WO2008081115A2

WO2008081115A2 - Method and device for taking shots using a plurality of optoelectronic sensors

Info

Publication number: WO2008081115A2
Application number: PCT/FR2007/002002
Authority: WO
Inventors: Didier Debons; Laurent Lucas; Yannick Remion
Original assignee: Telerelief; Universite De Reims Champagne Ardenne
Priority date: 2006-12-06
Filing date: 2007-12-05
Publication date: 2008-07-10
Also published as: FR2909778A1; FR2909778B1; WO2008081115A3

Abstract

The method of the invention can be used for taking shots using a plurality of optoelectronic sensors (Oi to O'i) in order to create a final sequence of combined images that can be displayed on screens adapted to this image combination. It comprises the physical or virtual distribution of the optoelectronic sensors (Oi to O'i) on a virtual arc of circle having a radius R corresponding to the distance between the optical centre of each of the sensors (Oi to O'i) and the point (C) in the area of the scene shot on which at least one of said sensors, or reference sensor, is focused, the length of the chord (Cori) between two successive optical centres being constant whatever the length of the radius R.

Description

METHOD AND DEVICE FOR TAKING OF REALIZATION G ¹ USING MULTIPLE SENSORS OPTOELECTRONIC.

The present invention relates to a method and a device for taking pictures with the aid of a plurality of optoelectronic sensors and for processing the information supplied by these sensors in order to create a final sequence of combined, displayable images. on screens adapted to this combination of images.

It applies in particular, but not exclusively, to the real-time broadcast of still images and / or animated images on auto-stereoscopic screens.

In general, it is known that the diffusion of relief images visible without glasses makes use of so-called "auto-stereoscopic" screen techniques whose general principle consists in simultaneously broadcasting one or more pairs of stereoscopic beams of so that the left sub-beam of each pair of beams for the left eye is not visible by the right eye and vice versa.

Contrary to the solution consisting in diffusing only a single pair of beams, a solution which implies that the user is in the axis of the screen to visualize the images in relief, the solution consisting in diffusing several pairs of simultaneous stereoscopic beams allows multiple people to simultaneously view images at a relatively large viewing angle.

To produce static photographic images or kinetic image sequences that can be broadcast on these types of screens, there are currently two types of means:

The use of computer graphics software. Bi-camera shots.

In fact, the common feature of CGI tools is to be able to produce three-dimensional images that are finally broadcast in two dimensions because, until now, display did not allow to consider a diffusion in three dimensions in relief; the images broadcast in stereoscopy allowed, with the help of glasses, to restore the relief effect.

Nevertheless, these computer-generated images are adaptable to the production intended for the new range of so-called auto-stereoscopic screens.

This is not the case with regard to the production of "real" images using images taken using a bi-camera system.

In fact, to obtain a realistic and natural effect of relief vision, it is necessary to be able to simultaneously broadcast several pairs of stereoscopic beams each having a different angle of view, ie a total shooting taken with the aid of several synchronized cameras. operating simultaneously.

At this stage, one encounters the difficulty of capturing several synchronous and homogeneous shots with an equidistant inter-objective space specific to each configuration of shots to obtain the desired relief effect.

The difficulties encountered are:

- of a physical nature: they essentially result from the crowding either of the cameras, the objectives, or both,

- Video order: difficulty in getting a synchronization of the signals picked up by the different cameras, as well as good homogeneity and appropriate geometry of the images delivered by each of the cameras.

The invention therefore more particularly aims to solve the problems mentioned above.

Thus, for the purpose of solving video-order problems, the method according to the invention consists of distributing, physically or virtually, a plurality of optoelectronic sensors on a line in order to create a final sequence of combined images that can be displayed on screens adapted to this combination of images, characterized in that it comprises the distribution and the orientation of a plurality of electronic sensors on a line, so that the angles formed between the optical axes of two sensors consecutive are equal and converge to a point C selected on the area of the scene to be captured on which is carried out the development of at least one of the sensors called reference sensor on which is carried out the focus of least one of said sensors referred to as reference sensor.

In the case of a virtual distribution, the optoelectronic sensors are equipped with lenses with variable focal length adjusted to compensate for variations in distance between each of the sensors and the aforementioned point on which the development of references is carried out.

The method comprises other features as defined in claims 1 to 13

The invention also relates to a device for implementing the previously defined method, this device involving:

a fixed support structure equipped with means for adjusting the horizontality, this support structure comprising a fixed support structure equipped with horizontality adjustment means, this support structure comprising "on behalf of the means of distribution of a plurality optoelectronic sensors (Oi, O'i) along a line so that the angles formed between the optical axes of two consecutive sensors are equal and convergent towards a point C chosen on the area of the scene to be captured and on which a setting of point of at least one of the sensors called reference sensor has been made.

Advantageously, the fixed support structure may comprise a rigid beam defining the first axis OX and along which the movable structures are slidably mounted. This beam can be equipped with micrometric transport systems to position very precisely and independently each of the mobile structures.

The mobile structures may comprise rigid arm guiding means constituting the aforesaid support elements along the second axis OY. Each of these arms may be provided with a micrometric transport system for sliding along the axis OY independently of the other arms. Each of the arms may comprise a micrometric system of rotation about a vertical axis on which is fixed a camera system whose optoelectronic sensor is centered on the axis of rotation.

Thanks to these arrangements, the device according to the invention makes it possible to guarantee: for a given angle α, formed between the axes of two consecutive optoelectronic sensors, a perfect equidistance between the focus reference area of the captured scene and each of the "N" optoelectronic sensors intended to capture the image of the scene, that the length of the cord of the circular arc specifically determined between two optical centers of two successive optoelectronic sensors remains identical regardless of the length of the radius of a circle constituted by between the "n" sensors and the filmed scene.

Of course, the invention relates to a processing of the images delivered by the optoelectronic sensors, this processing consisting of: to capture in an orderly manner, to process and to assemble the images of the "n" electronic sensors, to modify these images in digital forms, to assemble as is or by cutting the "x" images to create a final image corresponding to the target auto-stereoscopic screen type according to the manufacturer's recommendations.

One embodiment of the invention will be described below, by way of non-limiting example, with reference to the accompanying drawings in which:

Figure 1 is a schematic representation of an image pickup device;

Fig. 2 is an example of an arrangement of optoelectronic sensors with respect to a focus center;

Fig. 3 is a flowchart showing an operating sequence executed for the processing of the raw images;

Figure 4 is a schematic representation of a physical positioning system;

Fig. 5 is a block diagram showing the useful part of a raw image having a marker; Figure 6 is a block diagram of a self-positioning system of the useful frame.

In the example illustrated in FIG. 1, the device for taking pictures involves a support structure represented here schematically by a horizontal beam 1 of axis OX. This beam 1 can be mounted on a base by means for adjusting its orientation.

On this beam 1 is mounted axially sliding a plurality of carriages d (here, by way of non-limiting example three carriages Ci to C ₃ ) whose displacements are each controlled by a respective micrometric actuator Mj M ₁ to M ₃ independent of each other. others (here, by way of non-limiting example three screw jacks whose bodies are integral with the beam and whose rods are integral with the carriages).

Each carriage Cj comprises a system of slides in which slides an oriented arm Bj, respectively for the example Bi to B _3, along an axis OY perpendicular to the axis OX. The movements of the arm along this axis OY are provided by a micrometric actuator Aj, respectively for example A ₁ to A ₃ , for example rack, comprising a gear motor driving a pinion which meshes with a longitudinal rack fitted to the arm,

At one of the ends of each of the arms Bj there is arranged a rotary support Si, respectively for the example Si at S ₃ , pivoting about a corresponding vertical axis Zj (trace Z ^ Z ₃ ) on which a system is mounted. of shots whose optoelectronic sensor Oj, respectively for example O ₁ to O ₃ , is centered on the respective vertical axis Z ₁ to Z ₃ .

The actuators Mj, for example, Mi to M3 and Aj, for example A ₁ to A ₃ , are controlled by a calculation unit UC so as to be able to arrange the optoelectronic sensors according to one of the configurations illustrated in FIG. 2.

This figure represents the distribution of the sensors with respect to a dashed median line, and the shooting device of FIG. 2 comprises eight sensors of which only the four located to the left of the line. median and the fifth of the group of four, located to the right of this median line and symmetrically, are represented.

Five of the eight optoelectronic sensors Oi to O ₅ , arranged on a circle CE of center C, have thus been represented, it being understood that the number of sensors, in other exemplary embodiments, Oi, O _h On, may be different. Subsequently, he will be designated by "n",

. Point C is a point on which optoelectronic sensors are developed, for example by means of a self-focusing system.

. The optical center of each of the optoelectronic sensors Oj in the example Oi to O ₅ , is arranged on the circle CE, so that the distance between the optical centers of the optoelectronic sensors Oi in the example Oi to O ₅ , and the point C is equal.

The angles α formed between the optical axes passing through the optical centers of two consecutive optoelectronic sensors and the point C are equal, so that the length of the cord of the circular arc Cori in the example "COr ₁ to Cor ₄ "between two optical centers of two consecutive optoelectronic sensors Oj, O _{i +} i remains identical regardless of the length of the radius R of the circle CE (formed between the optoelectronic sensors Oj and the focusing center C of the scene filmed) .

The calculation unit UC is programmed so as to determine the positions of the sensors Oj in the example O ₁ to O ₅ on the circle (virtual) CE and their orientation according to the desired focus center C.

This focusing center C is determined, for example by telemetric sighting or possibly by means of one of the sensors Oj in the example O ₁ to O ₅ (reference sensors), with which a setting is made. on point. The telemetry device associated with the self-focusing system of this sensor makes it possible to determine the radius R of the virtual circle, then the value of the angle at the center α and / or of the chord Coη in the example Cori to Cor ₄ of the circle corresponding to the distance of the constant value separating each of the sensors Oj in the example Oi to O ₅ between them on the circumference of said circle CE.

The calculation unit UC then determines the coordinates OX OY of each of the optical centers of the optoelectronic sensors in the coordinate system XOY in which the actuators Mi, Mi to M ₃ and Ai, Ai to A ₃ operate, and then addresses to each of these actuators the corresponding commands.

Once the optical centers of the sensors Oj, in the example Oi to O ₅ , arranged in their respective locations on the circle C, the computing unit accordingly controls the orientation of the sensors O ₁ in the example Oi to O _5, so that they focus on the center of focus C.

The calculations made for this purpose by the calculation unit UC are simple trigonometric calculations that do not need to be explained in detail.

Synchronous raw images delivered by shooting systems associated with the sensors for example Oj Oi O ₅ optoelectronic subject to a processing sequence of capturing, processing and assembling the images of "n" electronic sensors , to modify these images in digital forms, to assemble as is or by cutting the "n" images for the purpose of creating a final image corresponding to the type of target auto-stereoscopic screen, according to the manufacturer's recommendations.

According to a particularly advantageous embodiment of the method according to the invention, the raw images of the "n" electronic sensors are recovered in gray value or in color. In the case of grayscale images, it is possible to use a method of converting grayscale images to color images of the type described in US Patent No. 3971065.

In the example illustrated in FIG. 2, a simplified embodiment of the system in which the optoelectronic sensors O'j in the example O'1 to O ' ₅ are also equipped with FIG. a lens with variable focal length and where the arms B _s have been removed. In this embodiment, these optoelectronic sensors are mounted pivoting on the carriages C ₁ in the exampleCI to C ₅ , mobile on the beam 1.

The sensors O'j are then positioned on the axis OX, so as to respect the equality of the angles α, as represented by the positions O'i to O ' _δ and the distance between each sensor is no longer constant.

The difference in the distance between the position of the sensors O 1 to O ₅ and the center of focus C with the length of the radius R is individually compensated by an adjustment of the focal length of the lens of each of the sensors.

Furthermore, the processing of the raw images delivered by the optoelectronic sensors involves, as illustrated in FIG. 3, a synchronization signal generator 3 which defines for each shot: the absolute time of the occurrence of said synchronization signal, an identical pause time for each optoelectronic sensor.

The synchronization signal generator 3 is controlled by a trigger 4 manual or automatic shooting and optionally by an operating control system including entering the number of processing sequences (burst) to run.

It gives an order of storage of the image to the electronic sensors Oj, in the example Oi to O ₅ (time delay / opening) (block 5).

Following this order of storage, the images stored in the "n" optoelectronic sensors Oi, in the example Oi to O ₅ , are collected sequentially in raw format (called "RAW") and are stored in temporary memory (memory raw image input buffer) (block 6) pending subsequent processing. This group of synchronous images, that is to say the set of raw images taken at the same time by the "n" optoelectronic sensors, is called "image train".

The processing system performs cropping (block 7) and then adjusts each raw image keeping only the useful part extracted from each of the "n" images in elementary images (conversion of the "n" useful images into elementary images) (block 8).

The "n" elementary, complete or cut images (block 9) are then distributed according to a distribution grid specific to each type of auto-stereoscopic screens for the production of the black and white composite image (block 10). The digital information relating to the composite image is converted to "Bayer" to form a final composite image in color (block 11).

This image is then the subject of a specific treatment in order to reduce more or less the amount of digital information characterizing it using standard programs (MJPEG, MPEG ...) or specifically developed. The choice of the treatment applied to each image constituting a sequence is a function of the subsequent use envisaged.

In the case of a simple live broadcast, for example over an IP network, the reduction may be high or very high if the degradation of the signal does not affect the readability of the image sequence. In the case where the sequence is recorded and therefore intended for subsequent reuse such as for example the editing of several different shots, it will be necessary to favor a "non-destructive" encoding mode so as to obtain the assurance that the quality of generations of successive images will not be affected.

This more or less compressed image (block 11 ') is then stored in the output buffer (block 12). Triggering the recording and / or sending to the broadcast network (block 13) can then be performed. The steps of conversion into elementary images (block 8) and fabrication of the composite B / W image (block 10) are controlled from specific parameters of the grid that is to be used (block 13 ').

The rendering of the relief effect depends on a set of adjustments specific to each scene captured whether it is fixed or moving, the support system itself being fixed or moving relative to the captured scene.

As a result, a number of sensor positioning adjustments are required before shooting and, if necessary, shooting.

So that the desired relief effect is obtained, it is necessary to perform a number of actions such as:

. physical actions on sensor positioning optoelectronic, physical actions on the optical settings of these sensors, virtual actions on the raw image delivered by the sensors.

As previously mentioned, the calculation unit UC determines, for each camera configuration, from the measurement of distance between the captured scene or element of the captured scene and the optical of the reference sensor taken in a physical manner or with the aid of a rangefinder independent or dependent on the entire carrier device or in the case of a complete system: the X coordinates on the beam of the positioning of each arm Bj to B3,

the coordinates Y on each of the arms B] to B3 of the positioning of each of the "n" micrometric rotation systems of each of the "n optoelectronic sensors CM to O ₃ ,

. the value of the rotation angle R of each of the "n" rotary supports of the sensors O _\ is, in the case of a simplified system (without arms) in which the displacements of the arms are replaced by an adjustment of the focal lengths of the zooms equipping the photoelectronic sensors:

. the coordinates of each optoelectronic sensor,. the value of the orientation angle R of each of the "n" sensors towards the shooting center C.

This rule provides the XYR information to the motorized carriages Ci in the example, C ₁ to C ₃ of each arm, Bj in the example, Bi to B ₃ and each of the "n" micrometric rotation systems, allowing the positioning precise "n" optoelectronic image sensors O _, in the example O ₁ to O ₃ .

In use, the human or robotic operator informs the management system of the raw image processing system of the value of the radius R of the circle CE so that subsequently the CPU calculation unit controls, via interface ( s), motors or actuators of the actuators. The use of a telemeter integral with the reference sensor, connected to the CPU calculation unit, allows the device to be implemented quickly for shooting and to be dynamic to recalculate the XYR coordinates in the case of movement and / or displacement . These coordinates are recalculated in real time if: the carrier device is moving relative to a fixed element of the captured scene, the carrier device is fixed or in motion and the relief effect is modified. In any case, the rangefinder can be placed at any point known by the computer. It can in particular be coupled to one of the "n" optoelectronic sensors because the computer knows the XYR coordinates of the sensor used.

The telemetry function can be integrated with one or more of the "n" optoelectronic sensors.

The positioning of the overlapping images for which the optical centers are to be confused takes place in at least two steps:

. a necessary and obligatory step of physically positioning the "n" optoelectronic sensors Oj in the example Oi to O ₃ , or O'-i established according to the scene captured and the desired relief effect obtained by the intermediate of the support structure described above, a complementary step considering the one or two optoelectronic sensor (s) positioned (s) most in the center of the support structure 1 as referent; this step consists in refining the preceding step by an operation performed in the raw image recovered from one of the other optoelectronic sensors, during the processing of the raw images, by moving the virtual framing, this within the limit of the number of pixels remaining between the frame of the useful image and the out-of-frame of the raw image, whether in height and / or in width. This step, repeated for as many optoelectronic sensors as necessary, is called "virtual positioning". This virtual positioning can be done:

. by manual displacement of the framing of the useful image in the raw image; by a pattern recognition method according to one or more marks added or determined in the scene inside or outside the frame of the useful image Moreover, the physical positioning of the sensors (automated carrier device) can be ensured notably by: a pattern recognition method according to one or more selected markers in the captured scene, integrated into the useful image; to a tracking system of a moving landmark chosen in the captured scene, known as Anglo-Saxon "tracking".

Each system associated with an optoelectronic sensor can be autonomous in its calculation of image retrieval and repositioning, insofar as all the systems communicate with each other their positioning information, while being in solidarity with a physical or virtual structure.

This is particularly valid when the radius R is important, for example of the order of several kilometers or more (shooting distant landscapes).

Each optoelectronic sensor can be extremely small size for microscopic shots. It may, for example, include an optical system comprising one or more microlenses connected to an optoelectronic cell via a light guide such as one or more optical fibers. Thanks to these arrangements, only the optics of the sensor will be arranged on the circle CE.

The device described above may include the following adjustment modes:

Focus mode: This is the focus control of the image formed on the "n" optoelectronic sensors Oj to O ₅ . This development can be done:

- manually, by action on each of the objectives associated with the sensors,

- automatically, thanks to an autofocus system equipping each objective,

by controlling the debugging servocontrols associated with each sensor, by means of a general command having a single value for all the objectives associated with the "n" sensors electronic,

by controlling the debugging servocontrols associated with each sensor, by means of commands presenting a specific value for each of the objectives associated with the "n" electronic sensors.

Iris mode: This mode concerns the iris control of the optics associated with each of the sensors. It allows a direct manual control on the lens of each of these optics or a slave control.

Exposure mode: This setting mode is used to set the exposure time or exposure time common to each camera system.

Burst mode: This mode allows you to select the number of images to be processed in a fraction of a second (basic period) to obtain the final composite image.

Rendering mode Relief: This is to allow an adjustment of the length of the radius R (focusing distance) and the length of the rope Cor (determining a convergence index).

To perform the physical positioning of the optoelectronic sensors, the computing unit UC may comprise a physical positioning calculator comprising means for:

- receive metric or telemetry information,

receiving information modifying the telemetric information so that the desired relief effect is obtained,

calculate the values of the positions of each optoelectronic sensor,

- drive the actuators Mj to M ₃ and Aj to A ₃ or the like,

optionally, control the motors of the zoom functions of the lenses of the optoelectronic sensors,

- optionally, memorize all or part of this information,

- optionally, make this information available to third parties by standard digital communications techniques in accordance with the standards in force or by proprietary communication interface, - Optionally, control the motors that provide the focus, iris and zoom functions of the optoelectronic sensors.

FIG. 4 shows a theoretical block diagram of a physical positioning system involving a control interface IC of the actuators AX ₁ AY, AR along the axes OX, OY and in rotation of each of the optoelectronic sensors O.

This IC interface is controlled by a CPU calculation unit and / or an EX operating control system. The calculation unit UC is coupled to a TE rangefinder for providing information relating to the focusing distance of at least one of the optoelectronic sensors O. The operating control system can itself be coupled to an input or calculation device OS of radius R (focusing distance).

The calculation unit UC may furthermore comprise virtual positioning means whose function is to display the positioning performed by a positioning computer of the optoelectronic sensors by interacting on the virtual framing of the useful image in the raw image. . These virtual positioning means are intended in particular to analyze the raw image or the reference useful image, to extract a predetermined mark or chosen by the operator, then to recognize this marker in the images from other sensors optoelectronic to act on the displacement or replacement of the virtual frame.

This virtual positioning means makes it possible, in particular, to position the frame of the useful image in the raw image coming from the reference optoelectronic sensor by means of a raw image processing method PTI, possibly supplemented by a method of image analysis incorporating a shape recognition function.

In manual mode, the operator positions via a displacement command (in the XOY plane) the virtual frame of the reference useful area in the raw image of the optoelectronic reference sensor. The virtual positioning means analyze the image, make a recognition of one or more shapes and calculate the coordinates of this or these forms with respect to the frame of the useful zone. The virtual positioning means analyze the raw images of the other electronic sensors and, by image trains, proceed to the recognition of the form (s) identified in each raw image and proceed to the positioning of the frames of the zones. useful in the raw images of the "n" optoelectronic sensors according to the xy coordinates stored at each image train.

In automatic mode, the operator visualizes the scene captured from the useful part or the raw image of the optoelectronic reference sensor (frame / out of frame). The system then positions the useful frames in the "n" raw images. This positioning may, for example, be centered, that is to say with the framework of the useful area centered on the raw reference image. The virtual positioning means determines, by image analysis and shape recognition, one or more Obi markers (from pre-established data) in the raw (or useful) reference image (FIG. 5). In principle, only the marks in the useful part of all the images are exploitable because it is possible that one or more references existing in the raw reference image is absent in one or more raw images of other optoelectronic sensors. The positioning means use the coordinates Pix, Pxy of the first mark for the positioning of the frame CAi of useful area in the raw image of each of the "n" electronic sensors and repeats this operation for as many bursts.

The virtual positioning means in particular make it possible to determine the distances QN, QE, QO and QS between the edges of the frame of the useful image CAi and the edges of the raw image CA2 (FIG. 5). If one of these distances is too small, or even negative, less than positioning tolerances, this means that the optoelectronic sensor that formed the image is incorrectly positioned. The operator or the physical positioning means may, in this case, reposition the sensor. The positioning tolerances may be different depending on the scene to be captured, depending on the raw image, even if the latter is different from one optoelectronic sensor to another and especially when the scene is animated and / or according to the desired relief effect.

As illustrated in FIG. 6, the raw images delivered by the optoelectronic sensors (block 21) are first transferred to an input buffer (block 22). These images are then transmitted to a form recognition system SF comprising two processing chains, namely:

A first processing chain of the raw image of the reference sensor which comprises successively: an analysis of the raw image of the reference sensor (block 23), a recognition of the mark or references (block 24), a storage of the form of each marker (block 25), a calculation of the coordinates of each marker (block 26),

A second processing chain of the raw images of all the sensors, this chain including: the analysis of the raw images of "n" sensors (block 27), the recognition of each reference mark according to the x and y coordinates determined at the block; 26 (block 28).

After recognizing the markers present in the raw images, the system performs a framing of the useful area in the raw image (block 29), then determines the elementary images (block 30). Then, he assembles the elementary images in a distribution grid to form the final black and white image (block 31). By a "Bayer" conversion, it transforms this final image into a color image (block 32). The system finally places this color image in the output buffer (block 33).

The recognition system of the objective marker (s) determines whether the QE, QN and QS values are greater than a positioning tolerance threshold. If these values are below said threshold, the system performs a positioning calculation of the optoelectronic sensor concerned, so as to correct its positioning (block 34).

It should be obvious to those skilled in the art that the present invention allows embodiments in many other specific forms without departing from the scope of the invention. the invention as claimed. Therefore, the present embodiments should be considered by way of illustration, but may be modified within the scope defined by the scope of the appended claims, and the invention should not be limited to the details given above.

Claims

claims

A method for performing shooting with a plurality of optoelectronic sensors (Oi) to create a final sequence of combined images displayable on screens adapted to this combination of images, characterized in that it comprises the distribution and the orientation of a plurality of electronic sensors on a line, so that the angles formed between the optical axes of two consecutive sensors are equal and converge towards a point C chosen on the zone of the scene to be sensed on which is carried out the development of at least one of the sensors called reference sensor,

it also includes

a grayscale capture step, in an orderly and synchronized manner of each raw image supplied by the "n" electronic sensors (Oi) to form a raw image stream;

a step of framing the useful part of each raw image of a train and the extraction of this useful part which then constitutes an elementary image;

a final transformation step of the black and white image into a color image by a "Bayer" transformation;

2. Method according to claim 1, characterized in that it further comprises the operation of assembling the images of a stream of images for the purpose of creating a final image corresponding to the type of target screen to produce an effect relief corresponds to a distribution operation of the elementary images according to a distribution grid specific to a type of auto stereoscopic screen determined so as to obtain a final image.

3. Method according to claim 2, characterized in that to obtain the desired relief effect, the method comprises:

- actions on the distribution and orientation of the optoelectronic sensors and / or actions on the optical settings of these sensors.

4. Method according to one of the preceding claims, characterized in that the processing and modification step comprises a step of virtual positioning of the overlapping images for which the optical centers must be merged, this step comprising the physical positioning of the optoelectronic sensors in function of the captured scene, of the sought-after relief effect, the refining of this positioning by a processing of the raw image coming from at least one sensor other than the reference sensor, this processing consisting in moving a virtual frame in the limit of the number of pixels remaining between the frame of the useful image and the outermost frame of the raw image.

5. Method according to the preceding claim, characterized in that the aforesaid virtual positioning is performed by manual movements of the framing of the useful image in the raw image.

6. Method according to one of the preceding claims, characterized in that the synchronization of the capture of the images from the optoelectronic sensors is obtained by a common synchronization signal with definition for each shot: the absolute time of the occurrence of said signal of synchronization and an identical exposure time for each optoelectronic sensor.

7. Method according to one of the preceding claims, characterized in that a computing device associated with a telemeter integral with the reference sensor recalculates the coordinates and angles to redefine the distribution and orientation of the sensors. in the case of movements or relative movements of the shooting system with respect to the area of the scene to be captured.

8. Method according to one of the preceding claims, characterized in that the step of framing uses a pattern recognition method according to one or more pins added or determined in the scene inside or outside the frame of the useful image.

9. Method according to claim 1, characterized in that in the case of a distribution on a straight line, the optoelectronic sensors (Oi) are equipped with lenses with variable focal length adjusted to compensate for variations in distance between each of the sensors and the point C on which the reference focus is performed.

10. Method according to claim 1, characterized in that in the case of a distribution on a circular line the length of the cord (CORi) separating two consecutive optical centers of adjacent optoelectronic sensors (Oi), (Oi + 1) is kept constant whatever the distance R between the optical center of the sensor and the point C on which the focusing is carried out.

11. Method according to one of the preceding claims, characterized in that the focus center C is determined by the reference sensor and in that the distance to the point of convergence is determined by a telemetry device associated with the car system. focusing of said reference sensor.

12. Method according to one of the preceding claims, characterized in that said sensors (Oi, O'i) and their possible objectives are actuated by actuators and the method comprises determining the coordinates of each of the optical centers of the electronic sensors (Oi , O'i) in the coordinate system in which the actuators operate and then the transmission to the actuators of the corresponding commands.

13. Device for implementing the method according to one of the preceding claims, characterized in that it comprises:

a fixed support structure equipped with horizontal adjustment means, said support structure comprising "at least means for distributing a plurality of optoelectronic sensors (Oi, O'i) in a line such that the angles formed between the optical axes of two consecutive sensors are equal and convergent towards a point C chosen on the area of the scene to be captured and on which a focus of at least one of the sensors referred to as the reference sensor has been performed ";

means for triggering grayscale capture in an orderly and synchronized manner of each raw image produced by the N electronic sensors (Oi) to form a stored raw image stream;

means for framing the useful part of each raw image of a train;

extraction means of this useful part which then constitutes an elementary image

means for transforming the black and white image into a final color image by a "Bayer" transformation.

14. Device according to claim 13 characterized in that it comprises means for assembling images of a stream of images for the purpose of creating a final image corresponding to the target screen type to produce a relief effect made by means of distribution of the elementary images according to a distribution grid specific to a type of auto stereoscopic screen determined to obtain a final image.

15. Device according to claim 13 or 14, characterized in that the distribution means of the optoelectronic sensors are actuators, actuating the positioning of the sensors and the possible objectives of these sensors, these actuators being controlled by a central unit, a unit calculation method determining the coordinates of each of the optical centers of the optoelectronic sensors in the coordinate system in which the actuators operate and then transmitting the corresponding commands to the actuators.

16. Device according to one of claims 13 to 15, characterized in that the computation unit controls the synchronization of the optoelectronic sensors by a common synchronization signal with definition for each shot of an image stream of the absolute time of the occurrence of said synchronization signal, and an identical exposure time for each optoelectronic sensor (Oi).

17. Device according to one of claims 13 to 16, characterized in that the computing unit performs the digital processing of raw images, processing which comprises a virtual positioning of the overlapping images, the refinement of this positioning by means of a shape recognition system associated with one or more markers added or determined in the scene inside or outside the useful image.

18. Device according to claim 13, characterized in that the fixed support structure comprises a rigid beam (1) defining the first axis OX and along which the movable structures (Ci) are slidably mounted, this beam (1) being equipped of systems micrometric transport for positioning very precisely and independently each of the mobile structures (Ci).

19. Device according to claim 18, characterized in that the said movable structures (Ci) comprise rigid arm guide means (Bi) which constitute the said support elements, along the second axis OY, each of these arms (Bi). ) being provided with a micrometric transport system allowing it to slide along the axis OY independently of the other arms.

20. Device according to claim 19, characterized in that each of the arms (Bi) comprises a micrometric system of rotation about a vertical axis on which is fixed a camera system whose optoelectronic sensor (Oi) is centered on the axis of rotation (Zj).

21. Device according to one of claims 13 to 21, characterized in that the said optoelectronic sensors (Oi, O'i) are of reduced size and each comprise an optics comprising one or more microlenses connected to an optoelectronic cell by the intermediate of a light guide