EP4457559A1 - Device for projecting an image into the eye of a user - Google Patents

Abstract

A device (1) for projecting an image onto the retina of an eye, the device comprising light sources and an emission support (10) having different emission points, each emission point being configured to emit light in an angular direction of propagation, the emission points being distributed in different subsets of emission points, each emission point subset emitting light in the same angular direction, so as to converge towards the same pixel of the image. The device is such that a same light source is connected to several emission subsets, called twin emission subsets.

Description

Device for projecting an image into the user's eye

Description

TECHNICAL AREA

The technical field of the invention relates to the projection of an image on an eye, in applications of the augmented reality type.

PRIOR ART

Wearable augmented reality devices, such as goggles, make it possible to observe a real scene while viewing additional information. This type of device is frequently based on micro-screens, making it possible to form an image in the immediate vicinity of a user's eye. Such micro-screens can for example be integrated into a bezel. An optical system, comprising a set of lenses, allows the perception of a sharp image by the eye.

Patent US9632317 describes a device allowing projection, on the retina of an eye, without screen or optical system. The device comprises a transparent integrated optical circuit composed of an array of nanometric light guides, an array of electrodes and a holographic film. Such a device is compact, and makes it possible to obtain an extended field of vision. In addition, this makes it possible not to use a bulky optical system of complex design.

The nanometric light guides make it possible to define a set of emission points on the holographic film, each point being able to be illuminated by a light extracted from a light guide. The set of emission points is subdivided into different emission subsets, each emission subset comprising emission points distributed, as randomly as possible, on the holographic film. The emission points of the same emission subset can be simultaneously illuminated by the different light guides. Under the effect of the illumination, each emission point of the same emission subset emits a light wave propagating in the same angular direction as far as the pupil of the eye, so as to form a luminous point. unique to the retina. In this way, each subset of emission points allows formation of a pixel of the image perceived by the user. An image can be formed by successively illuminating different subsets of emission points, so as to form an image comprising a large number of pixels. The illumination frequency of each subset of emission points is dimensioned so that under the effect of retinal persistence, the user can feel the formation of a fixed image, despite the sequential formation of the various pixels of the image.

As previously indicated, it is preferable that in the same subset, the emission points are distributed as randomly as possible, so as to avoid effects of repetition of patterns by periodic or quasi-periodic structures during the formation of the image on the retina.

In US9632317, light is extracted from each light guide by arranging electrically modulated diffraction gratings along the light guide. The diffraction gratings are spaced from each other and define point light extraction zones. Under the effect of electrical activation, each diffraction grating allows extraction of the light propagating in a light guide. In US9632317, the light guides are sinusoidal in shape, and extend parallel to each other. It is however necessary to have a large number of laser light sources, typically several hundred, if it is desired to have a sufficient number of pixels formed simultaneously on the image.

The inventors propose an optimized configuration, so as to maintain a high number of pixels, while reducing the number of laser light sources and preserving the quality of the images formed on the retina.

DISCLOSURE OF THE INVENTION

A first object of the invention is a device for projecting an image onto the retina of an eye, the image being composed of several pixels, the device comprising:

- light sources, preferably laser;

- a transmission medium, comprising:

• light guides, each light guide being connected to a light source and to a plurality of diffraction gratings, distributed along the light guide, each diffraction grating being electrically adjustable to extract part of the propagating light in the light guide to which it is connected;

• electrodes, each electrode being configured to modulate diffraction gratings connected to different light guides;

• emission points, each emission point being formed at the level of a diffraction grating, and extending between an electrode and a light guide, each point emission being configured to emit light, in an angular direction, following an extraction of a light propagating in the light guide; the device being such that: different emission points are configured to emit the light according to the same angular direction, so as to converge on the retina, by forming the same pixel, the said emission points defining an emission subset , so that each emission sub-assembly corresponds to an angular direction of emission of the light; two adjacent pixels of the image are formed by two emission subsets, respectively associated with two different angular directions and spaced apart by an elementary angular difference; the device being characterized in that:

- the same light source is connected to different light guides, said light guides defining different emission sub-assemblies, called twins, configured to simultaneously form different pixels, called twin pixels, when the light source is activated, the twin emission sub-assemblies being associated with the light source;

- the respective angular directions of each twin emission sub-assembly are spaced from each other by an angular difference greater than or equal to a minimum angular difference, the minimum angular difference corresponding to k times the elementary angular difference, k being greater than or equal to 2;

- in such a way that the twin pixels, simultaneously formed on the retina, are shifted by at least the minimum angular deviation.

The emission points of different twin emission sub-assemblies are preferably connected to a same electrode or to an electrode of a same group of electrodes.

According to one embodiment:

- several light sources are respectively connected to different light guides, to form different twin emission sub-assemblies, each twin emission sub-assembly being associated with the same light source;

- the respective angular directions of each twin emission sub-assembly are spaced from each other by an angular deviation greater than the minimum angular deviation.

Preferably, the waveguides, connected to the same light source, and forming a twin emission subset are different from other waveguides connected to said light source, and forming another subset twin emission associated with the same light source. The minimum angular deviation can be greater than or equal to 0.5° or 1°.

According to one embodiment:

- the diameter or the diagonal of the emission points is less than or equal to 10 pm, the minimum angular deviation being greater than 1.5;

- the diameter or the diagonal of the emission points is between 10 μm and 20 μm, the minimum angular difference being greater than 0.8;

- the diameter or the diagonal of the emission points is greater than 20 pm, the minimum angular deviation being greater than 0.5;

Preferably, the diameter or diagonal of each emission point is greater than 5 µm.

According to one possibility:

- the light guides connected to the same light source define first twin emission sub-assemblies and second twin emission sub-assemblies, so as to respectively form first twin pixels and second twin pixels;

- the emission points of each first twin emission subset are connected to first electrodes;

- The emission points of each second twin emission subassembly are connected to second electrodes, different from the first electrodes;

- the device comprises a switch, to supply either the first electrodes, or the second electrodes, so that the light emitted by the light source is extracted sequentially either by the emission points of the first emission subassemblies, or by the emission points of the second emission subsets, to form sequentially, on the image, the first twin pixels and the second twin pixels.

Advantageously, for at least one light source, different modulators are respectively interposed between a light source and different light guides connected to the light source, so as to modulate an intensity of the light propagating in said light guides independently for different light guides connected to the light source.

According to one possibility:

- A light source is connected to N different twin emission sub-assemblies, configured to simultaneously form N twin pixels, N being an integer greater than or equal to 2; - the light source is connected to N modulators, each modulator extending between the light source and the waveguides forming the same emission sub-assembly.

The number of twin emission subsets formed by at least one light source is greater than or equal to 20.

The device may include fewer than 20 different light sources to form an image, or part of an image, of at least 200 pixels by 200 pixels.

According to one embodiment, the device comprises a holographic film, subdivided into different holograms, each hologram being associated with a diffraction grating, and configured to emit the light wave, according to the direction of emission, under the effect of light extracted by the diffraction grating with which it is associated, each association between a hologram and a diffraction grating forming a point of light emission.

A second object of the invention is a bezel, comprising bezel lenses, the bezel comprising a device according to any one of the preceding claims, the stack being formed on at least one bezel lens.

A third object of the invention is a method for parameterizing a device according to the first object of the invention, the parameterization comprising: a) modeling of a spatial distribution of the noise between two twin pixels, by modifying the angular difference minimal; b) estimation of the minimum angular deviation as a function of the spatial distribution.

Step b) can comprise an estimation of the signal-to-noise ratio as a function of the minimum angular deviation.

Another object of the invention is a device according to the first object of the invention, the minimum angular deviation of which is parameterized by the method according to the second object of the invention.

The invention will be better understood on reading the description of the embodiments presented, in the remainder of the description, in connection with the figures listed below.

FIGURES

Figures IA and IB schematize the principle of self-focusing of emission points on a user's retina. In FIG. 1A, a set of emission points has been represented, converging towards a pixel. In FIG. IB, another set of emission points has been represented, converging towards an adjacent pixel of the pixel described in connection with FIG. IA.

FIG. 1 C shows an example of a subset of emission points distributed over an emission medium.

FIG. 1D illustrates a step for recording holograms on a holographic film. FIG. 1E illustrates a step of emission of a light wave by emission points formed on the device.

FIG. 2 corresponds to an example of implantation of light guides and electrodes.

FIG. 3A represents a model of a light wave, produced by a subset of emission points, and reaching a pixel. Figure 3A shows the intensity of the light wave incident on the retina as a function of position along a direction passing through the center of the pixel.

FIG. 3B corresponds to FIG. 3A, the unit of the abscissa axis being angular.

Figure 3C illustrates a calculation of the signal-to-noise ratio (SNR) associated with a light wave produced by an emission subassembly and reaching a pixel.

FIG. 4 represents pixels of an image formed on the retina of an eye, as well as a circular perimeter around pixels, the diameter of which determines a minimum angular difference between two pixels, called twins, formed from sub- emission sets simultaneously illuminated by the same light source.

FIG. 5A shows an evolution of the signal-to-noise ratio of a pixel, called a central pixel, surrounded by 4 neighboring pixels, for different configurations, as a function of the angular difference between the central pixel and each neighboring pixel.

FIG. 5B shows images of a central pixel, surrounded by 4 neighboring pixels, for three different angular deviations between the central pixel and each neighboring pixel. In FIG. 5B, the pixels result from the simultaneous illumination of emission subsets by the same light source: these are twin pixels.

FIG. 5C shows images of a central pixel, surrounded by 4 neighboring pixels, for three different angular deviations between the central pixel and each neighboring pixel. In FIG. 5C, the pixels result from the illumination of emission subsets by light sources different from each other.

FIG. 6A schematizes a configuration of light sources, light guides and electrodes making it possible to define 16 emission subsets, each emission subset comprising 4 emission points.

Figure 6B illustrates 16 pixels that can be formed by the configuration shown in Figure 6A.

FIG. 7A schematizes a configuration of light sources, light guides and electrodes making it possible to define 16 emission subsets, each subset emission comprising 4 emission points, while reducing the number of light sources compared to the configuration of FIG. 6A.

Figure 7B illustrates 16 pixels that can be formed by the configuration shown in Figure 7A.

FIG. 8A schematizes the configuration described in connection with FIG. 7A, with simultaneous activation of the two light sources and activation of half of the electrodes. Figure 8B shows the pixels displayed by the configuration of Figure 8A.

FIG. 9A schematizes the configuration described in connection with FIG. 7A, with simultaneous activation of the two light sources and activation of the other half of the electrodes. Figure 9B shows the pixels displayed by the configuration of Figure 9A.

FIG. 10 represents an evolution of the signal-to-noise ratio of the image as a function of the angular spacing between a central pixel and 4 neighboring pixels, all of the pixels being twin pixels. In FIG. 10, the size of each emission point has also been varied.

FIG. 11 shows an example of an arrangement of twin pixels, formed by the same light source.

Figure 12 schematizes a compromise between the signal-to-noise ratio of the image (left ordinate axis) and the number of light sources used (right ordinate axis) as a function of the angular difference between the twin pixels ( horizontal axis).

DESCRIPTION OF PARTICULAR EMBODIMENTS

Figures IA to 1E illustrate the principles on which the invention is based. These principles are described in the publication Martinez C, “See-through holography retinal projection display concept”, Optica, Vol. 5, No. 10, October 2018. A projection device 1 is placed facing a pupil P of an eye, the latter focusing the light on the retina R of the eye. The projection device comprises an emission medium 10. The emission medium 10 is formed of a set of emission points EP, each emission point being able to be illuminated by a light source. The structure of transmission medium 10 is described in connection with FIGS. 1C and 1D. The emission medium 10 is placed at a distance Zo from the pupil P. The pupil P is placed at the focal distance f _Q of the retina R.

Each emission point EP represented in FIG. IA is configured to emit light in an angular direction of emission 0, _y towards the pupil P. In general, the emission points EP of the emission surface are divided into different EPDj emission subsets, (Emission Point Distribution) such that the emission points of the same emission subset all emit light each other. propagating parallel to the same angular direction of emission 0^. The angular direction of emission is defined with respect to a reference direction, for example a direction passing through the center of the pupil and perpendicular to the latter. The reference direction is represented by a dashed line in Figures IA and IB.

Each emission point EP of a same emission subassembly EPDjj is configured to emit, under the effect of illumination by a light source, a light wave propagating in an angular direction of propagation towards the pupil P. The phase relationship between the different light waves, respectively emitted by the emission points EP of a same emission subset EPDj , makes it possible to form, upstream of the pupil P, a plane wave , of wave vector kjj or, more generally, a wave whose wavefront is controlled. Thus, the wave front can have a flat shape, or else spherical or parabolic. Downstream of the pupil, each light wave is focused by the pupil and converges to the same pixel R(i,j) on the retina R.

The projection device 1 thus makes it possible to form an image, on the retina R, the image being discretized into several pixels R(i,j), each pixel R(i,j) corresponding to the pixel of an image formed, by the projection device 1, on the retina R. In the example shown, the image has a square shape, and the same applies to each pixel composing it. An image can have several hundred pixels per side.

The light intensity at the level of the pixel R(i,j) results from the contribution of each light wave, propagating, upstream of the pupil, in the same angular direction 0, _y resulting from each point of emission EP d' the same emission subset EPDjj, after focusing by the lens. The index ij corresponds to an angular coordinate of the light wave which, after focusing by the pupil (or self-focusing), converges towards the pixel R(i,j) on the retina.

The transmission medium is segmented into different emission subsets EPDjj, each subset comprising several emission points EP distributed over the emission surface, each emission subset being associated with the same direction of propagation angular Qjj, given by the wave vector kjj, towards the pupil.

FIG. 1B represents another EPD _i+1 subset. The emission subset EPDj+1 comprises different emission points EP, each point being associated with a same direction of angular propagation 0 _i+1J -, given by a same wave vector k _i+1 . Downstream of the pupil, each light wave, emitted by each point of the EPD _i+1 emission subset, converges to the same pixel R(i + l,j) on the retina R. The pixel R(i + l,j) is adjacent to pixel R(i,j) shown in Figure IA. The angular directions 0, _y , Qt+ which respectively make it possible to form two adjacent pixels on the retina, are spaced apart by an elementary angular deviation /?. The elementary angular deviation corresponds to an angle, upstream of the pupil, between two angular directions of propagation of light waves converging, downstream of the pupil, towards two adjacent pixels R(i + 1, j), R(i , j) on the retina R. The elementary angular deviation /3 is typically less than 0.1° or 0.2°. It is for example equal to 0.06°.

As described in connection with the prior art, it is advantageous for the emission points EP of the same emission subset EPDij to be distributed in a pseudo-random manner on the emission medium 10. FIG. a subset of emission points configured to emit light propagating towards the pupil P, parallel to the same angular direction θ^. The emission medium extends parallel to the pupil, along an X axis and a Y axis. Each emission point EP has coordinates (x _EP , y _E p) along these axes.

FIG. 1D diagrams the structure of the projection device 1. The projection device comprises the emission medium 10, the latter being connected to several light sources 21 and several switches 23. Preferably, each light source is a source of laser light. Each switch makes it possible to connect an electrode 13 to a power supply. The emission medium is formed of several transparent layers superimposed on each other: a first layer, in which light guides 11 are formed. Each light guide is configured to receive coherent light, emitted by a laser source 21, and to propagate the coherent light along the emission medium. Each light guide can for example be formed from SiN (silicon nitride) deposited on glass. Each guide can extend along a thickness and a width comprised between 200 nm and 300 nm. a second layer, in which are formed diffraction gratings 12, such that each diffraction grating 12 is coupled with a light guide 11. Each diffraction grating is configured to extract part of the light propagating through the light guide 11. Each diffraction grating 12 corresponds to a periodic variation of refractive index, capable of being electrically modulated. Each diffraction grating 12 can be formed from inclusions, defining a periodic pattern, for example a Bragg grating, in silicon oxide (SiOz). Each inclusion is formed of a material whose refractive index is electrically adjustable, for example a liquid crystal. The diffraction gratings 12 coupled to the same light guide 11 are spaced apart along the light guide, and are considered to be point-like. When the wavelength of the light is 532 nm, the period of the pattern of the diffraction grating 12 can be between 300 nm and 400 nm. A diffraction grating can extend according to 10 periodic patterns, and thus extend over a length of a few microns, for example between 2 and 10 μm. a third layer, in which transparent electrodes 13 are formed, the electrodes being configured to electrically modulate the refractive index of a material forming the diffraction gratings. The transparent electrodes can be formed from a transparent conductive material, for example ITO (indium tin oxide). Each electrode can thus activate a diffraction grating under the effect of the electrical modulation. a fourth layer, corresponding to a holographic film 14. By holographic film is meant a photosensitive medium capable of recording a hologram. The holographic film is assumed to be thin enough to be assimilated to an emission surface. The holographic film can be a photopolymer, for example poly(methyl methacrylate), or a photoresist.

The transmission medium 10 is preferably formed on a plate. It may be a transparent plate 15 when the device is intended to be integrated into a bezel. It may for example be a plate of glass or polycarbonate.

Under the effect of a polarization by an electrode 13, each point diffraction grating 12 connected to the electrode can be activated, in the sense that it allows an extraction of part of the light propagating in a light guide 11 to which it is coupled. The extracted light propagates towards an elementary zone of the holographic film 14, the elementary zone storing a hologram. Under the effect of the illumination, the hologram forms an emission point EP, emitting a light wave according to a predetermined wave vector k and phase O. Thus, each diffraction grating 12, controlled by an electrode 13, allows coupling of a waveguide 11 with a hologram previously stored in the holographic film. The phase of the light wave emitted by the emission point depends on the phase information stored in the hologram. The light waves emitted by different emission points EP of the same emission subset EPDj are out of phase with respect to each other so that all of these light waves form a coherent wave, with the same vector of wave k^, for example plane, propagating towards the pupil P in the same angular direction 0 _f .

Thus, each emission point EP corresponds to a superposition, parallel to the emission surface, between a point diffraction grating 12 coupled to a light guide 11, and an electrode 13, facing a hologram of the holographic film 14.

The emission points are distributed along an emission surface S. The light guides 11 are arranged parallel to the emission surface S. The same applies to the electrodes 13. Thus, the electrodes 13 are superimposed on the light guides. light 11. Parallel to the emission surface S, each electrode “crosses” several light guides, so as to define several intersections, each intersection corresponding to a position of an emission point EP. The term "cross" is to be interpreted as designating a superposition of an electrode and a light guide. In FIG. 1D, for simplification purposes, a light guide 11 has been shown coupled to three diffraction gratings 12, the latter being connected to three different electrodes 13, the latter being oriented perpendicular to the light guide 11.

The holographic film 14 has previously been the subject of a holographic recording, shown schematically in FIG. 1D. In known manner, a hologram is formed by interference between two light waves emitted by a coherent light source: an object light wave and a reference light wave. The interference fringes generated are stored physically or chemically in the holographic film 14. During the recording phase, the light extracted from a light guide 11 acts as a reference beam. In FIG. 1D, there is represented: by dashed arrows, light emitted by the laser light source and propagating in a light guide 11; by solid dashed arrows, the light extracted from each light guide 11, respectively by three diffraction gratings 12, under the effect of activation by the three electrodes 13. The object beam is a beam propagating towards the holographic film according to a wave vector k^, so as to converge at a point of the retina by autofocusing. During the recording step, the object beam can be emitted by a collimated source.

In FIG. 1D, three object beams are represented propagating according to the same wave vector k _j materialized by dashed arrows. The object beams and each reference beam are emitted by the same laser light source. The recording phase consists in storing holograms, in different elementary zones of the holographic film 14, each recording resulting from an interference between the object beam, propagating according to the same wave vector k _j (and therefore the same angular direction 0^) and a reference beam extracted from a light guide 11. During recording: different elementary zones of the holographic film 14 are exposed to an object beam of the same wave vector kÿ, so as to form a emission subset EPDjy of emission points EP corresponding to the same direction of propagation. During this exposure, the emission points of the emission subassembly EPD^ are illuminated by the reference beam, the latter being extracted from the light guides. The other elementary zones of the holographic film 14, located opposite points belonging to other emission subsets, are masked. different elementary zones of the holographic film 14 are respectively exposed to different object beams propagating according to different wave vectors, so as to form emission points associated with different emission directions e 0 ^ .

FIG. 1E illustrates the phase of use of the device 1. In this figure, there is shown, in shaded form, the elementary zones of the holographic film 14 having recorded a hologram during the holographic recording step. Under the effect of an activation by electrodes 13, part of the laser light propagating in the light guide 11 is extracted and propagates towards the holographic film 14, as described in connection with figure ID. Under the effect of the extracted illumination, each hologram stored in the photosensitive film diffracts a wave which corresponds to the object wave at the time of recording, and in particular along the direction of propagation k^. The diffracted wave is represented by mixed dashes in Figure 1E. When using the device: the emission points EP corresponding to the same emission subset EPDij, associated with the same wave vector /c, '-J ,, are activated simultaneously: thus forming coherent diffracted waves, propagating parallel to the same direction of propagation 0^ towards the pupil P: the waves refracted by the pupil converge in the same pixel R(i, ) on the retina.

Simultaneously, another emission subset EPD^ji can produce coherent diffracted waves propagating towards the pupil P, according to direction of propagation 9^ the waves refracted by the pupil converging in the same pixel on the image formed at the level of the retina, the emission points of the EPD^ji emission subset being illuminated by light extracted from other light guides illuminated by another laser light source or the same laser light source and activated by the same group of electrodes.

Emission points EP of different emission subset EPDjj can be activated sequentially, so as to converge sequentially on different pixels on the retina.

The electrodes 13 and the light guides 11 are distributed forming groups. An emission sub-assembly EPDj corresponds to emission points EP formed by superimposing the electrodes of a group of electrodes on light guides of the group of light guides. Thus, each emission subset EPDjj, forming the same pixel 7?(i,j), is associated with a single group of light guides 11 and with a single group of electrodes 13.

Due to retinal persistence, when the frequency of activation of the different emission subsets is sufficiently fast, the user perceives an image formed by the different pixels R(i,f).

The design of a projection device as previously described is confronted with constraints of compactness: it is considered that the diameter D _p of a pupil is of the order of 4 mm. Furthermore, the emission surface S is greater than the surface of a pupil, so that the eye can move vis-à-vis the surface, without any degradation of the image formed on the retina is seen.

Another requirement concerns the density of emission points. Each pixel R(i, of the image projected on the retina corresponds to an emission subset EPDjj of emission points EP associated with the same direction of propagation9j . The more the number of emission points EP contributing to the same pixel is high, the better the quality of the pixel.The number of emission points EP of a same emission subset EPDjj is preferably greater than 40, or even greater than 55 or 60. Furthermore, the number of subsets of emission points EPDj corresponds to the number of pixels formed on the image. In order to increase the number of pixels of the image formed on the retina, it is necessary to increase the number of emission subsets EPDij respectively associated with different directions of propagation θ^.

Figure 2 illustrates an arrangement of electrodes and light guides as described in US9632317. According to this arrangement: the light guides 11 extend, along a surface parallel to the emission surface, according to sinusoidal curves, along a longitudinal axis X. Two adjacent light guides are translated relative to each other. the other, perpendicular to the axis X. the electrodes 13 extend, according to a surface parallel to the emission surface, according to sinusoidal curves, along a lateral axis Y. Two adjacent electrodes are translated the one relative to the other, perpendicular to the Y axis.

The light guides are connected to laser light sources 21, for example laser diodes. Electrodes 13 are connected to switches 23. Each switch is configured to activate or deactivate each electrode to which it is connected.

Other arrangements of light guides and electrodes are possible, for example the arrangement described in application FR2105017.

However, regardless of the arrangement of light guides and electrodes, the number of different laser light sources to be used is high. For example, the device described in US9632317 comprises 400 emission sub-assemblies, each emission sub-assembly being formed by 10 waveguides. This requires the use of 400 laser light sources.

In order to reduce the number of laser light sources, the inventors propose to couple different light guides 11 to the same laser source 21, so as to simultaneously form different pixels, called twin pixels, spaced from each other. The intensity of the light propagating in the light guides connected to the same light source can be modulated by placing a modulator 22 between each light guide connected to the laser source. Thus, the intensity of each twin pixel can be adjusted independently. In general, the number of modulator associated with each laser light source corresponds to the number of twin pixels that can be simultaneously formed by the laser light source. Each twin pixel corresponds to an emission subset, referred to as a twin emission subset. However, the inventors have observed that the use of the same laser light source to simultaneously form twin pixels can lead to the appearance of interferences at the level of the image formed on the retina. This is because the same laser light source is the source of different coherent light waves which respectively propagate to different pixels, which can generate cross-interference.

Figures 3A to 3C illustrate spatial intensity distributions of light waves propagating to a retinal pixel. Experimental studies have shown that the light intensity, formed by a subset of EPDij emission, around a retinal pixel can be modeled by a sum of two two-dimensional Gaussians: a first two-dimensional Gaussian, of intensity I _lt and the half-waist w _lt which represents the useful signal; a second two-dimensional Gaussian, of intensity I ₂ , and half-waist w ₂ , which represents the noise. The width w ₂ is due to diffraction effects by the emission points EP. The smaller the latter, the more the diffraction effects increase.

The term “waist”, literally meaning size, corresponds to a characteristic width. It is usually the width of the Gaussian corresponding to a height equal to the maximum height multiplied by e~ ² .

If x denotes a coordinate along a line drawn on the image projected on the retina, centered on a pixel 7?(i,y), the intensity of the light produced by the emission subset EPDj can be expressed according to the relationship: with f ₀ is the focal length of the eye;

D _p is the diameter or diagonal of the support as shown in Figure IC;

A is the wavelength.

FIG. 3A represents the intensity /(%) (axis of ordinates) after normalization by the maximum intensity measured at the coordinate x=0, as a function of x (axis of abscissas—unit pm). In FIG. 3A, the modeled intensity has been represented (curve a), as well as the two Gaussians corresponding respectively to the useful signal and to the noise: curves b and c. Expression (1) can be expressed as a function of an angle ip _x with respect to a reference direction connecting the pupil to the center of the pixel. The angle ip _x is obtained from x according to: with

FIG. 3B represents the intensity I( _x ) (axis of the ordinates) after normalization by the maximum intensity modeled at the coordinate ip _x =0, as a function of ip _x (axis of the abscissas-unit degree). In FIG. 3B, the modeled intensity has been represented (curve a), as well as the two Gaussians corresponding respectively to the useful signal and to the noise: curves b and c.

FIGS. 3A and 3B have been established by considering that each emission point extends along a disk 10 μm in diameter.

When the same laser light source illuminates two different twin emission subsets EPDjj, EPDjiji in order to avoid the appearance of noise on the image due to cross interference, it is preferable that the two Gaussians modeling the noise , corresponding to curves c) of FIGS. 3A and 3B, do not overlap, or as little as possible. Thus, the angular directions respectively associated with the two twin transmission sub-assemblies

EPDj, EPDjiji are preferably spaced apart by an angular difference O- _min greater than or equal to the angle 2fl ₂ , that is to say twice the half-waist of the Gaussian representative of the noise.

The pixels 7?(i, j) and R(i'>j') respectively defined by the two emission subsets EPDjj, EPDj>j> are designated by the term “twin pixels”, because they are generated simultaneously by the same laser light source. The emission subsets EPDjj, EPDj>j> respectively associated with the twin pixels are called twin transmit subsets. The twin pixels are separated, on the image formed on the retina, by a minimum angle greater than the angle 2fl ₂ , so as to avoid cross interference. The twin emission subsets are configured to be simultaneously activated by electrodes, so as to simultaneously display the twin pixels.

FIG. 4 illustrates projections, on the retina, of rays from the same point of the pupil P. The pixels of the image formed on the retina have been represented, in the form of a grid. Four “twin” pixels have also been shown, corresponding to dark pixels. Around of each twin pixel, a dashed circle has been drawn, the diameter of which is equal to 2 times the half-waist w ₂ of the Gaussian representative of the noise. In order to avoid an overlap of two different Gaussians, representative of the noise, which would be conducive to the formation of cross interference, the distance between two twin pixels must be greater than a minimum distance w _min such that w _min = 2w ₂ .

The elementary angular difference /3 between two adjacent pixels, mentioned in connection with FIG. 1B, has also been shown.

The minimum distance w _min can be expressed as minimum angular deviation Sl _min . Using (3), we can write

The minimum angular deviation Sl _min is equal to 2fl ₂ with fl ₂ = asin -y (6')

Thus, an important aspect of the invention is that two twin pixels are separated from each other by an angular deviation fl greater than or equal to the minimum angular deviation Sl _min . The minimum angular deviation is defined by modeling the noise affecting the pixel when several twin pixels are formed. The minimum angular deviation is k times the elementary angular deviation P separating two adjacent pixels, where k is an integer greater than or equal to 2.

The inventors have quantified the signal-to-noise ratio associated with the intensity projected by each pixel. The projected intensity corresponds to the percussion response of the system formed by the emission medium, the pupil and the lens, usually designated by the acronym PSF (Point Spread Function). The PSF corresponds to the image of a point on the retina. Figure 3C illustrates PSF, depicting a spatial intensity distribution on the retina. In FIG. 3C, the ordinate axis corresponds to the light intensity /(%) of the image formed on the retina and the abscissa axis corresponds to a distance x from the origin, the origin corresponding at the center of the pixel formed on the retina.

The signal to noise ratio (SNR) is such that:

/(O) is the height of the central peak; w is the half-waist of the Gaussian representative of the useful signal, described in connection with FIGS. 3A and 3B; max (/(%)) represents the maximum value of the intensity outside the central peak |x|>W of intensity.

The inventors have calculated the signal-to-noise ratio of a pixel obtained by an emission subset of 80 emission points on an emission surface of dimensions 6 mm by 6 mm. In order to simulate the effect of twin pixels neighboring the modeled pixel, they assumed that each EP emission point contributing to a twin pixel forms a plane wave, whose intensity is modulated by a Gaussian envelope, as described in link with FIGS. 3A to 3C (gaussian representative of the noise).

The electric field of the plane wave is such that: x ² +y ²

E(x, y) oc e ^{i kxX+k} y ^y ' e ^i<t> e ^{w 2} (8) with where: x _EP and y _EP correspond to the coordinates of the emission point EP on the emission medium 10, as represented in FIG. IC;

Z _o is the distance between the emission support 10 and the pupil; x and y are coordinates on the retina, along the X and Y axes shown in Figure 1C. It is assumed that the emission medium is parallel to the retina or can be considered as such.

O is a phase shift of the light wave emitted by the emission point EP, this phase shift being adjusted as a function of the coordinates of each emission point so that the wave resulting from the different emission points of the same subset emission is a plane wave. where θ and Q _y are the components of the angle of propagation θ i with respect to the X and Y axes respectively. The inventors used expression (8) to simulate an image formed solely by a central pixel and four neighboring pixels regularly distributed around the central pixel. By neighboring pixel is meant a pixel situated in the vicinity of the central pixel. It is not the adjacent pixel. In order to estimate the noise in the image, the useful signal of the four neighboring pixels has been removed. We kept the background noise generated around each pixel: central pixel and neighboring pixels. The four neighboring pixels have been placed at different distances from the central pixel.

In a first configuration, neighboring pixels formed by waves coherent with the light wave forming the central pixel have been simulated. This case is representative of the use of the same laser light source to respectively illuminate each emission subset forming each pixel. This is the configuration according to the invention, in which it is sought to maximize the number of emission sub-assemblies illuminated by the same laser light source. It is also the least favorable configuration with respect to the formation of cross-interferences.

In a second configuration, neighboring pixels formed by light waves that are not coherent with the light wave forming the central pixel have been simulated. This case is representative of the use of five different laser light sources to respectively illuminate each emission subset forming each pixel. This is a reference configuration, representative of the prior art.

Figure 5A shows the evolution of the SNR signal-to-noise ratio, as defined in expression (7) (ordinate axis) as a function of the angular distance between the central pixel and each neighboring pixel (abscissa axis - unit degree ). Curve a) corresponds to the configuration according to the invention: the central pixel and each neighboring pixel are twin pixels. The dotted line corresponds to the signal-to-noise ratio of the central pixel without taking into account the neighboring pixels. Curve b) corresponds to the reference configuration.

FIG. 5B shows simulated images considering distances between the central pixel and the neighboring pixel of 5 μm, 25 μm and 525 μm. These distances correspond respectively to angular deviations of 0.0125°, 0.062°, and 1.35°. In FIG. 5B, the configuration according to the invention has been taken into account. We observe the background noise and its attenuation as the distance between twin pixels increases.

Figure 5C shows images similar to those shown in Figure 5B, taking into account the reference configuration. Figure 5A shows that twin pixels that are too close together result in a large decrease in SNR. By moving the twin pixels apart by an angle greater than 1°, or even 1.5°, the impact on the SNR is lower. Beyond 2°, the impact on the SNR is negligible.

FIGS. 6A and 6B schematize an addressing method according to the prior art. To simplify, these figures show light guides 11 and linear electrodes 13, the electrodes being perpendicular to the light guides. There are 16 light guides Hi, II2...II16 and four electrodes 13i, 132, 13a, 13 ₄ . Each light guide is connected to a laser light source 21. Each electrode 13 is connected to a switch 23: the electrodes 13i, 13a are connected to a first switch 23i and the electrodes 132, ₁₃₄ are connected to a second switch 232 The switches are conductive and connected to a power supply.

Each emission point EP, schematized by an oval shape, is placed at the intersection of a light guide and an electrode. Each emission point is assigned a label, between 1 and 16, identifying the EPD emission subset to which the emission point belongs. In this example, each EPD transmit subset has four pixels. There are 16 different transmission subsets. FIG. 6B represents the 16 pixels respectively formed by the 16 emission subsets after self-focusing. The emission subassemblies 1 and 9 are activated when the electrodes 13i and ₁₃₃ are activated, and when the laser connected to the waveguides 111 and ₁₁₈ is activated. In FIG. 6B, a bold frame shows the pixels formed when the electrodes 132 and ₁₃₄ are activated.

FIGS. 7A and 7B schematize an addressing method according to the invention. Only two lasers 21i, 2I2 are used, the latter being coupled to 8 different waveguides. The device comprises modulators 22 arranged between each laser and a set of waveguides. The modulators allow modulation of the intensity of the light propagating towards each emission subset. Intensity modulation makes it possible to form pixels of different light intensities with the same laser. In FIG. 7A, the shaded emission points are coupled to the laser 2I2. It is the same in FIG. 7B: the shaded pixels are emitted by an illumination of emission subsets by the laser 2I2.

FIG. 7B shows the pixels obtained: following the activation of the laser 21i and of the switch 23 ₄ : this allows display of the twin pixels 1, 2, 3 and 4: pixels displayed in white and framed by a thin line. following the activation of the laser 21i and the switch 232: this allows a display of the twin pixels 9, 10, 11 and 12: pixels displayed in gray and framed by a thin line, following the activation of the laser 212 and the switch 23i: this allows display of twin pixels 5, 6, 7, 8: pixels displayed in white and framed by a bold line. following the activation of the laser 21 ₂ and of the switch 23 ₂ : this allows display of the twin pixels 13, 14, 15 and 16: pixels displayed in gray and framed by a bold line.

It is thus possible to form 16 pixels using only two laser diodes.

The modulators 22 make it possible to modulate the intensity of the pixels 1, 2, 3, 4, 9, 10, 11 and 12, activated by the first laser 21i, as well as the intensity of the pixels 5, 6, 7, 8, 13 , 14, 15 and 16 activated by the second laser 2I2. So that the intensities of pixels 1, 2, 3 4 on the one hand and 9, 10, 11, 12 on the other hand can be modulated independently of each other, the power supplies 23i and 232 are successively activated.

Pixels 1 and 9 are respectively generated by a first emission subset, bringing together the emission points bearing, in FIG. 7A, the label 1 and a second emission subset, bringing together the emission points bearing, in FIG. 7A, the label 9. The emission points belonging respectively to the first and second emission sub-assemblies are formed on the same waveguides, connected to the same modulator 22. In order to be able to independently adjust the intensities of pixels 1 and 9, the first emission subset (pixel 1) is activated by first electrodes (13i, 183) while the second emission subset (pixel 9) is activated by second electrodes (132, 134). The sequential activation of the first and second electrodes allows the sequential display of pixels 1 and 9. This allows the modulation of their respective intensities by adjusting the modulator during each display. This aspect also concerns pixels 2 and 10, 3 and 11, 4 and 12, 5 and 13, 6 and 14, 7 and 15, 8 and 16. Thus, the use of sequential activation of different electrodes is relevant when the points emission of different emission subsets are distributed over the same waveguides. The selection of an electrode makes it possible to activate only one emission subset on said waveguides. This allows the sequential display of the pixels respectively associated with the emission subsets formed on common waveguides.

Figures 8A and 8B show an arrangement similar to that described in connection with Figures 7A and 7B. In these figures, only the switch 23i is on: this allows a display simultaneous pixels 1, 2, 3, 4 (with the laser source 21i) and pixels 5, 6, 7, 8 (with the laser source 21 ₂ ). Pixels 1, 2, 3 and 4 are twins: pixels shown in white in FIG. 8B, framed by a thin line. The same applies to pixels 5, 6, 1, 8: pixels shown in gray in FIG. 8B, framed by a fine line. The non-displayed pixels are surrounded by a dotted frame: these are pixels 9, 10, 11, 12, 13, 14 and 15.

FIGS. 9A and 9B represent a symmetrical configuration: only switch 23 ₂ is on: this allows display of pixels 9, 10, 11, 12 (with laser source 21i) and pixels 13, 14, 15, 16 (with the laser source 21 ₂ ). Pixels 9, 10, 11 and 12 are twins: pixels shown in white in FIG. 9B, framed by a bold line. The same applies to pixels 13, 14, 15 and 16: pixels represented in gray in FIG. 9B, framed by a bold line. The non-displayed pixels are surrounded by a dotted frame: these are pixels 1 to 8.

The configurations schematized in FIGS. 8A-8B as well as 9A-9B can be used successively, according to a sufficiently high repetition period so that, under the effect of retinal persistence, all the pixels appear simultaneously displayed. The passage from one configuration to the other is carried out by switching the switches 23i and 23 ₂ . The switching makes it possible to activate only one transmission sub-assembly among different transmission sub-assemblies arranged on common waveguides. The emission sub-assemblies connected to the same laser light source, and distributed on different waveguides from each other, activated by common electrodes, form twin pixels: they are simultaneously activated, by the electrodes, so as to simultaneously form several pixels.

The width of the two-dimensional Gaussian w ₂ representing the noise around a pixel depends on the size of the emission points EP. The more the size of the emission points decreases, the more the diffraction effects increase. The inventors have studied the influence of the size of the emission points on the SNR. For this, they carried out the modeling described in connection with FIG. 5A (curve a) for different sizes of emission points.

FIG. 10 represents the evolution of the SNR, as a function of the angular difference between the central pixel and four twin neighboring pixels, for different sizes comprised between 5 μm and 40 μm. We have also represented the SNR obtained with a single pixel for each size: cf. dotted lines. Figure 10 was drawn based on disc-shaped emission points It is observed that the smaller the size of the emission points, the more the SNR is degraded. It is particularly preferable for the size of each emission point to be greater than or equal to 10 μm when circular emission points are used.

Sizing example.

Considering the foregoing, it is possible to dimension a device 1 comprising an emission medium 10 defining a field of observation of 15°. The emission support extends over a surface of 6mm x 6 mm. Its diameter D _p is considered to be 6 mm. Each waveguide 11, 300 nm wide, was spaced apart by a period dn=1.5 μm. Each electrode is spaced from one another by a period d ₁₃ =5 μm. A number of waveguides Nu = 9 and a number of electrodes N _i3 = 9 are used per emission subassembly EPD.

We can thus form a number of pixels N _pix such that

Given the values previously indicated, the number of pixels N _pix is equal to 59000, i.e. 243 x 243.

Using a configuration as shown in Figures 6A and 6B, the number of laser light sources would be 444.

By taking into account emission points of size 10 μm×10 μm, a minimum deviation = 2fl ₂ = 1.87° is obtained, the twin pixels can be distributed according to a square tiling, as shown in FIG. 11. A first group of twin pixels can comprise 8×8 pixels, regularly distributed according to a square mesh according to the spatial pitch of 750 μm, ie a total image width of 5.5 mm. The other groups of pixels are arranged regularly in a frame of side 5.5 mm, either in groups of 8x8, or in groups of 7x7. This leads to a total of ten groups of twin pixels, which requires only ten laser diodes. FIG. 12 shows the first group of pixels, containing 8×8 pixels, and a second group of pixels, comprising 7×7 pixels.

The number of laser light sources required to tile the image regularly can be estimated by the expression: where FOV corresponds to the field of view and 11 corresponds to the angular difference between two twin pixels. N _taser corresponds to the number of lasers in the reference configuration, i.e. 444 lasers.

The lower the number of laser light sources, the greater the number of twin pixels, which, for a fixed image size, is likely to increase noise. The inventors have calculated a noise factor r], which corresponds to a ratio between the noise obtained by a configuration of a central pixel surrounded by 4 neighboring pixels, respectively with 5 different laser sources and with the same laser source. This amounts to comparing the noises respectively obtained in the configurations according to the invention and of the reference described in connection with FIGS. 5A to 5C.

The noise factor is expressed according to the expression: where B and B ₂ are respectively the noise amplitudes respectively in the first configuration (invention) and the reference configuration. Each noise amplitude is such that B=max (/(%)) in the configuration concerned. This is the denominator of the expression (7).

The noise factor r], expressed as a percentage, is represented in FIG. 12: curve a: left ordinate axis. In FIG. 12, the abscissa axis corresponds to the minimum angle between twin pixels, unit degree. Curve b: represents the number of laser light sources necessary for the system (right ordinate axis) as a function of the angular pitch between two twin pixels (abscissa axis). The more the angular pitch decreases, the more the number of light sources also decreases. It can be seen that by considering an angular pitch comprised between 1.5° and 2°, for example 1.8 or 1.9°, an acceptable noise figure is obtained for a reasonable number of laser sources. By reducing the angle between two twin pixels, it is possible to implement a reduced number of laser light sources, to the detriment of strong cross-interference effects. By increasing the angle between two twin pixels, the effects of cross interference are reduced, to the detriment of the number of laser light sources to be used. In Figure 12, the curve has been extended to the configuration of the prior art (444 laser sources)

The invention may be implemented on portable augmented reality devices, in particular glasses, or visors, or screens, and more generally, on surfaces, intended to be arranged facing an eye so as to form an image superimposed on the image perceived by the eye on the surface or through the surface.

Claims

26 CLAIMS 1. Device (1) for projecting an image onto the retina of an eye, the image being composed of several pixels, the device comprising: - laser light sources (21); - a transmission medium (10), comprising: • light guides (11), each light guide being connected to a light source and to a plurality of diffraction gratings (12), distributed along the light guide, each diffraction grating being electrically adjustable to extract a part of the light propagating in the light guide to which it is connected; • electrodes (13), each electrode being configured to modulate diffraction gratings connected to different light guides; • emission points (EP), each emission point being formed at the level of a diffraction grating, and extending between an electrode and a light guide, each emission point being configured to emit light , in an angular direction (Qÿ), following an extraction of a light propagating in the light guide; the device being such that: different emission points (EP) are configured to emit light in the same angular direction (0^), so as to converge on the retina, forming the same pixel (R(i,j) ), said emission points defining an emission subset (EPDj ), so that each emission subset corresponds to an angular direction of emission of the light; two adjacent pixels (R(j,j), (R(i+1,/))) of the image are formed by two emission subsets (EPDij, EPD _i+1 j)), respectively associated with two angular directions (0, _y , 0 _i+1 ) different and spaced apart by an elementary angular difference (/3); the device being characterized in that: - the same light source is connected to different light guides, said light guides defining different emission sub-assemblies, called twins, configured to simultaneously form different pixels, called twin pixels, when the light source is activated, the twin emission sub-assemblies being associated with the light source; - the respective angular directions of each twin emission sub-assembly are spaced from each other by an angular difference (fl) greater than or equal to a difference minimum angular the minimum angular deviation corresponding to k times the elementary angular deviation (/3), k being greater than or equal to 2; - in such a way that the twin pixels, simultaneously formed on the retina, are shifted by at least the minimum angular deviation. 2. Device according to claim 1, in which the emission points of different twin emission sub-assemblies are connected to the same electrode. 3. Device according to claim 1, wherein: - several light sources are respectively connected to different light guides, to form different twin emission sub-assemblies, each twin emission sub-assembly being associated with the same light source; - the respective angular directions of each twin emission sub-assembly are spaced from each other by an angular deviation greater than the minimum angular deviation. 4. Device according to any one of the preceding claims, in which the waveguides, connected to the same light source, and forming a twin emission subassembly are different from other waveguides, connected to said light source, and forming another twin emission subassembly. 5. Device according to any one of the preceding claims, in which the minimum angular deviation (n _m i„) is greater than or equal to 0.5° or 1°. 6. Device according to any one of the preceding claims, in which: - the diameter or the diagonal of the emission points is less than or equal to 10 pm, the minimum angular deviation being greater than 1.5; - the diameter or the diagonal of the emission points is between 10 μm and 20 μm, the minimum angular difference being greater than 0.8; - the diameter or the diagonal of the emission points is greater than 20 pm, the minimum angular deviation being greater than 0.5; 7. Device according to any one of the preceding claims, in which the diameter or the diagonal of each emission point is greater than 5 μm. 8. Device according to any one of the preceding claims, in which - the light guides connected to the same light source define first twin emission sub-assemblies and second twin emission sub-assemblies, so as to respectively form first twin pixels and second twin pixels; - the emission points of each first twin emission subset are connected to first electrodes; - The emission points of each second twin emission subassembly are connected to second electrodes, different from the first electrodes; - the device comprises a switch (33), for supplying either the first electrodes or the second electrodes, so that the light emitted by the light source is extracted sequentially either by the emission points of the first subset of emission, or by the emission points of the second emission subsets, to form sequentially, on the image, the first twin pixels and the second twin pixels. 9. Device according to any one of the preceding claims, in which for at least one light source, different modulators (22) are respectively interposed between a laser light source and different light guides connected to the light source, so as to modulating an intensity of light propagating in said light guides independently for different light guides connected to the light source. 10. Device according to claim 9, in which: - A light source is connected to N different twin emission sub-assemblies, configured to simultaneously form N twin pixels, N being an integer greater than or equal to 2; - the light source is connected to N modulators (22), each modulator extending between the light source and the waveguides forming the same emission sub-assembly. 11. Device according to any one of the preceding claims, in which the number of twin emission subassemblies formed by at least one light source is greater than or equal to 20. 12. Device according to any one of the preceding claims, comprising less than 20 different light sources to form an image, or part of an image, of at least 200 pixels by 200 pixels. 29 13. Device according to any one of the preceding claims, comprising a holographic film (14), subdivided into different holograms, each hologram being associated with a diffraction grating (12), and configured to emit the light wave, according to the direction emission point (0^), under the effect of light extracted by the diffraction grating with which it is associated, each association between a hologram and a diffraction grating forming an emission point (EP) of light. 14. Bezel, comprising bezel lenses, the bezel comprising a device according to any one of the preceding claims, the stack (10) being formed on at least one bezel lens. 15. Method for parameterizing a device according to any one of claims 1 to 13, the parameterization comprising: a) modeling of a spatial distribution of the noise between two twin pixels, by modifying the minimum angular deviation; b) estimation of the minimum angular deviation as a function of the spatial distribution.