The present invention claims the priority of European Patent Application EP20180182654, filed on Jul. 10, 2018, the content of which is incorporated here by reference.
The present invention concerns a micro-optical component for generating an image. The present invention concerns also an imaging device comprising this micro-optical component, a method for manufacturing the micro-optical component and a method for manufacturing the imaging device.
Nowadays, imaging devices, as for example and in a non-limiting way micro cameras, endoscopic cameras, light-weight vision systems for flying robots, and similar devices comprise image sensors connected, preferably directly connected, to ultra-thin or very flat optical components, so as to obtain an optic height in the order of few millimeters.
In this context, the optic height is the height (or the thickness) of optical component from the surface of the optical component cooperating with the image sensor to the interface of the optical component with the external surface.
In this context, a micro-optical component (or ultra-thin optical component or very flat optical component) is an optical component having an optic height in the order of few millimeters, e.g. 2 mm, 1 mm or smaller.
Such micro-optical components can be manufactured in precise and cost efficient wafer-scale processes like resist reflow for mastering and uv-replication for reproduction, which makes these technologies ideal candidates in order to obtain optic heights in the order of few millimeters. Other possible mastering methods are direct laser writing, grey-scale lithography, ultra-precision diamond turning/micro-machining. The considered lens, which is in general a single spherical lens, is replicated on a substrate, e.g. a glass chip, with an appropriate aperture and thereafter attached, in particular directly attached, to an image sensor.
In this context, a spherical lens is a lens having a spherical profile. In other words, a spherical lens is a convex lens whose cut section comprises an arc of circle.
An example of a known micro-optical component 100′ is visible on
The aperture 12 is for example a circular aperture defined by a substrate obscured area or layer 120 created on the first substrate 21 before the lens 10 is placed on the first surface 21. The lens 10 is created on the first substrate 21 so as to cover, in particular completely cover, the aperture 12, e.g. by using the uv-embossing fabrication process. This substrate obscured area 120 can for example be obtained by coating at least a part of the first surface 21 with a material preventing the light to pass through it, this coating being performed before creating the lens 20 on the first surface 21 of the substrate 20.
The lens 10 of
The lens 10 of
The micro-optical component 100′ has an optic height h typically equal or less than 1 mm. The thickness of the substrate is a function of the focal length of the lens, e.g. its radius. The more flat the lens, the longer the focal length, and the thicker the substrate.
The known micro-optical component of
However, the known micro-optical component of
In this context, the expression “field of view” is a measure (e.g. in degrees) of the angular range of a scene or a subject the lens can take in.
In this context, the expression “angle of incidence” (AOI) indicates the angle between a ray incident on the lens and the normal of the substrate at the point of incidence.
First, large field of views suffer most from the intensity fall-off at the image edge.
Secondly, for wide viewing angles a single spherical lens suffers from spherical aberrations, visible in
Therefore, there is a need for a micro-optical component in which the problems of intensity fall-off at the image edge and/or the image blur due to spherical aberrations for wide field of view can be mitigated or reduced with regard to the known solution.
There is also a need for a micro-optical component which equalizes the intensities for central and marginal AOIs while reducing the aberration by suppressing aberrated and blinding rays.
There is also a need for a micro-optical component adapted to an image sensor having intelligent pixels with a low fill factor.
According to the invention, these aims are achieved by means of the optical component according to claim 1. These aims are also achieved by means of the imaging device according to claim 13, by means of the method for manufacturing the micro-optical component according to claim 15 and by means of the method for manufacturing the imaging device according to claim 17.
The micro-optical component for generating an image on an array of light detective elements, e.g. pixels, according to the invention comprises:
The invention will be better understood with the aid of the description of an embodiment given by way of example and illustrated by the figures, in which:
The presence of a substrate 20 allows to avoid internal reflection in the lens 10, in particular for large AOIs.
In one preferred embodiment, the refractive index of the substrate 20 is substantially equal to the refractive index of the lens 10.
According to the invention, the micro-optical component 100 of
According to the invention, the area of the obscuration element is less than the area of the lens 10, so that at least a first part of a light beam incident to the lens pass through the lens, thereby generating the image on an array of light detective elements 300, this obscuration element preventing at least a second part of said light beam to pass through the lens, so as to improve the intensity homogeneity of the image and/or reduce image blur due to spherical aberrations.
In one preferred embodiment, an example of which is illustrated in
In another (not illustrated) embodiment, the obscuration element is on the internal surface of the lens 10. However it must be understood that the obscuration element is not necessarily on a surface of the lens, but can be in alterative or in complement in the lens: e.g. if the lens replication is done not in a single layer but in multiple layers, the obscuration can be processed on any surface level inside the lens or even in front of the lens.
In another (not illustrated) embodiment, the obscuration element in front of the lens or behind the lens.
In the embodiment of
The light blocking function of the obscuration is angle dependent. In other words, the obscuration element improves the image resolution for a first range of the angles of incidence while lowers the image resolution for a second range of the angles of incidence.
First considering a lens without obscuration as shown in
If Δ≤|r−robs| then there is the maximum blocking effect
If |r−robs|<Δ≤r+robs the blocking effect is increasing
If Δ>|r+robs| there is no blocking effect
Typical values are (robs,0)2=ε*r02, ε=20%-50%, zobs=0.3 mm, r0=0.15 mm, so that the blocking effect is decreasing within the range 11°-27°<AOI<41°-50°. In other words, the obscuration element affects the intensity of almost all viewing angles except those larger than 41°-50°.
Therefore, the transmitted power of viewing angles close to the optical axis is reduced most. Thus, the benefit of such an obscuration is to decrease the intensity difference of central and marginal parts of the image.
The realization of an obscured area for a lens is surprising, because an obscuration at the entrance pupil has a negative effect on the resolution of the lens 10.
In this context, the obscuration level 6 is defined as the ratio between the obscured area of the entrance pupil EP of the lens and the total area of the entrance pupil EP of the lens.
As visible in
The point spread function of a lens is known as “Airy pattern” and the relation of the central peak to the function side lobes is called “Strehl ratio”.
As visible in
By resuming, by increasing the obscured area, the Strehl ratio and thereby the optical quality of the image are decreased.
In one preferred embodiment, the area of the obscuration element allows to have a micro-optical component adapted for wide fields of view, without decreasing the optical quality of the image due to the presence of the obscured area.
In one preferred embodiment, the invention is therefore based on such a trade-off. The optical resolution of the lens is defined by the cut-off frequency of the point-spread function (cf Airy pattern and diffraction limit), defining a limit for the transmitted spatial frequency of a lens having a F-number F for a minimum MTF contrast of 50% at a given wavelength λ, according to the following formula:
f
cut-off_o=1 line/(λF) (1)
Typical values for the optical resolution of the lens fcut-off_o can be in the order of several hundred lines per mm; especially low power systems with the necessity of small F # numbers (F #<2) might have cut-off frequencies fcut-off_o)>600 lines/mm, and a MTF contrast of 50% is achieved for spatial frequencies in the order of 150 lines/mm. In practical systems, such a resolution is not required. Thus, the lens according to the invention is an applicable solution if the required resolution of the pixels is smaller than the optical resolution defined by equation (1), and the required MTF contrast.
In one embodiment, the applicant has discovered that the range from 20% to 50% for the obscuration level 6 allows to obtain satisfying results.
Since the obscuration element prevents at least a part of a light beam to pass through the lens, the intensity homogeneity of the image is improved and/or the image blur due to spherical aberrations reduced. On the other hand, the area of the obscuration element allows also that the optical resolution of the micro-optical component is equal or larger than the image resolution given by the light detective elements.
By properly selecting this area, it is possible to realize a micro-optical component in which the problems of intensity fall-off at the image edge and/or the image blur due to spherical aberrations for wide fields of view of the known micro-optical component using single spherical lens (with no obscuration) are reduced, without compromising the optical quality of the image due to the presence of the obscured area of the lens.
Moreover, especially for large sensor pixels, i.e. for pixels having a pitch equal or larger than 3 μm, e.g. 5 μm, the required resolution by the sensor's pixel size is much less than the resolution offered by the unobscured lens; in this case an obscuration in the entrance pupil diameter will not show a visible effect in the image.
This range has been selected by considering both the cut-off frequency by the point-spread function (equation (1)), and the cut-off frequency by pixel pitch, defining the maximum spatial frequency that can be transmitted by an image sensor with pixel pitch p, according to following formula:
f
cut-off_p=1 line/(2p) (2)
Advantageously, the total optical track length (TOTR) of the micro-optical component according to the invention, i.e. the sum of the back focal length (required optical path length after the lens) and the rear focal length (required optical path length in the lens) is 1 mm or less.
In the embodiment illustrated in
In the embodiment illustrated in
This shifted aperture allows to increase the resolution at parts of the image at the cost of image resolution in other parts. The spherical lenses with shifted aperture are in general placed on the array 300 in such a way that only the best resolution parts of the image (or sub-images) are on the photosensitive pixels.
In one preferred embodiment, the final image is obtained by stitching together of those sub-images.
As the aperture underneath the spherical lens 10 is shifted out of the center C of the spherical lens 10, aberrations are reduced for the diametrical part of the image at the price of higher aberrations in the rest of the image. Moreover, the illumination for all the AOIs is equalized and/or double images are avoided with regard to the embodiment of
In one embodiment, the back focal length can also be optimized for the outer field of view.
In one preferred embodiment, if the lens 10 of the micro-optical component according to the invention is spherical and with a shifted aperture (as in the example of
The presence of the shifted aperture causes the capture of an image wherein only a portion of the image is usable, the size of the portion depending on the shifting of the aperture with respect to the axis of symmetry A of the lens 10.
Therefore, in one embodiment, the micro-optical component according to the invention comprises more than one spherical and obscured lens 10 with a shifted aperture, the at least two lenses 20 being identical. The apertures under each of the at least two lenses are identical as well.
In this case, the image sensor is arranged to communicate with a processor arranged to shift and/or accurately combine the images from each lens, so as to form a full image. In fact, the use of more than one lens requires image reconstruction.
In one preferred embodiment, image processing is minimized in order to economize energy.
In case of very large pixels, the image resolution of the very flat optics is already sufficient to resolve the active diode area of the image sensor only and the full resolution information is immediately available in the images.
In case of a low resolution optics, i.e. a resolution that is not lower than the pixel pitch, image deblurring techniques like deconvolution algorithms can be applied to achieve the desired resolution.
In one preferred embodiment, the number of those lenses 10 is even. In another preferred embodiment, those lenses are arranged in a
M×M square, wherein M is the number of lenses per side.
As will be discussed, the use of multiple lenses allows to increase the resolution of the final image, for example by shifted sub-sampling or by spatial oversampling.
In one preferred embodiment, M=2 so that there are four identical spherical and obscured lenses 10 on the substrate 20, each lens cooperating with a shifted aperture. In other words, the scene is collected by four separated lenses 10. The shift of their aperture 12 is selected so that each lens 10 gives an image wherein only a quadrant (i.e. ¼ of the image) is usable.
In one preferred embodiment, as will be discussed, the pitch between two consecutive lenses LP (i.e. the distance between the axes of symmetry of two consecutive lenses) is selected according to the following formula:
LP=(N+½)·p (3)
wherein N is the number of pixels per row of the pixel array 300 with N*N pixels and p is the pitch of each pixel.
For a number M of lens, the previous formula can be generalised as following:
LP=(N/M+1/M)·p (4)
In one preferred embodiment, the pitch is selected so that it can enable the image recomposition by simple pixel interleaving.
In one preferred embodiment, those lenses share the same substrate 20, so that the micro-optical component can be easily manufactured (as the lens can be easily replicated e.g. via uv-replication) and can be safely connected, e.g. via a glue or an adhesive layer, to the image sensor with an accurate pitch towards each lens.
The micro-optical component 100 of
The substrate will be connected to an image sensor (not illustrated) via an adhesive or glue layer 130. Another adhesive or glue layer 140 allows to fix the lens 10 to the cover 400. In one embodiment, the image sensor has a size of about 4 mm×4 mm.
The first surface 21 of the substrate 20 is coated with the layer 120, defining the apertures (not visible) under each lens 10.
In one preferred embodiment, the aperture layer 120 is manufactured by lithography on the substrate 20. The preferred material is the black chromium. In one preferred embodiment, the lithography reticule allows different aperture values. The varying of the aperture allows to compensate fabrication tolerances, e.g. the tolerance of the substrate's thickness. For example, if the substrate is too thick, a compensation can be achieved by choosing a smaller aperture, for example a smaller aperture diameter. On the contrary, if the substrate is too thin, it is possible to increase the thickness of the glue layer for compensating it.
In one preferred embodiment, the diameter of the aperture can be varied in a range from 0.20 mm to 0.40 mm.
In the embodiment of
This mask cover 400 has a particular 3D shape allowing on one side to perform the obscuration of the lenses 10 and on the other side to protect the micro-optical component 100 as a sort of partial package.
The particularity of the embodiment of
Advantageously, the total height of the micro-optical component 100 of
The micro-optical component 100 of
The AOI=0° point of the assembly of four lenses, e.g. the assembly illustrated in
The aspherical lens 10′ of
However, aspherical lenses 10′ cooperating with shifted apertures could be imagined as well.
According to the invention, the aspherical lens 10′ comprises an obscured area 40 at the entrance pupil of the lens 10′, so that the ratio between the unobscured area of the entrance pupil and the total area of at the entrance pupil is comprised in the range from 20% to 50%.
In the embodiment of
One can imagine that the two ideal semi-circumferences are cut-sections of corresponding ideal spherical lenses, each of which cooperates with a shifted aperture. The distance between the centers C1 and C2 has been selected so that the aspherical lens 10′ is centered on the aperture 12. In other word, the aspherical profile of the lens 10′ is generated by cutting out the shift of the shifted aperture of the corresponding ideal spherical lenses of centers C1 and C2.
However, the aspherical profile of the aspherical lens 10′ induces a decrease in the tangential resolution. Further optimization allows to balance this tangential resolution and also radial resolution. In one embodiment, such optimization is based on building up a merit function with the Strehl ratio, the modulation contrast and the intensity distribution as parameters. By variation of the lens geometry, e.g. by variation of some polynomial parameters (the lens geometry can be described best by a polynomial function), the obscuration diameter, the lens thickness, etc., a merit function is recomputed, thereby optimizing the lens profile.
As the covered top of an aspherical lens 10′ has no optical purpose, in one preferred embodiment, illustrated in
Therefore, in one embodiment, the aspherical lens 10′ comprises a flattened and obscured top 17. In one preferred embodiment, this cut is realised at a height h, of about 0.30 mm from the substrate 20.
The embodiment with the flattened and obscured top 17 allows to have a micro-optical component 100 having a still more reduced optic height h, in particular a height of 1 mm or less (depending on the choice of the materials).
In particular, in
The illustrated aperture is circular, but other possible shapes could be imagined (e.g. in case of an astigmatic lens). The layer 120 can be fabricated with a process (e.g. the lithography) allowing the variation of the size of the aperture 12 (e.g. its diameter) for compensation as previously discussed.
Then, as illustrated in
Then, as illustrated in
In particular, in
As anticipated, the number of lenses is a feature useful to increase the resolution of the final image, for example by shifted subpixel sampling. This is a technique know from super-resolution imaging: a camera with a low resolution grid of pixels takes images, but in-between the images the camera is shifted laterally by a distance that is smaller than the pixel pitch (subpixel). These low resolution images are superimposed to a high resolution image.
According to an important embodiment of the invention, instead of shifting a camera, M×M lenses of the type shown in
Multiple images can be captured in parallel, in particular with ultra-flat lenses: for example if the image sensor has a total pixel number of 200×200 pixels, and if the subpixel sampling is performed with 2×2 lenses of
N/2×N/2=100×100 pixels, the resolution of the final image of the interlaced subpixel sampled images becomes:
(2×N/2)×(2×N/2)=200×200 pixels.
It is the identical resolution as if a single (larger) lens would have been used for imaging the image on the initial N×N pixels.
According to an important embodiment of the invention, the use of multiple but identical very flat optic lenses, each positioned accurately to allow sub-pixel sampling allows to compensate low fill factor pixels, i.e. the ratio between active photodiode area and pixel area.
In fact, especially for intelligent pixels, the pixel size is quite large and the pixels suffer from a low fill factor, since intelligence is incorporated into each pixel instead of being located at the circumference of the pixel field. The advantage of such sensors with sparse photo sensitive pixels is that the signal pre-processing in the pixel economises the time for data transfer to the sensor periphery; it goes in parallel with a decrease in the number of interconnects and a reduced noise sensitivity. Such an intelligent pixel may have a fill factor far below 70%, e.g. less than 50%. Here it is proposed to compensate the low fill factor by using the above-mentioned sub-pixel sampling.
In
The micro-optical component 100 of
With the micro-optical component 100 of
The micro-optical component according to the invention is well adapted to cooperate with pixels having a low fill factor, for example with pixels wherein the active area is allocated in one quadrant only. Each of the 2×2 lenses can be positioned to image an identical object point to one of the four quadrants of the pixel; thus no image processing for distant objects is necessary to interlace the images; and in theory, no loss in contrast will occur. In one preferred embodiment, the increased resolution is visible, if the optical resolution is designed for the subpixel size.
The scaling of the focal length and thereby the system height is a function of the number of incorporated lenses. The nominal FoV of an image sensor is resolved by N0 pixels of a sensor with pixel pitch p0 in a distance z and equipped with a single lens of focal lens f0. In case of M×M lenses, the focal length fM, the size of the subpixel pM and the number of pixel per lens NM scale with the number of lenses M, according to the following formulas:
f
M
=f
0
/M (5)
p
M
=p
0
/M (6)
N
M
=N
0
/M (7)
An ideal condition for subpixel sampling is a pixel array with a fill factor ff equal to
ff=p
M
2
/p
0
2.
Although the number of pixels (with a low fill-factor) is decreasing per lens, the final image that is reconstructed out of M×M images is of full-resolution. The scaling of the system height is only limited by the required optical resolution of the lens, which has to fulfil the subpixel size pM and is diffraction limited.
Another possibility to increase the resolution of the final image is to use multiple lenses, e.g. arranged in a square or in a rectangle, so as to perform spatial oversampling, as many identical images reduces noise and therefore increase the signal-to-noise ratio. In this case, the lenses' height allows only small focal lengths and hence only a limited number of pixels per image. However, the oversampling (taking several identical images by identical images) is requires high computational time. The favourite method for pixels with a low fill factor is sub-pixel sampling.
The drawback of the sub-sampling is a limited minimum object distance due to the parallax error and an increased effort in image processing for interleaving the sub-images.
Δr=(r2−r1)=off-set·f/z≤p/4 (8)
As the off-set is at least (N+1/M)·p:
(N+1/M)·p·f/z≤p/4 (9)
Therefore the minimum object distance (MOD) z is:
MOD≥4·f·(N+1/M)· (10)
For example, for a number of pixels N of 160 (i.e. for a resolution of 160×160) and a focal length f of 1 mm, the minimum object distance (MOD) is 642 mm for M=2. For a number of pixels N of 320 (i.e. for a resolution of 320×320) and a focal length f of 1 mm, the minimum object distance (MOD) is 1284 mm.
2×2 aspherical lenses 10′, with flattened and coated top 17. A black frame 70 surrounds the lenses 10′. Two buffles 80 with black filler are arranged over the image sensor 300 in a cross shape. They cross at the image sensor center. The thickness of the substrate 20 is less than 1 mm, e.g. equal to 0.7 mm.
In the embodiment of
The micro-optical component 100 of
As for the example of
The imaging device is therefore composed by three elements requiring high-precision assembly, e.g. using adhesive or gluing layers:
The present invention concerns also an imaging device comprising
In one embodiment, the pixels have a fill factor less than 70% and/or the pitch of pixels is larger than 3 μm, e.g. 5 μm.
The present invention concerns also a method for manufacturing the micro-optical component as described above, comprising the following steps:
In one embodiment, the method further comprises:
The present invention concerns also a method for manufacturing the imaging device as described above, comprising the following steps:
Number | Date | Country | Kind |
---|---|---|---|
18182654.6 | Jul 2018 | EP | regional |