HEAD-UP DISPLAY TYPE AUTOSTEREOSCOPIC SYSTEM

Abstract

AUTOSTEREOSCOPIC SYSTEM OF THE HEAD-UP DISPLAY TYPE The invention relates to a system for the dematerialized projection of an autostereoscopic image intended for an observer whose eyes are positioned within a predetermined three-dimensional space, called an eye box, said system comprising: an optical device comprising at least one converging optical component arranged relative to said eye box so as to be able to direct the light rays that it receives towards said eye box; an autostereoscopic screen arranged relative to said optical device so that the light rays that it emits can reach said converging optical component, and comprising: a matrix of pixels and a lenticular network surmounting said matrix, said network having a pitch P determined from a known pitch P' of a screen intended to be observed in direct vision. Figure for the abstract: figure 1

Description

Title of the invention: AUTOSTEREOSCOPIC SYSTEM OF THE HEAD-UP DISPLAY TYPE Technical field of the invention

[0001] The invention relates to an autostereoscopic system, of the head-up display type, allowing at least one observer positioned in front of the display system to observe an aerial and dematerialized projection of an autostereoscopic image. Technological background

[0002] Autostereoscopy is a technique for displaying relief images without requiring the observer to wear special glasses. This technique is known, in particular from patent documents WO2006 / 024764, WO2014 / 041504, WO2013 / 140363, WO2014 / 016768, WO2019207235 and WO2022175052 in the name of the applicant.

[0003] Generally speaking, an auto stereoscopic image consists of a plurality of bands of elementary images nested according to a predetermined mixing scheme, corresponding to views of the same object or the same scene from different points of view. A selector device, typically consisting of a network of cylindrical lenticules or a parallax barrier, is arranged in front of the display screen so as to allow the projection of a pair of elementary images corresponding to two different points of view of the scene, respectively towards the two eyes of the observer, which creates an impression of relief in the brain of the observer.

[0004] Autostereoscopic displays generally result from the addition of a glass optical component to the surface of an LCD panel. The size, thickness and weight of the glass are proportional to the size of the display. In practice, manufacturing and mounting a lenticular array on a display more than 2 m wide is so restrictive that manufacturers rarely offer larger solutions.

[0005] Furthermore, it is very rarely possible to display content whose total depth exceeds the diagonal of the screen, with clarity and without ghosting. The content must therefore be designed in such a way as to limit its volume. Consequently, multi-camera systems, whether physical hardware for shooting or software cameras, are often associated with geometric parameterization models, which correlate the stereoscopic base and the focal length, in order to control the distance at which the most distant backgrounds are located in a scene. The largest autostereoscopic screens available can display a scene with a maximum depth of 3 m.

[0006] The inventors sought to overcome the limits of current autostereoscopic screens to propose a solution allowing the projection of large images with a very large depth of field. Objectives of the invention

[0007] The invention thus aims to provide an autostereoscopic system which makes it possible to project large images (for example 4 meters wide) with a very large depth of field (for example 100 meters deep).

[0008] The invention also aims to provide an autostereoscopic system which allows an observer to benefit from binocular vision close to reality.

[0009] The invention also aims to provide such a system which makes it possible to display content which can retain orthostereoscopic proportions.

[0010] The invention also aims to provide, in at least one embodiment, such a system which allows consistency of the proportions of the autostereoscopic image projected by the system with the proportions of the real elements of the environment close to the system.

[0011] The invention also aims to provide, in at least one embodiment, such a system which can be used for augmented reality applications.

[0012] The invention also aims to provide, in at least one embodiment, a compact autostereoscopic system, which can for example be used for automotive applications.

[0013] The invention also aims to provide a versatile system which can be used in different configurations for various applications. Statement of the invention

[0014] To do this, the invention relates to a system for dematerialized projection of an autostereoscopic image intended for an observer whose eyes are positioned within a predetermined three-dimensional space, called an eye box, having a predetermined width L, at least greater than the average inter-pupillary distance of an observer and preferably greater than the average width of a human being's head.

[0015] The system according to the invention is characterized in that it comprises: - an optical device comprising at least one converging optical component with behavior identical to a concave mirror of radius of curvature R and focal length F, said optical device being arranged relative to said eye box so as to be able to direct the light rays that it receives towards a viewing window observable from said eye box, - an autostereoscopic screen arranged in relation to said optical device so that the light rays which it emits can reach said window of visualization, after reflection on said convergent optical component, and having traveled a distance d less than F, said auto stereoscopic screen further comprising: • a matrix of pixels of N rows and M columns, each pixel being composed of a plurality of sub-pixels of different colors, • a lenticular network surmounting said matrix and having a pitch P determined from a pitch P', called nominal pitch, of an autostereoscopic screen, called nominal screen, having usage characteristics identical to those of the screen of the system and intended to be observed in direct vision at a distance corresponding to the usage distance of said screen of the system, by the formula P=P\(\+CI(SM )) where C is the number of parts of viewpoints visible simultaneously from said usage distance inside said eye box, when said nominal screen of pitch P' is arranged in said projection system in place of said screen, and S is the number of different viewpoints of the image addressed by each pixel of said pixel matrix

[0016] The system according to the invention thus allows an observer placing his eyes within a predetermined three-dimensional space, called an eye box, to perceive a large-scale image, from a smaller-scale autostereoscopic screen (for example a 5.5-inch (13.97 cm) or 13.3-inch (33.782 cm) screen), and whose image has a large depth of field.

[0017] The observer benefits from an amplitude of relief unequalled by the autostereoscopic screens of the prior art and from a large-sized image whereas the autostereoscopic screen, which he does not perceive directly, may be of reduced size.

[0018] To obtain this result, the invention uses a specific autostereoscopic screen, arranged in relation to the optical device, in such a way that its reflection appears very enlarged behind the converging optical component, or behind a front or semi-transparent plane mirror making it possible to carry out an angle return of the screen's reflection.

[0019] The viewing window observable from said eye box and through which the observer observes the scene in relief can be physically delimited by the converging optical component or be formed by the reflection of the converging optical component in a front-facing mirror or a semi-transparent mirror inclined facing the observer and the converging optical component, preferably at an angle close to 45°.

[0020] Furthermore, the autostereoscopic screen is separated from the converging optical component (which behaves identically to a concave mirror of radius R and focal length F) by a distance d less than the focal length F of the optical component. When the distance from the screen to the optical component is strictly less than its focal length (equal to R / 2 where R denotes the radius of the mirror), it thus acts as a magnifying mirror. The screen is then reflected behind the optical component, in a more or less enlarged manner depending on its distance.

[0021] The system according to the invention thus implements a specific autostereoscopic screen specially designed for this sole purpose, to allow observation of the dematerialized reflection of an autostereoscopic image reflected by a convergent optical component. Such a convergent optical component is for example a concave mirror or a holographic optic. Throughout the following, the terminology of concave mirror is mainly used to facilitate reading, but it is understood that this concave mirror can be replaced by a holographic optic without modifying the technical effects of the invention.

[0022] In particular, an autostereoscopic screen designed for direct observation is not suitable for use within a device implementing a reflection of the screen on a concave mirror.

[0023] Indeed, such a screen dematerialization device must make it possible to find in the main lobe of the screen all the desired characteristics: number of viewpoints, flat tint distance and distribution of viewpoints on the screen pixel matrix.

[0024] The flat tone distance, which results from the moiré theory, designates the distance from the screen from which it is possible to observe only one of the multiple viewpoints of the image on the entire surface of the screen (regardless of the number of viewpoints of the image). This distance is unique for each model of autostereoscopic screen.

[0025] At the flat tint distance, an observer can see successively, with a single eye, the series of viewpoints used, which is repeated across the entire width of the screen. Each of the series thus observed corresponds to a lobe. Care must be taken to ensure that the observer's two eyes are in the same lobe.

[0026] For example, we want the screen to be observed at the preferred distance of 1 m (in other words, the screen has a flat tint distance of 1 m). We consider that the average gap between the two eyes of the observer is 6.5 cm. We want a succession of viewpoints every 3.25 cm of movement parallel to the plane of the screen at the flat tint distance. With 10 viewpoints, the lobe is 3.25 cm X 10, or 32.5 cm, this is the central lobe 0. After which, the series of 10 viewpoints reappears in the same order on each side of the main lobe. This forms the two lobes -1 and +1. And so on, moving further and further away from lobe 0 in continuity and right / left symmetry. Lobes -2 and +2 etc.

[0027] Practice shows that a concave mirror alters all of these characteristics so that an autostereoscopic screen designed for direct vision is little or not at all suitable for use in a device implementing a concave mirror.

[0028] The deformations induced by the concave mirror require the introduction of new rules for defining the parameters of the lenticular network, different from those of a network in direct vision. In order for the reflected immaterial double of the screen to have the desired autostereoscopic properties, it is necessary to modify the characteristics of the lenticular network of the screen. The presence of a concave mirror transforms all the parameters described previously. To maintain the desired characteristics, the pitch of the network must be increased compared to the pitch of an equivalent screen which would be observed directly without reflection by a concave mirror. The inventors have succeeded in determining the level of correction of the pitch to be made to allow this observation reflected in a concave mirror.

[0029] This increase in the pitch is obtained according to the invention from the pitch P' of an autostereoscopic screen, called a nominal screen, having usage characteristics identical to those of the system screen (i.e., same flat tint distance, same number of viewpoints, same focal length, etc.) and intended to be observed in direct vision at a distance corresponding to the usage distance of said system screen, by the formula P=P'.(1+C / (SM )) where C is the number of parts of viewpoints visible simultaneously from said usage distance (flat tint distance) inside the eye box, when said nominal screen of pitch P' is arranged in said projection system in place of said screen, S is the number of different viewpoints of the image addressed by a pixel of the pixel matrix and M the horizontal resolution of the pixel matrix.

[0030] As an example, for addressing to the sub-pixel from a matrix of pixels where each pixel comprises 3 sub-pixels of different color, then S is equal to 3. Thus, the pitch P is obtained by the formula P=P'.(1+C / 3.M).

[0031] If the same matrix is ​​used for whole pixel addressing, then S is equal to 1 and the pitch P is obtained by the formula P=P'.(1+C / M).

[0032] Throughout the following and to facilitate reading, we consider sub-pixel addressing and a matrix in which each pixel is formed of 3 sub-pixels, it being understood that the technical teachings provided apply mutatis mutandis to whole-pixel addressing and / or to matrices comprising a number of sub-pixels distinct from 3.

[0033] If a color chart displayed on a nominal autostereoscopic screen (whose usage parameters are identical to those of the system) is reflected in a concave mirror, the reflected image of this color chart, in return, presents at the flat tint initially chosen a number of color bars corresponding to a projected step having gained several lenses for the entire network compared to the screen.

[0034] For example, for a 10-viewpoint screen, if the observer perceives, at any distance from the screen, 10 color bars (assuming the screen displays one color per viewpoint), this means that he has deviated from the flat tint distance to the point that the projected pitch has increased or decreased by 10 sub-pixels, i.e., by the pitch of one lens. If the observer perceives 20 color bars, this means that the projected pitch has increased or decreased by 20 sub-pixels, i.e., by two lenses. If the screen displays a single color, this means that the observer is precisely at the flat tint distance.

[0035] To the extent that it is possible to know (by manufacturing) very exactly the pitch of the lenticular network of a nominal screen, for example to the nearest 100th of a micron, it becomes possible to correct the effect of the concave mirror on the optical path which connects the surface of the real screen to the user's eyes via the real dematerialized conjugated image, by machining a new lenticular network with the corrected pitch.

[0036] We see that this amounts to creating a flat tint in direct vision outside the projection device at a very long distance from the screen. This therefore amounts to increasing the pitch of the lenticular network compared to use in direct vision.

[0037] If a nominal screen with 10 viewpoints placed in the projection device allows 12 color bars to be viewed at the desired flat tint distance of 1 m for example (assuming the screen displays 10 distinct colors, each color corresponding to a viewpoint), and since the pitch of a lens corresponds to 10 colors in direct vision, 12 color bars in reflection in the concave mirror represent 12 / 10ths of the pitch of the lenses. It is necessary to increase the pitch of the lenticular network by this value to adapt it to the device according to the invention and to divide it by the total number of lenses to calculate the exact incidence on the pitch which will be applied to the machining of the lenticular network of the screen intended to be arranged in the projection device.

[0038] This makes it possible to obtain a system whose dematerialized auto stereoscopic image and reflected by the mirror allows us to find all the characteristics expected in the central lobe.

[0039] For the first time, it becomes possible to simply determine the pitch correction to be made to a nominal autostereoscopic screen, for its integration into a device implementing a converging optical component such as a concave mirror, by knowing the pitch of the nominal screen and by determining the number of viewpoint parts visible simultaneously from said usage distance. inside said eye box, when this screen is housed in the device according to the invention.

[0040] It therefore becomes possible for the first time to use an auto stereoscopic screen in association with a concave mirror.

[0041] As seen previously, this determination of the number of parts of viewpoints visible simultaneously can be facilitated by displaying a color chart on the screen, which makes it possible to count, from said eye box corresponding to the distance of use of the screen, the number of color bands visible simultaneously.

[0042] Advantageously and according to the invention, said lenticular network of said autostereoscopic screen forms a plurality of zones in front of the screen, called lobes, in which all the different viewpoints of the image follow one another at a preferred distance from the screen, called flat tint distance, corresponding to the distance at which each eye of the observer perceives only one viewpoint among all the viewpoints of the image over the entire surface of the screen and said network is perfectly focused in the central lobe so as to avoid any overlapping of viewpoints only in the central lobe.

[0043] According to this advantageous variant, the lenticular network is perfectly focused in the central lobe. The notion of a network perfectly focused in the central lobe means that there is no overlap of the viewpoints in the central lobe.

[0044] Indeed, it is known that an autostereoscopic screen is characterized in particular by its separating power, that is to say by its capacity to limit or eliminate the overlap between points of view, generating ghosts and double images. The LCD screens which are used to design autostereoscopic screens are generally not designed by the original manufacturers with a view to subsequently accommodating a lenticular network on their surface. The structure of their panel, consisting of pixels and sub-pixels of colors, can have very varied shapes. Each detail of a pixel, its geometry, its topology, that of its sub-pixels as well as the black mask which surrounds them is magnified when it crosses the optical axis of the lenses forming the micro-lenticular network placed on its surface. Contagions can occur and generate mixtures between neighboring points of view or luminous inhomogeneities.The final optical quality of a screen therefore depends on its resolving power.

[0045] Furthermore, optical components do not have homogeneous optical properties under all angles of incidence of the observer's gaze, due to variations in their front and side focusing. Optically speaking, the lobes follow one another but do not exactly resemble each other. In general, designers of auto stereoscopic screens favor the two secondary lobes adjacent to the central lobe, so as to optimize two lobes rather than just one. The focal length of the central lobe (No. 0) is a little shorter, those of the adjoining lobes n°-l and n°+l are perfect, while those of the lobes n°-2 and n°+2 are a little longer. The overlap between points of view which results from these local variations in focus does not allow the ideal focal length to be obtained from all angles.

[0046] In the case of the present invention, it was decided to optimize the central lobe only and to ensure that its focal length is perfect, to the detriment of the other lobes.

[0047] The system according to the invention introduces a constraint on the placement of the observer's eyes in order to be able to optimize the separating power of the screen and offer an experience of unequaled relief and depth.

[0048] Used in flat tint and viewed from the front, the resolving power of a lenticular network perfectly focused in the central lobe is optimum and allows exceptional depths of field to be obtained without ghosting.

[0049] The virtual image of the autostereoscopic screen of a system according to the invention, more or less enlarged, is visible in the central lobe but not from the side in the secondary lobes.

[0050] The depth of field made possible under these conditions makes it possible to overturn the rules of composition of relief scenes which are usually imposed in autostereoscopy.

[0051] The virtual image perceived through the "porthole" formed by the concave mirror can even reproduce the proportions of the surrounding real physical space. The orthoscopy between the real and the virtual is of quality.

[0052] Under these conditions and in an augmented reality configuration with a semi-transparent plane mirror, it becomes possible to project virtual objects into physical space on the scale of the real world, at depths unimaginable until now.

[0053] The relief performance of the system is very favorable: a small disparity results in the perception of a very large volume.

[0054] While the same relief scene with little disparity in the most opposite planes of the image gives a modest sensation of volume on a small screen seen from close up, it delivers a considerable depth of field and perfectly superimposable on the real world, without effort or fatigue, when the virtual image of this same screen is enlarged and moved back thanks to its reflection in the concave mirror.

[0055] Whether the concave mirror is viewed directly or by reflection in a front-facing plane mirror, a semi-transparent mirror or even the windshield of a vehicle, the size of the window through and beyond which the viewer can observe the scene is limited by the physical size of the concave mirror. Consequently, it is not advisable to use mirrors that are too small. A 30 cm wide mirror placed at about 1 m seems to be a minimum size for perceiving a scene in relief by binocular vision, whether in augmented reality in combination with a semi-transparent mirror or in direct vision.

[0056] Preferably, the lenticular network is further configured so that the successive viewpoints of the projected central lobe cover the width L of the eye box. Thus, no passage of lobes is observable in the eye box.

[0057] Different configurations are possible to benefit from the advantages of a system according to the invention.

[0058] Thus and according to a first embodiment, said converging optical component and said autostereoscopic screen extend horizontally and are arranged on the same vertical axis, said optical device further comprising a semi-transparent mirror having a reflecting surface oriented towards said eye box and towards said converging optical component, said semi-transparent mirror being inclined at an angle of 45° relative to the optical axis of said converging optical component which extends vertically, and arranged between said screen and said converging optical component.

[0059] Thus and according to this embodiment, the screen and the converging optical component (concave mirror or holographic optics) face each other and are each arranged on the same axis which corresponds to the optical axis of the converging optical component. A second semi-transparent mirror is arranged at 45° between the converging optical component and the autostereoscopic screen, semi-reflecting front face oriented towards the converging optical component and turned towards the eye box.

[0060] Thus, the light emitted by the screen passes through the semi-transparent mirror on the non-reflecting side. The light is then reflected by the concave mirror and this reflection is in turn reflected at 45° by the reflective front face of the semi-transparent mirror.

[0061] It should be noted that a portion of the light is lost at each passage through the semi-transparent mirror. Therefore, it is preferable to have a very bright screen to compensate for these successive line light losses, which are equivalent to 75% of the original light of the screen.

[0062] Viewed through the semi-transparent mirror, the virtual images are superimposed on the real world and can produce an illusion of augmented reality.

[0063] If the system is placed in front of a black background, the reflectivity of the mirror is much better, the relief image is much brighter and more contrasted, but augmented reality is no longer possible.

[0064] This configuration has a drawback which results from the “on-axis” architecture of the system and which results in the fact that the reflection of the observer’s face can interfere with the relief image. When the observer is himself illuminated, the light which he reflects follows the same optical path in the opposite direction. It is reflected on the reflective surface of the semi-transparent mirror at 45° facing him, is partly returned towards the concave mirror, then combined with the reflection of the screen and it returns towards the eyes of the spectator. When he stands at a distance less than the focal length of the concave mirror, he then perceives the greatly enlarged parasitic reflection of his face in the background of the scene. If he stands exactly at the focal length of the mirror, the reflection of his face is sent back to infinity and is no longer perceptible. On the other hand, if he moves back, the inverted reflection of his face in reduced size again parasitizes the scene in relief.

[0065] For certain use cases and in particular when the ambient light is controlled, this system nevertheless presents a very practical interest because it allows easy adjustment of all the parameters of the relief scene so that they coincide with the real environment. In addition, since the screen and the concave mirror are centered on the axis, this architecture minimizes the spherical deformations inherent in the use of a concave mirror used off-axis.

[0066] According to a second embodiment, said converging optical component and said autostereoscopic screen extend vertically, facing one another, vertically offset from one another, and arranged relative to the eye box such that the eye box is located behind and above the autostereoscopic screen, said converging optical component whose optical axis extends horizontally being arranged facing the eye box.

[0067] This second configuration makes it possible to eliminate the parasitic reflections of the observer observable with the first configuration. In addition, this second configuration preserves the initial brightness of the screen but does not allow use in augmented reality.

[0068] The screen and the concave mirror are arranged horizontally in a line, one behind the other. The screen is turned with its back to the viewer and the concave mirror faces it.

[0069] The user positions himself behind the screen and looks at the concave mirror located behind, a little higher than the screen. The "eye box" is located above the screen.

[0070] The system operates "off-axis". The concave mirror is oriented so that its central horizontal axis cuts the space into two superimposed volumes. The screen, turned towards the mirror, occupies the lower volume. It is inclined so that its reflection in the concave mirror is upright and allows the user to see a perfectly vertical enlarged virtual image.

[0071] It is possible to further offset the system to enlarge the "eye box". The screen can be lowered a little more and reoriented relative to the concave mirror. This results in greater distortions of the imaged scene. When using a concave mirror with a longer focal length, these distortions remain acceptable.

[0072] The viewer looks at the enlarged reflection of the screen through the "window" of the concave mirror and ignores the physical screen.

[0073] Thanks to this "off-axis" architecture, parasitic reflections are eliminated. Even when the user is illuminated, the light that he reflects takes the reverse optical path and ends up on the screen, which reflects almost nothing in return.

[0074] According to a third embodiment, said optical device further comprises a semi-transparent mirror arranged on the optical axis of said converging optical component, having a reflecting surface oriented towards said eye box and towards the converging optical component, said semi-transparent mirror being inclined at an angle of 45° relative to the optical axis of said converging optical component which extends vertically, said screen being arranged relative to the eye box in such a way that the eye box is located at the rear of said screen which faces said semi-transparent mirror, and vertically offset relative to the reflection of the converging optical component in said semi-transparent mirror.

[0075] This third configuration makes it possible to eliminate parasitic reflections, while offering the possibility of projecting relief content in augmented reality.

[0076] Thus and as in the second configuration, the screen is arranged with its back to the eye box and slightly inclined upwards relative to the vertical.

[0077] Instead of facing it, the concave mirror is here placed flat in front of the screen.

[0078] A semi-transparent plane mirror inclined at 45° is arranged above the concave mirror, with its reflective face oriented towards this mirror and towards the screen. It forms by reflection a virtual image of the concave mirror at the very place where it is physically located in the second configuration.

[0079] The screen is reflected in the virtual image of the concave mirror under the same conditions as previously, but this time in augmented reality.

[0080] The viewer sees the environment through the semi-transparent mirror inclined at 45°, on which the objects on a black background displayed in relief on the screen are superimposed.

[0081] As in the first configuration, this is done to the detriment of the original brightness of the screen, divided here by a factor of 4. Not only is the reflection of the concave mirror in the semi-transparent mirror at 45° attenuated by a factor of 2 but the light emitted by the screen also passes through this same mirror, which absorbs half of it.

[0082] It is therefore preferable to use a screen 4 times brighter to achieve a light intensity comparable to that of a screen seen live.

[0083] According to a fourth embodiment, said converging optical component and said autostereoscopic screen extend opposite each other, horizontally offset from each other, and inclined relative to the horizontal, said optical device further comprising a semi-transparent mirror having a reflecting surface oriented towards said eye box and towards said converging optical component, said semi-transparent mirror being inclined at an angle of 45° relative to the vertical and arranged in so that the light rays from the screen can reach the converging optical component without passing through the semi-transparent mirror.

[0084] This fourth configuration allows use in augmented reality while limiting the reduction in brightness.

[0085] This is an “off-axis” variant of the first embodiment. Thus, the semi-transparent plane mirror remains essential for generating an augmented reality experience but it is no longer aligned with the screen, so as to prevent the light from the screen from passing through the plane mirror before being reflected in the concave mirror.

[0086] The offset of the screen relative to the plane mirror is accomplished after having carried out a slight forward inclination of the concave mirror / plane mirror pair. The screen is shifted forward, until it no longer overhangs the plane mirror. It is then inclined backwards so as to remain facing the concave mirror and remain parallel to it. The rear of the screen is found to be substantially aligned with the front of the concave mirror.

[0087] The orthogonal projection of the center of the screen onto the concave mirror is no longer at its center. We are therefore in the conditions of an “off-axis” projection. Just as in the second and third embodiments, this avoids contaminating the reflection of the screen with the parasitic reflection of the viewer’s face.

[0088] The light from the screen no longer passes through the plane mirror before being reflected in the concave mirror and is therefore no longer attenuated.

[0089] The enlarged and receding reflection of the screen in the concave mirror is returned to the semi-transparent plane mirror, forming an angle close to 45° with respect to the vertical. Facing the plane mirror, the user sees this enlarged reflection through the virtual window formed by the reflection of the concave mirror. Its brightness has indeed been reduced this time but the total loss is now only 50%, and not 75%.

[0090] The distance and perceived size of the projected screen relative to this virtual window depends on the distance between the screen and the concave mirror. For example, as a 13.3-inch screen is moved away from the concave mirror, its projection can be likened to a 32-inch screen seen at 2 m, a 55-inch screen seen at 3 m, a 65-inch screen seen at 4 m, or a 5 m wide screen seen at 12 m.

[0091] The "off-axis" reflection of the screen in the concave mirror, however, causes a geometric distortion of the content, similar to a slight bean-shaped distortion. It is difficult to identify and not very troublesome in most use cases. If necessary, it would be possible to make minor corrections to the images to compensate for this distortion.

[0092] Since the light from the screen completely escapes the plane mirror and no longer needs to pass through it, it is no longer essential to use a semi-transparent mirror to reflect the concave mirror. It is perfectly possible to use a 100% front-facing mirror if augmented reality is not the desired effect and to maintain full screen brightness. The space behind the semi-transparent mirror can also be blacked out to enhance contrasts and accentuate its reflectivity.

[0093] According to a variant of this fourth embodiment, the system further comprises a stereoscopic camera arranged behind the semi-transparent mirror facing said eye box so as to be able to film said observer, said semi-transparent mirror having a black rear face so that the camera is not visible to the observer.

[0094] According to this variant, a stereoscopic camera is placed behind the semi-transparent plane mirror, facing the user. The rear face of the mirror around the lens is completely blacked out, so that no light can pass through it. The user therefore does not see the camera filming him from the front.

[0095] The shooting axis is adjusted so as to correspond to the axis of the observer's gaze. It then becomes possible to film in stereoscopy without worrying about the "false token" gaze, because without his knowledge, the user is always looking at the camera from the front. In doing so, his interlocutor has the impression of being looked at in the eyes and vice versa.

[0096] A tracking camera can be placed at the lower limit of the plane mirror, facing the user. Pupil detection makes it possible to dynamically adjust the content displayed by the autostereoscopic screen.

[0097] According to a fifth embodiment and for an automobile, railway or aeronautical application, said optical device further comprises a windshield inclined relative to a vertical axis towards said eye box and a semi-transparent mirror arranged under the windshield and having a reflective surface oriented towards said windshield, and said converging optical component extends vertically under the windshield and said autostereoscopic screen extends horizontally facing the semi-transparent mirror.

[0098] This configuration makes it possible to form an augmented reality device for a vehicle which can superimpose information on the road.

[0099] The proposed configuration is a variant of the third embodiment to obtain an augmented reality effect without parasitic reflection.

[0100] The screen is fixed horizontally, for example under the dashboard, face downwards.

[0101] The concave mirror is placed vertically in line with the front edge of the screen, with the reflective face facing backwards.

[0102] The plane mirror is inclined at 45°, with the reflecting face facing upwards. It starts from the bottom of the concave mirror and goes up towards the rear of the opening in the dashboard.

[0103] The light rays emitted by the screen are reflected a first time by the 45° mirror towards the concave mirror, which in turn reflects them a second time in the concave mirror. 45°. Thanks to this angle, the rays return upwards and pass through the dashboard through an opening made there. They then meet the windshield, which finally reflects them towards the driver's eyes. The latter then sees an enlarged three-dimensional image in front of the windshield. The driver's driving position defines the position of the eye box, from which the other parameters are determined. List of figures

[0104] Other aims, characteristics and advantages of the invention will appear on reading the following description given solely for non-limiting purposes and which refers to the appended figures in which: • [Fig.l] is a schematic view of a projection system according to a first embodiment of the invention, • [Fig.2] is a schematic view of a projection system according to a second embodiment of the invention, • [Fig. 3] is a schematic view of a projection system according to a third embodiment of the invention, • [Fig.4] is a schematic view of a projection system according to a fourth embodiment of the invention, • [Fig.5] is a schematic view of a projection system according to a fifth embodiment of the invention, • [Fig.6] is a schematic view of an autostereoscopic screen of a projection system according to one embodiment of the invention.

[0105] Detailed description of an embodiment of the invention

[0106] In the figures, the scales and proportions are not strictly respected, for the purposes of illustration and clarity.

[0107] Identical, similar or analogous elements are designated by the same references in all the figures.

[0108] Whatever the embodiment, the autostereoscopic projection system according to the invention comprises an optical device comprising at least one concave mirror 30 having a radius of curvature R and a focal length F. This optical device is arranged relative to a predetermined three-dimensional space (eye box 50) which defines the space within which the observer must place his eyes to perceive the autostereoscopic image projected by the system according to the invention, so as to be able to direct the light rays that he receives towards a viewing window observable from the eye box. As indicated previously, according to another embodiment, the concave mirror can be replaced by a holographic optic which then forms the convergent optical component.

[0109] According to the embodiment of figures 1, 3, 4 and 5, the optical device further comprises a semi-transparent mirror 20.

[0110] The system according to the invention also comprises, whatever the embodiment, an auto stereoscopic screen 10 arranged relative to the optical device so that the light rays that it emits can reach the concave mirror 30, either directly or after reflection on the semi-transparent mirror 20.

[0111] The auto stereoscopic screen 10 is separated from the concave mirror 30 by a distance d less than F.

[0112] The screen 10 further comprises a matrix 10a of pixels arranged in rows and columns, each pixel being composed of a plurality of sub-pixels of different colors each assigned to a point of view of the image.

[0113] The screen also comprises a lenticular network 10b which surmounts the matrix and each lens of which is inclined at a non-zero angle relative to the direction of the columns of the matrix. This aspect of the invention is known and is not shown in the figures. It is also possible to have a vertical network.

[0114] The lenticular network of the screen 10 also has the particularity of being perfectly focused in the central lobe so as to avoid any overlapping of viewpoints only in the central lobe.

[0115] In particular, the autostereoscopic screen comprises a matrix of pixels arranged in rows and columns, each pixel being composed of a plurality of sub-pixels of different colors each assigned to a viewpoint of the autostereoscopic image to be displayed and a lenticular network which surmounts the matrix. Each lens is inclined, in the embodiment of the figures, by a non-zero angle relative to the direction of the columns of the matrix (for example 18°). As indicated previously, however, nothing prevents the use of a vertical network. As indicated previously, nothing prevents the use of pixel addressing of the viewpoints of the image. Throughout the following, it is considered that S is equal to 3 for sub-pixel addressing of pixels each formed of three sub-pixels of different colors.

[0116] [Fig.6] schematically illustrates an auto stereoscopic screen 10 comprising a matrix of pixels 10a arranged in rows and columns, each pixel being composed of a plurality of sub-pixels of different colors each assigned to a viewpoint of the auto stereoscopic image to be displayed. The screen also comprises a lenticular network 10b which surmounts the matrix 10a.

[0117] The pitch P of the screen network is determined from the pitch P', called the nominal pitch, of an autostereoscopic screen, called the nominal screen, having usage characteristics identical to those of the system screen and intended to be observed in direct vision at a distance corresponding to the usage distance of the system screen, by the formula P=P'.(1+C / (SM)) where C is the number of parts of visible viewpoints simultaneously from the distance of use inside the eye box, when said nominal screen of pitch P' is arranged in said projection system in place of said screen, S is the number of viewpoints addressed by each pixel of the pixel matrix and M the horizontal resolution of the matrix.

[0118] For example, if we consider a 13.3 inch (33.782 cm) screen with 10 viewpoints and a resolution of 3840 x 2160. This screen is placed at a distance of 700 mm from the concave mirror in rectangular format of 350 mm by 250 mm, whose focal length is 750 mm and the radius of curvature is 1500 mm. Once placed in the system, the observer views from the eye box 12 color bars (assuming that one color per viewpoint is displayed on the screen). The pitch of a lens corresponding to 10 colors in direct vision, 12 color bars in reflection in the concave mirror represent 12 / 10ths of the pitch of the lenses.

[0119] Since one image viewpoint is displayed per sub-pixel, the number of lenses in the network of the screen considered is equal to (3840 pixels x 3) / 10, i.e. 1152 lenses.

[0120] The total pitch correction is therefore equal to 12 / 10 = 1.2 lenses.

[0121] This corresponds to a correction per lens equal to 1.2 / 1152 = 0.00104.

[0122] In other words, the pitch of the lenses must therefore be increased by 0.1%.

[0123] If the known pitch P' of a lens is 0.240 mm, the necessary pitch correction amounts to adding 0.00024 mm per lens.

[0124] Although this value of 0.24 microns seems relatively small, its incidence is in reality considerable for optimal observation of the autostereoscopic image reflected by the concave mirror. This correction is decisive for being able to exploit an autostereoscopic screen in reflection in a concave mirror.

[0125] According to one embodiment, a screen with 14 viewpoints is used in steps of 2, which makes it possible to obtain a lobe of 45.5 cm, sufficient to occupy the entire width of the "eye box". The exceptional resolving power obtained in the central lobe makes it possible to very comfortably see relief scenes extending continuously from 1 m to 100 m deep. Within the limits of a lateral displacement of about twenty cm on either side of the center of the "eye box", the 14 viewpoints also offer a horizontal parallax in accordance with reality. When one moves away from the flat tint distance, they also allow the images to recompose themselves in accordance with physiological expectations. These properties are very useful for giving the user a certain latitude of positioning. However, the quality of the sensations is so exceptional at the flat tint that it is intuitive to place oneself there spontaneously.

[0126] According to one embodiment, the concave mirror has a radius of 1.5 m and a focal length of 0.75 m. The distance d between the screen 10 and the concave mirror 20 is fixed at 0.8 m.

[0127] The various components of the system according to the invention (concave mirror, autostereoscopic screen and semi-transparent mirror) can be housed in a housing not shown in the figures or arranged relative to each other in a dedicated observation space corresponding to the intended use. The observation area can depend on the intended use. Thus, in the embodiment of [Fig.5] intended for automotive use, the eye box 50 is determined by the position of the driver in the motor vehicle in which the system according to the invention is installed.

[0128] In [Fig. 1], the system comprises a concave mirror 30 and a screen autostereoscopic 10 extend horizontally and arranged on the same vertical axis represented by a dotted line. In addition, the system comprises a semi-transparent mirror 20 having a reflective surface oriented towards the eye box 50 and towards the concave mirror 30, inclined at an angle of 45° relative to the optical axis of the concave mirror and arranged between the screen 10 and the concave mirror 30.

[0129] As indicated previously, according to this embodiment, the light emitted by the screen 10 passes through the semi-transparent mirror 20 on the non-reflecting side. The light is then reflected by the concave mirror 20 and this reflection is in turn reflected at 45° by the reflective front face of the semi-transparent mirror 20. The viewer whose eyes are placed in the eye box 50 can thus perceive the image reflected by the semi-transparent mirror 20.

[0130] According to one embodiment, the screen is a 13.3-inch 4K screen and the concave mirror is a mirror having a focal length of 75 cm, open at F / 2 and sufficiently wide (37.5 cm for example).

[0131] In [Fig.2], the system comprises a concave mirror 30 and a screen autostereoscopic 10 which each extend vertically, facing one another. They are further vertically offset from one another, and arranged relative to the eye box 50 such that the eye box is located behind and above the autostereoscopic screen 10. The concave mirror 30 is arranged facing the eye box 50.

[0132] The concave mirror is oriented so that its median horizontal axis, represented in [Fig.2] by a dotted line, cuts the space into two superimposed volumes. The screen 10 turned towards the mirror 30, occupies the lower volume. It can be slightly inclined so that its reflection in the concave mirror 30 is straightened and allows the user to see a perfectly vertical enlarged virtual image.

[0133] In [Fig.3], the system comprises a semi-transparent mirror 20 arranged on the axis optics of the concave mirror 30 and which has a reflective surface oriented towards the eye box 50 and towards the concave mirror 30.

[0134] This semi-transparent mirror 20 is inclined at an angle of 45° relative to the optical axis of the concave mirror, represented in the figure by a vertical dotted line. The screen 10 is arranged relative to the eye box 50 in such a way that the eye box located at the rear of the screen which faces the semi-transparent mirror 20. Furthermore, the screen 10 is vertically offset relative to the reflection of the concave mirror 30 in the semi-transparent mirror 20.

[0135] As indicated previously, this embodiment makes it possible to eliminate parasitic reflections while offering the possibility of projecting relief content in augmented reality.

[0136] In [Fig.4], the system comprises a concave mirror 30 and an autostereoscopic screen 10 which extend opposite each other, horizontally offset from each other, and inclined relative to the horizontal.

[0137] The system further comprises a semi-transparent mirror 20 having a reflective surface oriented towards the eye box 50 and towards the concave mirror 30. The semi-transparent mirror 20 is inclined at an angle of 45° relative to the vertical and arranged so that the light rays from the screen can reach the concave mirror without passing through the semi-transparent mirror.

[0138] As indicated previously, this embodiment allows use in augmented reality while limiting the reduction in brightness.

[0139] It is also possible, as illustrated in [Fig.4], to have an opaque mask 60 behind the semi-transparent mirror to enhance the contrasts and accentuate the reflectivity. Of course, according to this variant, augmented reality is not possible.

[0140] The system of [Fig.5] further comprises a windshield 40 inclined relative to a vertical axis, for example at an angle of 37°, towards the eye box 50 and a semi-transparent mirror 20 arranged under the windshield 40 and having a reflective surface oriented towards the windshield 40. Furthermore, the concave mirror 30 extends vertically under the windshield 40 and the autostereoscopic screen 10 extends horizontally facing the semi-transparent mirror 20.

[0141] When used in a motor vehicle, the screen 10 is fixed horizontally under the dashboard of the vehicle, facing downwards.

[0142] The plane mirror is inclined at 45°, with the reflecting face facing upwards. It starts from the bottom of the concave mirror and goes up towards the rear of the opening in the dashboard.

[0143] This embodiment thus makes it possible to form an augmented reality device for a vehicle which can superimpose information on the road.

[0144] The light rays emitted by the screen 10 are reflected a first time by the 45° mirror towards the concave mirror 30, which in turn reflects them a second time in the 45° mirror. Thanks to this angle return, the rays go back upwards and pass through the dashboard through the opening which has been arranged there. They then meet the windshield 40, which finally reflects them towards the driver's eyes arranged in the eye box 50. The driver then sees an enlarged three-dimensional image in front of the windshield 40

Claims

1. Claims Dematerialized image projection system stereoscopic auto intended for an observer whose eyes are positioned within a predetermined three-dimensional space, called an eye box (50), having a predetermined width L, at least greater than the average inter-pupillary distance of an observer and preferably greater than the average width of a human being's head, said system being characterized in that it comprises: an optical device comprising at least one converging optical component (30) with behavior identical to a concave mirror having a radius of curvature R and a focal length F, said optical device being arranged relative to said eye box (50) so as to be able to direct the light rays that it receives towards a viewing window observable from said eye box, an autostereoscopic screen (10) arranged relative to said optical device so that the light rays that it emits can reach said viewing window, after reflection on said converging optical component (30), and having traveled a distance d less than F, said autostereoscopic screen (10) further comprising: • a matrix of pixels of N rows and M columns, each pixel being composed of a plurality of sub-pixels of different colors, • a lenticular network surmounting said pixel matrix, and having a pitch P determined from a pitch P', called nominal pitch, of an autostereoscopic screen, called nominal screen, having usage characteristics identical to those of the system screen and intended to be observed in direct vision at a distance corresponding to the usage distance of said system screen, by the formula P=P'.(1+C / (SM )) where C is the number of parts of viewpoints visible simultaneously from said usage distance inside said eye box, when said nominal screen of pitch P' is arranged in said projection system instead of said screen, and S is the number of different viewpoints addressed by each pixel of said matrix.

2. System according to claim 1, characterized in that said lenticular network of said autostereoscopic screen (10) forms a plurality of zones in front of the screen, called lobes, in which all the different viewpoints of the image follow one another at a preferred distance from the screen, called flat tint distance, corresponding to the distance at which each eye of the observer perceives only one viewpoint among all the viewpoints of the image over the entire surface of the screen and in that said network is perfectly focused in the central lobe so as to avoid any overlapping of viewpoints only in the central lobe.

3. System according to one of claims 1 or 2, characterized in that said converging optical component (30) and said auto stereoscopic screen (10) extend horizontally and are arranged on the same vertical axis, said optical device further comprising a semi-transparent mirror (20) having a reflective surface oriented towards said eye box and towards said converging optical component (30), said semi-transparent mirror (20) being inclined at an angle of 45° relative to the optical axis of said converging optical component which extends vertically, and arranged between said screen and said converging optical component.

4. System according to one of claims 1 or 2, characterized in that said converging optical component (30) and said autostereoscopic screen (10) extend vertically, facing each other, vertically offset from each other, and arranged relative to the eye box in such a way that the eye box is located behind and above the autostereoscopic screen, said converging optical component whose optical axis extends horizontally being arranged facing the eye box.

5. System according to one of claims 1 or 2, characterized in that said optical device further comprises a semi-transparent mirror (20) arranged on the optical axis of said converging optical component (30), having a reflecting surface oriented towards said eye box and towards the converging optical component (30), said semi-transparent mirror (20) being inclined at an angle of 45° relative to the optical axis of said converging optical component (30) which extends vertically, said screen (10) being arranged relative to the eye box such that the eye box is located at the rear of said screen (10) which faces said semi-transparent mirror, and vertically offset relative to the reflection of the converging optical component in said semi-transparent mirror.

6. System according to one of claims 1 or 2, characterized in that said converging optical component (30) and said auto stereoscopic screen (10) extend opposite each other, horizontally offset from each other, and inclined with respect to the horizontal, said optical device further comprising a semi-transparent mirror (20) having a reflecting surface oriented towards said eye box and towards said converging optical component, said semi-transparent mirror being inclined at an angle of 45° with respect to the vertical and arranged so that the light rays from the screen can reach the converging optical component without passing through the semi-transparent mirror.

7. System according to claim 6, characterized in that it further comprises a stereoscopic camera arranged behind the semi-transparent mirror (20) facing said eye box so as to be able to film said observer, said semi-transparent mirror having a black rear face so that the camera is not visible to the observer.

8. System according to one of claims 1 or 2, characterized in that said optical device further comprises a windshield (40) inclined relative to a vertical axis towards said eye box and a semi-transparent mirror (20) arranged under the windshield and having a reflective surface oriented towards said windshield, and in that said converging optical component (30) extends vertically under the windshield and said autostereoscopic screen (10) extends horizontally facing the semi-transparent mirror.

9. System according to one of the preceding claims, characterized in that said autostereoscopic screen (10) is a 5.5 inch (13.97 cm) or 13.3 inch (33.782 cm) screen.

10. System according to one of claims 1 to 9, characterized in that said convergent optical component (30) is a concave mirror or holographic optics.

Images