scan 5, to scan the strip of land 4 laterally (or rather laterally), combined with o at least one step 7 parallel to the forward scan and said longitudinal step to scan the strip of land longitudinally (or rather longitudinally), o a bi-axis "stare" micro-motion to compensate for a translation movement of the line of sight during the acquisition of each image 1.
    A first banner of images is thus acquired on a first strip of land. More generally, if at least one other land strip is considered, the two or more land strips are scanned simultaneously in parallel, by means of additional lateral or front "step" movements to move from one strip to another. the other during the scan, realizing a set of initial image bands covering these different strips of terrain. Preferably, for this strip of images to be continuous, the adjacent images are partially overlapping with each other. A low overlap of adjacent images is sufficient; this recovery rate is greater than a predetermined low recovery rate, typically 20% or even 10%. Most of the time, we opt for a minimal recovery rate in order to maximize the number of bands stored in the processing unit.
    Once this initial banner (possibly supplemented by other initial bands if several bands of land have been predetermined) has been acquired, other bands of images are acquired for the same strip of land (the same strips of land in the general case), by repeating these scanning steps for another position P2 of the wearer on its trajectory 20. Each image of another strip is acquired for a portion of strip of land with a high recovery rate with the (or the image (s) of the initial strip for the same portion of the strip of land. Thus, the overlapping images from one strip to another, from the same portion of land are respectively acquired during these iterations in different directions of the line of sight because according to different positions of the wearer. Generally, at least one other predetermined terrain band is associated with the scene. In Figures 1a and 1b, there can be seen three separate strips of land 4a, 4b and 4c. Then, the "step-and-stare" mode further comprises at least one lateral movement 6 of change of terrain strip, to move from one strip of land to another strip of land. The image bands corresponding to separate strips of land can be acquired in parallel. 4 bands of images are shown in FIG. 2b: 40a, 40b, 40c, 40d.
    We will detail these steps now. The line of sight 3 of the sensor is slaved so as to: 1) make it scan forward these predetermined land strips 4. This forward scan of the line of sight (whose ground track can reach and even exceed a speed of 1500m / s in the case of a sensor on board an aircraft) is also combined with o a succession of micro - Two-way step-and-stare local movements. Indeed, because of the very fast speed of the forward scan of the line of sight (and the very fast speed of its realization on the ground), it is necessary to introduce in the general movement of the scan before microbalays "step- and-stare ": micro-scan" stare "to ensure during imaging, sufficient image stability and integration time and micro-scan" step "to switch from one image to another. It is a question of performing the "step" scanning by jumps 5 or 7 very fast between each image taking (= to ensure the passage of an image to the other on the same band 4) or lateral jumps 6 larger (to move from one band to another), then perform a "stare" of applying during the image taking a microstrip compensating the movement of the line of sight to achieve images, each with a long integration time during which the shooting is stabilized on the strip of land.
    These micro-movements 5, 6, 7 of translation are not all identical from one image to the next; they are controlled by the servo unit. The latter must make checks to ensure a controlled partial overlap between successive images on the same band, but also between adjacent images of the same band separated temporally by a lateral scan return applied during the passage of the first band to the other bands field and back to the first band. These verifications consist of measuring, by inertial means, the recovery rate between adjacent images of the same band and verifying that a minimal overlap (meeting a predetermined criterion by the system) exists between them.
    In general, the micro-movements of "step" are decomposed into a component 5 or 6 perpendicular to the direction of the forward scan (5 is a "step" of small amplitude corresponding to the transition from one image to the next to the inside a band, 6 is a "step" of greater amplitude corresponding to the passage from one band to another), and a component 7 parallel to the direction of the forward scan. "Steps" whose main component is parallel to the direction of the forward scan are called "steps" front; the "steps" whose main component is perpendicular to the direction of the forward scan are called lateral "steps". Different scanning patterns of these steps can be performed. When only one band is to be scanned, it is possible, for example, to perform a lace scan, with a succession of exclusively lateral steps (with lateral principal components) to traverse the strip 40 laterally as shown in FIG. 2a, followed by a "step". »7 exclusively frontal (front main component) to achieve a forward advance of the band, then a succession of" steps "exclusively lateral in the opposite direction to previous to travel the band laterally in the other direction, etc.. We can of course provide less simple scanning patterns using the same elementary "steps" but combined differently.
    Stare micro-movements to compensate for LOD movement during image capture are also bi-directional translation motions; they are not represented in the figures. To gain precision in the assembly of the images, a counter-rotating micro-scan is advantageously applied in rotation around the line of sight between each "step" movement and at each image acquisition during the "stare" function. . In addition to the assembly precision, this allows to offer a longer integration time (or illumination in the case of a lidar). o these micro-movements of "steps" are accompanied by additional lateral movements 6 and controlled from the line of sight to move from one band to another, these movements can be short if the two bands are close, or ample if they are distant. Similarly the movements to move from one band to the other can be decomposed into lateral components and frontal components; but these additional movements 6 have a lateral component greater than their frontal component. FIG. 2b shows four strips 40 (strip 40a, strip 40b, strip 40c, strip 40d) corresponding respectively to four stripes 4a, 4b, 4c and 4d (not shown in this figure). If the scene has only one band, these movements 6 to move from one band to another do not exist.
    FIG. 1c illustrates a scanning strategy in which lateral jumps (in azimuth) are preferred so as to first traverse all the strips in parallel in a manner, then frontal jumps on the strips furthest away from the wearer before return to closer bands, to minimize yaw jumps that may take longer to achieve than jumps in azimuth. In this way the separate land strips are scanned almost in parallel. More precisely, FIG. 1c illustrates in more detail 2 consecutive scans respectively carried out from two positions P1 and P2, on two near land strips 4a and 4b; on the images referenced 1 are indicated the temporal order of acquisition of the images during the scanning of the two bands: t1, t2, ..., t6. Once for a strip of land, a first image banner (initial strip) was generated from a position (position P1 in the figure with the two initial bands (strip 40a1 for strip 4a and strip 40b1 for band 4b dotted images), one (or more) other strip (s) of images of the same band of land will be generated during the following iterations, each other strip being obtained from another position of the carrier 2 (position P2 in the figure with these two other bands (band 40a2 for the band 4a and band 40b2 for the band 4b striped images) .This allows to acquire several times images of the same point of the ground with different points of view and this for all the points of a band of land of wide width The iterations determine the number of times that each band of ground is traversed.
    For each other strip, the servo processing of the line of sight of the sweeps on the same strip of land, must also ensure that the current image (= being acquired) projected on the ground has a high recovery rate with the corresponding image of the first generated banner. In other words, each image of another strip is acquired for a portion of this strip of land with a recovery rate with the image (s) of the first strip for the same portion of the strip of land. This recovery rate is greater than a predetermined high rate of recovery, for example 80%. Thus, the overlapping images (according to this rate) from one strip to another, originating from the same portion of terrain are respectively acquired during these iterations along different directions of the line of sight (considered in the 3D coordinate system of the scene), that is to say according to different angles of view. Indeed, for a recovery rate fixed for example at 80%, at least 80% of the points of the initial strip of land are "reviewed" during the acquisition of images of each iteration; In other words, at least 80% of the points of the strip of images initially acquired are found in each band respectively acquired during the iterations. Thus, for example in FIG. 1a, points of the scene at the bottom left of the band 4a are present in two images 1: one taken when the carrier 2 is in a first position P1 on its trajectory 20, the another catch when the wearer is in another position P2. The same principle is found in FIG. 1c, except that this time the numbers indicate an example of successive positions of the footprint of the images acquired by the wearer in position 2 during the scanning of the initial 2 bands, these successive imprints. corresponding to successive instants t1, t2, ..., t6 of acquisition of these images. Note that the images acquired during the second scan may very well not match the images acquired during the first scan and overlap, while ensuring maximum recovery of all images on the initial 2 bands. A rotation of the line of sight of the sensor can however also be applied to ensure a perfect alignment of the images produced. This rotation, if it is performed, is performed by servocontrol itself which can be controlled by image processing. The processing of the acquired images makes it possible to generate strips of extended images seen at different angles of incidence and each strip of images can be regenerated temporally so that each point of the scene imaged in this strip can be seen temporally in different directions. For example, until the scanned scene can no longer be traversed by the line of sight (sensor out of range or line of sight angle exceeding a permissible threshold).
    We can notice that a classic "step-and-stare" which is illustrated figure 2a, is known and applied to quickly sweep the ground in lateral continuous (with small frontal jumps): it thus generates a band of longitudinal images extended, but the banner is generated only once by these conventional means and is necessarily seen from a single point of view (this mode of scanning is called "strip-map"). It differs from the scan according to the invention which generates several bands seen from different points of view on the same area of land, and excludes the possibility of covering several strips of land. Another known mode is that of the "spot-light" mode which consists of enslaving the line of sight on the same place on the ground. It exploits, if necessary, image processing to enslave and stabilize the line of sight on the same place on the ground (the image processing correlates the different successive images of the sensor to maintain a fixed point on the ground). The result of this mode is a succession of images whose footprint does not exceed the size of each image projected on the ground. The intersection of the projected images then defines the area of the terrain viewed from different angles by the different successive images. Here again, this mode differs a lot from the scanning according to the invention since the restored terrain zone is very localized (restricted to the intersection of the projections of images on the ground) and does not correspond to an extensive strip of land as in the application that we propose. According to the invention, one or more bands of images are generated which traverse one or more strips of land with a large longitudinal extent, which are reiterated at least twice and preferably more by traversing the strip or strips of land by successive sweeps as shown. Figure 2b, which has the consequence that each point of the scene present in successive bands is seen temporally in different directions of angles of the line of sight of the sensor.
    For each other band, the servo processing of the line of sight sweeps on the same strip of land, must also ensure that within each other strip, adjacent overlapping images are preferably with a low recovery rate predetermined minimum, so that each other banner of images is also a continuous banner, as already indicated for the initial banner.
    This treatment of the line of sight ensures that the images taken on the same portion of a strip of land at different times and lines of sight, are all overlapping with a high recovery rate (80% in our example). One can thus extract a common area of land containing all the images and which is sufficiently large and exploitable to obtain on a long stretch of ground which corresponds to the intersection of the longitudinal strips restored (closed or not on themselves, see figures 1 a or 1 b) any area on which we can view images with the different points of view of the wearer viewing this area. For a given area of land covered by the various strips returned, it is very easy, knowing the geographical coordinates of the chosen area, to extract in the different banners the corresponding images which visualize the area according to the different points of view of the wearer. swept this area. Thus a same point of ground is seen several times at different times and with different points of view. The servo processing applied to control the movements of the line of sight can be realized according to different embodiments.
    According to a first embodiment based on the use of inertial means, this treatment consists of: projecting the images on the ground using the inertial means and calculating the overlaps between two adjacent images projected on the ground when these images are part of the same band of images (the technique used calculates for example the ground projection of the current image and the adjacent image to measure the geometric transformation existing between these two projected images and measure their recovery rate as well as their effective alignment - do the same thing between the image commonly acquired on the current image banner (assuming that the current banner comes from a scan of the line of sight other than the initial scan) and that (or ) of the initial image strip resulting from the first scan on the same terrain area as that of the current image, - verify that the coverage is less than to the setpoint (= predetermined rate) targeted for the low overlap between two images of the same strip (less than 10% for example) and that it is higher than the instruction relative to the strong recovery that must exist between the images of a another banner with those of the initial banner (greater than 80% for example), - check the same thing for the desired alignment between images, - if the recovery is lower than the target, slow down the relative movement of the line of sight and accelerate it in the opposite case (or the opposite depending on whether the overlap is on the left in the direction of movement or on the right, or whether it is the overlap between images on the same area of terrain taken from different instants and for a different scan, or the overlap i of adjacent images inside the same banner and during a same scan), - same for the alignment of the images, correcting it if born necessary by acting on the rotational movement of the line of sight (when this movement is possible). )
    This type of servocontrol is sufficient when the line of sight angular position information is sufficiently precise relative to the movement of the carrier (absolute position of line of sight having a low drift with respect to the movement of the carrier) and that has a sufficiently precise knowledge of the environment to be able to calculate a projection of the ground image, from a terrain model provided by the navigation system for example. These conditions are met in many applications, for example for a terrestrial vehicle having a GPS and an IMU MEMS, equipped with a medium field camera (typically i 40 °) sweeping rapidly the nearby scene; but this is no longer the case for a carrier moving very quickly and having a very small field camera scanning a scene at a great distance with an inertial unit whose angular measurement or drift errors are much greater than the resolution of the pixel from the camera.
    In the case where the precision of information on the angular position of the line of sight and / or on the ground, obtained by the inertial means, is insufficient to allow projection of the images on the ground sufficiently precise to obtain the overlays and alignments required Preferably, a second type of servo based on image processing is employed: o Image matching by known image processing techniques (correlation of sub-image pixels, mapping of primitives ), calculating the geometric transformations existing between these images from the previous mappings, to directly measure the overlap and the alignment of these once projected on the ground. This mapping applies to adjacent images from a first scan on a strip of terrain, but also to overlapping images from successive scans on the same area of terrain. o Measurements of the correlations and geometrical transformations between projected images and those of the initial strip to accurately measure the position of the images projected in the reference of the initial bands, restitution from these measurements of the angular position of the movements of the line of sight (in relative and in absolute terms if one has a precise angular reference for the first strip), o Use of the measured angular position of the line of sight to apply the angular corrections accordingly with a possible adapted temporal filter. These angular corrections take into account the measured angular position of the line of sight and that which it should have by imposing the "weak" recovery criterion on the adjacent images and the "strong" recovery criterion that must exist between recovering images obtained during successive sweeps on the same area of ground underlying at different times.
    This results in a very fine positioning of the line of sight on the ground that could not have been achieved without the contribution of image processing despite the fact that the system is not equipped with a sufficiently precise direct measurement of angular position of the line of sight. Indeed, when the inertial means are insufficient and that the framework of the application can not be satisfied with the servoing technique presented above, the image processing makes it possible to guarantee in a precise way, that a minimum percentage of points Each of the selected terrain strips will be revisited by several successive sweeps under different angles of view and lines of sight and the successive images of each strip will be correctly joined and aligned. As a reminder: the line of sight is the direction connecting the image sensor of the wearer to a point of the terrain targeted by it and the angles of view are the angles of this line of sight in the terrestrial reference, that is to say the absolute angles of the line of sight. By definition, these angles are also the angles at which a point of the ground is seen by the sensor (and vice versa the angles according to which a point of the ground "sees" the sensor). 2) Once the terrain strip 4 has been scanned by this combined forward scan, or the terrain strips 4a, 4b, 4c have been scanned in parallel, scan the line of sight straight back as shown on FIG. 2b by the downward arrow, in dotted line, or the arrow 8 of FIG. 1a, so as to bring the line of sight back to (or almost at) the initial position chosen in the scene, but from a point of view different because of the advance of the wearer. This sweeping back or counter-sweep can be direct that is to say be performed in a single "step" (or jump): it is faster than the forward scan. This counter-scan may also not be direct. Indeed during the reverse scan, it may optionally perform "step-and-stare" operations as described above, depending on the speed and time given to it. This ground position of the LdV, reached at the end of this reverse scan (= return position) may not coincide exactly with the initial ground position of the LdV initially obtained on the first headband an offset, previously fixed by the system. and typically 20% of the size of the first image is admitted especially in the case of scrolling tape as will be seen later. This return position is determined by the image processing unit acquired until that moment. The forecast return position, as close as possible to the position of the initial line of sight on the first strip, takes into account the constraints that are prohibited for servocontrol (called dead-lock constraints) and the maximum clearance allowed by the system for the line of sight. A drift of the carrier may cause servo corrections by the system to bring the line of sight back to the forecast point. An image processing, if it is applied, also makes it possible to accurately measure the angular difference between the line of sight of the current image and that of the initial image in order to correct the enslavement accordingly.
    The foregoing description is based on a forward scan ie in the direction of the carrier's trajectory and a counter-scan in the opposite direction. More generally, a forward scan and a reverse counter-scan are considered which are not necessarily linked to the trajectory of the carrier, but whose meaning is chosen so as to minimize the deflection constraints of the line of sight and to optimize the number round trip i made by the line of sight. It is thus possible to adapt the choice of scanning direction to bands closed on themselves or strongly inclined with respect to the trajectory. One can also have as sweeping a scan in the direction opposite to the trajectory (= a backward scan), with then as counter-scanning, a forward scan. i
    The front-and-side "step-and-stare" movement on the line of sight allows the number of bands to be scanned, and determines the number of times it is traveled and the desired shape and orientation of each line. bandaged. ; According to a particular case illustrated in FIG. 1b, the stage on the ground forms one or more strips 4 which can be opened or closed, for example in the form of a circle or a rectangle. Several concentric forms can be established simultaneously around an initial form, joining the idea of distinct bands. The scanning of a closed band i can be carried out in a single forward scan repeated in a loop all along the closed bands (in the case where these surround the carrier as is the case in FIG. 1b, the scanning in loop of the line of sight is done around the wearer by successively scanning the different bands as illustrated more particularly in Figure 1c) or repeated round-trip scans between 2 points of the band, depending on the choice made on the servo . The bands can also cross each other.
    The method offers two modes of scanning round trip depending on whether or not to maximize the number of views on the ground.
    A first repeated scanning and counter-scanning mode is adapted to maximize the number of viewpoints on the ground for a given band size. FIG. 3 illustrates the movements of the line of sight in the case of site-and-field scans carried out for a ground scene of 10km long by 500m wide, from a POD on the airplane sweeping this scene. at a height of 6 km with a lateral distance of the belt from the POD of 8 km, and a maximum angular deflection of +/- 45 °. In the example presented, the different points of the ground can be seen up to 16 times at different times and different points of view.
    According to a second embodiment, the ground scene forms one or more bands that evolve during the forward movement of the wearer. These are, for example, drop-down strips that take place as the wearer advances, and which are therefore not limited in length. The bands of images that are generated are constantly updated according to different angular points of view throughout the duration of the carrier's movement without discontinuity. The length of the band therefore corresponds to almost the distance traveled by the carrier (several tens or hundreds of kilometers in the case of an aircraft).
    In the case of the second mode where we want to visualize several bands according to different points of view continuously, it is shown that it is sufficient to have a constant movement in the forward and backward scanning patterns, and that by properly synchronizing, the result is the production of strips corresponding to strips offset on the ground (according to the direction of the trajectory of the wearer) whose points are successively reviewed in time according to different angular points of view. The number of times that all the points of the band are seen depends directly on the maximum speed of scanning the beam on the ground, but will always be less than the first mode of scanning and counter-scanning described above. In our example, there will be 5 different angular views of each point of the terrain overflowed continuously.
    Prior to the implementation of this enslavement of the LdV according to the invention, the servo parameters are optimized as a function of: the average distance and the spatial extent of the scene to be scanned; bands associated with this scene, the desired average rate of recovery of the images of several bands from the same portion of the strip of land to be scanned, and the average recovery rate of two adjacent images of the same strip, the maximum speed and the maximum acceleration of the scan, - the inclination, the field of view, the resolution of the sensor and, - the integration time necessary to correctly image the scene, to determine: - the size of the floor images scanned by the sensor at a time t, - the minimum displacement (in lateral and frontal step) to be performed by the scanning to partially cover the different images between them, in order to produce a continuous banner of images. This displacement is calculated at each moment taking into account the ground disposition of 2 current adjacent images, but it can also be adjusted globally after a first complete angular sweep; the instantaneous scanning speed or the average scanning speed necessary to achieve the above condition and to respect the necessary integration time as well as the time required to perform the associated "step-and-stare" micro-scanning pattern; - the maximum angular deflection (= maximum angle traveled by the LdV) that can be achieved that ensures that the constraints of scanning speed and maximum acceleration of the sensor will not be exceeded; - the maximum number of round-trip sweeps to cover the ground scene and the maximum lateral deflection that can be achieved; the maximum average recovery rate that can actually be achieved in the recovery of the successive strips and the average minimum recovery rate between adjacent images of the same strip that must be requested.
    The forward scans are made at a speed taking into account as indicated above the various constraints of the system: trajectory of the carrier, size on the ground of the images, integration time required, maximum speed and acceleration of the servocontrollings according to different axes (for example site / deposit that do not necessarily have the same constraints).
    Backscatches (full scan in the reverse direction of the previous one) are done at full speed, regardless of integration times or system data related to imaging (unless the system allows it), and are only related at maximum speeds and accelerations> maximum servo control.
    The forward and backward sweeps must be angularly adapted to ensure a sweep or counter-sweep on the same portion of the ground. The purpose of this slaving is to allow the creation of contiguous or overlapping image strips, regenerated at regular time intervals according to different angular points of view. This allows the operator:
    To see the different elements of the scene according to different points of view. This allows in particular the unmasking of objects that could be hidden by elements of the scene, but also to see the different facets of an object of the scene to better recognize it.
    To reconstruct with appropriate algorithms and software and i known the 3D structures of the scene and to dress this reconstruction with the different images of the bands covering these structures.
    To allow other algorithms and software to render a consistent ortho-photography of different parts of the scene and following different plans. This aspect is not achieved or only in a very small extent by other scanning techniques.
    The method can be generalized to servo other types of sensors (radar, lidar, ...).
    The method is applicable to an aircraft as in the examples i presented, but is entirely applicable to any moving carrier (land vehicle, boat, short or long-range drone, ...). From the images acquired using this method, industrial applications can be envisaged:> - specific SAR or lidar applications requiring passages with different inclinations, the generation of conformed high resolution mosaic maps over a wide range, including at very great distance using low-field, high-resolution optics, generating these same cards at different times with different presentation angles, which makes it easier to detect low-contrast or partially-masked objects, the possibility of doing so. 3D reconstruction over a wide spatial extent, high resolution and long distance. )
    The present invention can be implemented from hardware and / or software elements. The servocontrol of the line of sight can in particular be implemented from a computer program product, this computer program comprising code instructions for performing the steps of the servo process. It is recorded on a computer readable medium. The support can be electronic, magnetic, optical, electromagnetic or be an infrared type of diffusion medium. Such supports are, for example, Random Memory Access Memory RAM, Read-Only Memory ROM), tapes, floppy disks or magnetic or optical disks (Compact Disk - Read Only Memory (CD-ROM), Compact Disk - Read / Write (CD-R / W) and DVD).
    权利要求:
    Claims (16)
    [1" id="c-fr-0001]
    A method of acquiring images of a predetermined ground scene from a carrier (2) traveling along a path (20) and equipped with a sensor having a line of sight (3), which comprises a step of acquisition by the sensor of successive images (1) of the scene during the movement of the carrier, and a step of servocontrolling the angular direction of line of sight by a processing unit connected to the sensor, characterized in that the acquisition is carried out: for a first position (P1) of the wearer on his trajectory, with a servo-control of the line of sight making it possible to scan a predetermined land strip (4, 4a) of the scene, said forward scan, combined with a step-and-stare scan with so-called "step" micro-motions which have an amplitude controlled by the processing unit, which comprises: o at least one "Step" with principal component the perpendicular to the forward and lateral scan (5), combined with o at least one "step" (7) with principal component parallel to the forward and said longitudinal scan, o a "stare" bi-axis micro-movement to compensate for a movement of translation of the line of sight during the acquisition of each image, a first image banner (40,40a1) thus being acquired, and in that at least one other image banner (40a2) is acquired on a same area of land as the first band by repeating these scanning steps for at least one other position (P2) of the wearer on its path, each image of another band being acquired for a portion of the land strip with a recovery rate with the image (s) of the first strip for the same portion of the strip of land, greater than a predetermined high rate of overlap, the overlapping images from one strip to the other from the same portion of land thus being respectively acquired during these iterations in different directions of the line of sight.
    [2" id="c-fr-0002]
    2. Image acquisition method according to the preceding claim, characterized in that the first position (P1) and the successive positions (P2) of the carrier serving as a basis for scanning are multiple and decomposed in as many positions as there has images during scanning when the wearer moves during scanning.
    [3" id="c-fr-0003]
    3. Image acquisition method according to one of the preceding claims, characterized in that the servocontrol is performed by image processing so that in the same band, adjacent images have a recovery rate less than a predetermined low recovery rate, so as to obtain a continuous strip.
    [4" id="c-fr-0004]
    4. Image acquisition method according to one of the preceding claims, characterized in that the servocontrol is performed by image processing so that within the same band, adjacent images are aligned with a quality alignment greater than a predefined quality.
    [5" id="c-fr-0005]
    5. Image acquisition method according to one of the preceding claims, characterized in that the bi-axis "stare" micro-movement is also associated with a counter-rotating movement of the determined line of sight (3). by the processing unit, to compensate for a rotational movement of the line of sight during the image acquisition.
    [6" id="c-fr-0006]
    6. Image acquisition method according to one of the preceding claims, characterized in that at least one other band (4b, 4c) of predetermined terrain is associated with the scene and is traversed by the scan and in that the "step-and-stare" mode further comprises at least one lateral movement (6) of changing the terrain strip, to move from one strip of land to another strip of land.
    [7" id="c-fr-0007]
    7. An image acquisition method according to the preceding claim, characterized in that the lateral movement (6) of change of the terrain band is combined with a front band change movement.
    [8" id="c-fr-0008]
    8. An image acquisition method according to one of the preceding claims, characterized in that the forward scan is performed in the direction of the trajectory (20).
    [9" id="c-fr-0009]
    9. An image acquisition method according to one of the preceding claims, characterized in that a counter-scanning reverse scan, said counter-scanning, is performed following the forward scan and before reiterations.
    [10" id="c-fr-0010]
    10. An image acquisition method according to the preceding claim, characterized in that the counter-scan is a direct counter-scan (8).
    [11" id="c-fr-0011]
    An image acquisition method according to claim 9, characterized in that images (1) are acquired during counter-scanning which is combined with step-and-stare scanning with micro-movements. so-called "steps" which have an amplitude controlled by the processing unit so that the successively acquired images (1) partially overlap, which comprises: a. at least one "step" perpendicular to the counter-scanning and said lateral, to scan the zone laterally, combined with b. at least one "step" parallel to the counter-scan and said longitudinal step to scan the area longitudinally, c. a bi-axis "stare" micro-movement to compensate for a translation movement of the line of sight during the acquisition of each image.
    [12" id="c-fr-0012]
    12. An image acquisition method according to one of the preceding claims, characterized in that at least one strip of land is held with the advance of the carrier.
    [13" id="c-fr-0013]
    13. An image acquisition method according to one of the preceding claims, characterized in that at least one strip of land (4a, 4b) is closed on itself.
    [14" id="c-fr-0014]
    14. An image acquisition method according to one of the preceding claims, characterized in that at least two strips of land (4) traversed by the line of sight intersect.
    [15" id="c-fr-0015]
    15. An image acquisition method according to one of the preceding claims, characterized in that the carrier (2) is an aircraft.
    [16" id="c-fr-0016]
    16. A computer program product, said computer program comprising code instructions for performing the steps of the image acquisition method according to any one of claims 1 to 15, when said program is executed on a computer.
    类似技术:
    公开号 | 公开日 | 专利标题
    EP3377854B1|2019-12-25|Method for acquiring images of a scene, from a sensor on board a moving carrier, with servocontrol of its line of sight
    EP3144881B1|2018-08-15|Method creating a 3-d panoramic mosaic of a scene
    EP2715662B1|2019-07-17|Method for localisation of a camera and 3d reconstruction in a partially known environment
    EP2687817B1|2019-03-20|Method of correlating images with terrain elevation maps for navigation
    US20140340427A1|2014-11-20|Method, device, and system for computing a spherical projection image based on two-dimensional images
    EP2984625B1|2017-11-08|Method for estimating the angular deviation of a mobile element relative to a reference direction
    EP2333481B1|2014-10-15|Optoelectronic system and method for creating three-dimensional identification images
    FR2853121A1|2004-10-01|DEVICE FOR MONITORING THE SURROUNDINGS OF A VEHICLE
    Mandlburger et al.2017|IMPROVED TOPOGRAPHIC MODELS VIA CONCURRENT AIRBORNE LIDAR AND DENSE IMAGE MATCHING.
    EP2239597A1|2010-10-13|Airborne radar for multimode ground surveillance
    EP2469299B1|2015-01-21|Method for enhancing images acquired by a radar with synthetic aperture
    Zienkiewicz et al.2015|Extrinsics autocalibration for dense planar visual odometry
    FR2952743A3|2011-05-20|METHOD OF ESTIMATING THE MOTION OF A SCREEN OBSERVING INSTRUMENT OVERLOOKING A CELESTED BODY
    WO2015185749A1|2015-12-10|Device for detecting an obstacle by means of intersecting planes and detection method using such a device
    EP1645888B1|2008-12-31|Device for assisting the steering of an all terrain vehicle
    EP2513592B1|2015-02-25|Method for designating a target for a weapon having terminal guidance via imaging
    Jutzi et al.2013|Improved UAV-borne 3D mapping by fusing optical and laserscanner data
    Toth et al.2008|Recovery of sensor platform trajectory from LiDAR data using reference surfaces
    EP0785414B1|2002-10-23|Method and device for determining a shift between the actual position and the planned position of an aircraft
    FR3113330A1|2022-02-11|Method for aligning at least two images formed from three-dimensional points
    FR3098929A1|2021-01-22|Method for determining extraneous calibration parameters of a measuring system
    Deschaud et al.2014|Colorisation et texturation temps r'eel d'environnements urbains par systeme mobile avec scanner laser et cam'era fish-eye
    Li et al.2009|3D Terrain Reconstruction Base on ALV Sensor
    同族专利:
    公开号 | 公开日
    CN108351208B|2021-05-07|
    EP3377854B1|2019-12-25|
    IL259208A|2021-07-29|
    US20180332232A1|2018-11-15|
    EP3377854A1|2018-09-26|
    US10616492B2|2020-04-07|
    CN108351208A|2018-07-31|
    CA3005390A1|2017-05-26|
    WO2017085259A1|2017-05-26|
    KR20180084804A|2018-07-25|
    RU2725776C2|2020-07-06|
    ES2774080T3|2020-07-16|
    FR3044101B1|2020-04-24|
    RU2018121631A|2019-12-25|
    RU2018121631A3|2020-01-21|
    IL259208D0|2018-07-31|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US20050177307A1|2002-05-30|2005-08-11|Rafael-Armament Development Authority Ltd|Airborne reconnaissance system|
    FR2884312A1|2005-04-08|2006-10-13|Thales Sa|SYSTEM FOR DESIGNATION AND / OR TARGET ILLUMINATION AND AIR RECONNAISSANCE|
    FR2944110A1|2009-04-07|2010-10-08|Thales Sa|MULTI-MODE FLOATING SURVEILLANCE AIRPORT RADAR|
    US20120249739A1|2009-05-21|2012-10-04|Elta Systems Ltd.|Method and system for stereoscopic scanning|
    WO2014122625A1|2013-02-08|2014-08-14|Thales Alenia Space Italia S.P.A. Con Unico Socio|Multiple-swath stripmap sar imaging|
    US20140266869A1|2013-03-15|2014-09-18|Mitsubishi Electric Research Laboratories, Inc.|Method and System for Random Steerable Sar Using Compressive Sensing|
    US5481479A|1992-12-10|1996-01-02|Loral Fairchild Corp.|Nonlinear scanning to optimize sector scan electro-optic reconnaissance system performance|
    US5557397A|1994-09-21|1996-09-17|Airborne Remote Mapping, Inc.|Aircraft-based topographical data collection and processing system|
    FR2752619B1|1996-08-23|1998-11-13|Thomson Csf|AIR-TO-GROUND RECOGNITION METHOD AND DEVICE FOR OPTRONIC EQUIPMENT|
    EP1920423A2|2005-09-01|2008-05-14|GeoSim Systems Ltd.|System and method for cost-effective, high-fidelity 3d-modeling of large-scale urban environments|
    US20070126867A1|2005-12-02|2007-06-07|Mccutchen David|High resolution surveillance camera|
    FR2899344B1|2006-04-03|2008-08-15|Eads Astrium Sas Soc Par Actio|METHOD FOR RESTITUTION OF MOVEMENTS OF THE OPTICAL LINE OF AN OPTICAL INSTRUMENT|
    FR2952743A3|2009-11-19|2011-05-20|Astrium Sas|METHOD OF ESTIMATING THE MOTION OF A SCREEN OBSERVING INSTRUMENT OVERLOOKING A CELESTED BODY|
    IL207590A|2010-08-12|2016-11-30|Rafael Advanced Defense Systems Ltd|Method and system for increasing the size of the area scanned by an airborne electro-optic reconnaissance system in a given time|
    RU107856U1|2011-05-16|2011-08-27|Государственное образовательное учреждение высшего профессионального образования "Московский государственный технический университет имени Н.Э. Баумана" |MOBILE AUTOMATED SYSTEM OF AEROVISUAL SURVEY OF TERRITORIES|
    RU2498378C1|2012-06-21|2013-11-10|Александр Николаевич Барышников|Method of obtaining image of earth's surface from moving vehicle and apparatus for realising said method|
    JP6211876B2|2013-10-01|2017-10-11|株式会社トプコン|Measuring method and measuring device|
    CN104061907B|2014-07-16|2016-08-24|中南大学|The most variable gait recognition method in visual angle based on the coupling synthesis of gait three-D profile|
    CN104567815B|2014-12-26|2017-04-19|北京航天控制仪器研究所|Image-matching-based automatic reconnaissance system of unmanned aerial vehicle mounted photoelectric stabilization platform|US11067996B2|2016-09-08|2021-07-20|Siemens Industry Software Inc.|Event-driven region of interest management|
    US10802450B2|2016-09-08|2020-10-13|Mentor Graphics Corporation|Sensor event detection and fusion|
    CN109541595A|2018-11-14|2019-03-29|北京遥感设备研究所|A kind of velocity error modification method and system based on Circular scanning radar images match|
    IL270565D0|2019-11-11|2021-05-31|Israel Aerospace Ind Ltd|Systems and methods of image acquisition|
    法律状态:
    2016-10-28| PLFP| Fee payment|Year of fee payment: 2 |
    2017-05-26| PLSC| Search report ready|Effective date: 20170526 |
    2017-10-26| PLFP| Fee payment|Year of fee payment: 3 |
    2018-10-26| PLFP| Fee payment|Year of fee payment: 4 |
    2019-10-29| PLFP| Fee payment|Year of fee payment: 5 |
    2020-10-26| PLFP| Fee payment|Year of fee payment: 6 |
    优先权:
    申请号 | 申请日 | 专利标题
    FR1502431A|FR3044101B1|2015-11-20|2015-11-20|PROCESS FOR ACQUIRING IMAGES OF A SCENE FROM A SENSOR ON BOARD A MOVING CARRIER, WITH CONTROL OF ITS SIGHT LINE|
    FR1502431|2015-11-20|FR1502431A| FR3044101B1|2015-11-20|2015-11-20|PROCESS FOR ACQUIRING IMAGES OF A SCENE FROM A SENSOR ON BOARD A MOVING CARRIER, WITH CONTROL OF ITS SIGHT LINE|
    KR1020187014228A| KR20180084804A|2015-11-20|2016-11-18|A method for acquiring images of a scene from a sensor mounted on a moving carrier,|
    RU2018121631A| RU2725776C2|2015-11-20|2016-11-18|Method of obtaining images of a survey object from a sensor onboard a moving carrier, with automatic adjustment of its line of sight|
    CA3005390A| CA3005390A1|2015-11-20|2016-11-18|Method for acquiring images of a scene, from a sensor on board a moving carrier, with servocontrol of its line of sight|
    PCT/EP2016/078146| WO2017085259A1|2015-11-20|2016-11-18|Method for acquiring images of a scene, from a sensor on board a moving carrier, with servocontrol of its line of sight|
    ES16801744T| ES2774080T3|2015-11-20|2016-11-18|Procedure for acquiring images of a scene, from a sensor on board a moving carrier, with servo control of its line of sight|
    US15/774,251| US10616492B2|2015-11-20|2016-11-18|Method for acquiring images of a scene, from a sensor on board a moving carrier, with servocontrol of its line of sight|
    CN201680068026.7A| CN108351208B|2015-11-20|2016-11-18|Method for acquiring scene images from a sensor of a moving carrier with automatic control of the line of sight|
    EP16801744.0A| EP3377854B1|2015-11-20|2016-11-18|Method for acquiring images of a scene, from a sensor on board a moving carrier, with servocontrol of its line of sight|
    IL259208A| IL259208A|2015-11-20|2018-05-08|Method for acquiring images of a scene, from a sensor on board a moving carrier, with servocontrol of its line of sight|
    [返回顶部]