JP7207883B2 - Optical low-pass filter and imager - Google Patents

Description

The present invention relates to an optical low-pass filter, and more particularly to a multi-point separation type optical low-pass filter used in imaging devices.

Imaging devices that use two-dimensional imaging elements such as CCD sensors and C-MOS sensors use optical low-pass filters (hereinafter also referred to as optical elements) to prevent the occurrence of false colors and moire fringes in images obtained by imaging. be done. The optical low-pass filter limits image information of high frequencies equal to or higher than the Nyquist frequency by controlling the point image distribution of the imaged light.

In Patent Document 1, 6 to 8 layers of birefringent optical elements (birefringent layers) are laminated to form a Gaussian point image distribution, thereby improving high-frequency resolution and suppressing false colors and moire in oblique directions. A suppressing optical low pass filter is disclosed. Japanese Patent Application Laid-Open No. 2002-200000 discloses setting a light beam separation width for four-point separation or two-point separation by an optical low-pass filter in consideration of the effects of aberration and diffraction limit of a photographing lens.

JP 2015-213306 A Japanese Patent No. 5212044

However, when the distance between the separated point images (light beam separation width) approaches the diffraction limit due to the reduction of the pixel pitch of the image sensor and the increase in the number of distributed point images, the point images interfere with each other. This changes the shape of the point spread distribution. Although Patent Document 1 discloses conditions for forming a point spread distribution, it does not refer to the phase difference generated in each birefringent element or the influence of the separation axis orientation of each birefringent element. Although Patent Document 2 refers to the influence of the spread of the point images due to the diffraction limit, it does not refer to the asymmetry of the point image distribution due to the interference between the point images.

The present invention provides an optical low-pass filter having a highly symmetrical beam splitting effect.

An optical low-pass filter as one aspect of the present invention has a plurality of optically anisotropic layers each separating an incident light beam into a plurality of light beams. The optical low-pass filter has a plurality of optically anisotropic layers, each of which is a first layer to an N-th layer ( 4 ≤ N ≤ 5) in order from the light incident side, and the first to (N−1)-th layers The total value of the beam separation width is Da [μm], and among the beams separated by the first to (N−1)th layers, the beam having the maximum phase difference and the beam having the minimum phase difference with respect to the incident beam When the distance between is Ds [μm],
Ds≤0.50 Da
and the angle formed by the light beam splitting directions of all two adjacent optically anisotropic layers among the first to Nth layers is 120° or more and 150° or less .

Note that an imaging apparatus including the above optical low-pass filter and an imaging device that photoelectrically converts an optical image formed by light passing through the optical low-pass filter also constitutes another aspect of the present invention.

According to the present invention, it is possible to realize an optical low-pass filter having a highly symmetrical light beam splitting effect.

1 is a diagram showing the configuration of an optical low-pass filter that is a representative embodiment of the present invention; FIG. FIG. 4 is a diagram for explaining point image separation by a birefringent optical element; The figure explaining the interference effect between point images. FIG. 4 is a diagram for explaining the point spread distribution and phase difference distribution of the optical low-pass filter of the example; FIG. 5 is a diagram for explaining point spread distribution and phase difference distribution of an optical low-pass filter as a typical comparative example; 1 is a point spread diagram of an optical low-pass filter that is Embodiment 1 of the present invention. FIG. 4 is a point spread diagram of an optical low-pass filter that is Comparative Example 1. FIG. 4 is a diagram showing spatial frequency characteristics of Example 1. FIG. FIG. 10 is a diagram showing spatial frequency characteristics of Comparative Example 1; FIG. 4 is a point spread diagram of an optical low-pass filter that is Embodiment 2 of the present invention. FIG. 10 is a diagram showing spatial frequency characteristics of Example 2; FIG. 4 is a point spread diagram of an optical low-pass filter that is Embodiment 3 of the present invention. FIG. 10 is a diagram showing spatial frequency characteristics of Example 3; FIG. 4 is a diagram showing the configuration of an optical low-pass filter that is Embodiment 4 of the present invention. FIG. 10 is a point spread diagram of the optical low-pass filter of Example 4; FIG. 11 is a diagram showing spatial frequency characteristics of Example 4; FIG. 5 is a diagram showing the configuration of an optical low-pass filter that is Embodiment 5 of the present invention. FIG. 11 is a point spread diagram of the optical low-pass filter of Example 5; FIG. 11 is a diagram showing spatial frequency characteristics of Example 5; The figure which shows the imaging device which is Example 6 of this invention.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 shows the configuration of a multi-layer (multipoint separation type) optical low-pass filter 10, which is a representative embodiment of the present invention. The optical low-pass filter 10 has a structure in which four layers of birefringent optical elements (birefringent plates) 1 to 4 as optical anisotropic elements (optical anisotropic layers) are laminated. However, the number of optically anisotropic layers is not limited to four, and the effect of the present invention can be obtained even with five layers .

In the following description, the axis extending in the long-side direction (horizontal direction) of the optical low-pass filter 10 is defined as the x-axis, and its direction is defined as the x-direction. The axis extending in the short side direction (vertical direction) of the optical low-pass filter 10 is defined as the y-axis, and the direction is defined as the y-direction. Further, a plane along the x-axis and the y-axis, that is, the xy plane is called an element plane of the optical low-pass filter 10 . The axis perpendicular to the element surface is called an optical axis, and the direction in which the optical axis extends is called the optical axis direction (z direction). Four layers of birefringent optical elements 1 to 4 are stacked in the optical axis direction.

The optical axis direction is the direction in which light enters and passes through the optical low-pass filter 10 . Although the thickness of each birefringent optical element in the optical axis direction is shown to be thicker than the actual thickness for explanation purposes, the actual thickness is about several hundred micrometers.

Each of the birefringent optical elements 1 to 4 is made of uniaxial crystal (anisotropic material) such as quartz or lithium niobate. As shown in FIG. 2A, when each birefringent optical element is viewed from the optical axis direction, the optical axis is oriented at 90° to the y direction. Also, as shown in FIG. 2B, when each birefringent optical element is viewed from the y direction, its optical axis is inclined by an angle θ (90°−θ with respect to the element surface) with respect to the optical axis direction. ing. In the following description, θ is called the tilt angle. Although θ is generally set to an angle around 45°, it is set in the range of 10° or more and 80° or less depending on the material and the desired phase difference.

Also, it is generally known that a light beam incident on a birefringent optical element having a parallel plate shape with an oblique optical axis is split into two light beams, an ordinary ray and an extraordinary ray. That is, as shown in FIG. 2(c), a light ray 6 incident on the birefringent optical element from the optical axis direction is an ordinary ray 7 that is transmitted straight, and an ordinary ray 7 that is transmitted in the direction shown in the figure. and the extraordinary ray 8 . In the following description, the direction in which the ordinary ray 7 and the extraordinary ray 8 are separated is called the ray separation direction.

The ordinary ray 7 becomes polarized light with an electric field oscillating in a direction perpendicular to the beam splitting direction, and the extraordinary ray 8 becomes polarized light with an electric field oscillating in a direction parallel to the beam splitting direction. That is, the polarization directions of the ordinary ray 7 and the extraordinary ray 8 are orthogonal.

The ray separation width w, which is the distance between the ordinary ray 7 and the extraordinary ray 8 when they are emitted from the birefringent optical element, is the magnitude nΔ of refractive index anisotropy unique to the material of the birefringent optical element. , the tilt angle θ of the optical axis, and the thickness d of the birefringent optical element. If the material and the inclination angle θ are determined, the beam separation width w is proportional to the thickness d.

By stacking such birefringent optical elements in the optical axis direction (stacking direction) so that the light beam separation direction is 45° or 135°, the number of layers N (4 ≤ N ≤ 5) is 2. A point spread distribution (ray distribution) divided into N-th power is formed. The optical low-pass filter 10 can be made to function by geometrically optically designing the spread of the point spread in consideration of the light beam separation direction, the light beam separation width, and the number of layers of the birefringent optical elements 1 to 4 .

However, when the distance between multiple point images (light rays) is close to the spread of the point image distribution itself, wave optics interference occurs and the point image distribution obtained by geometrical optics may not be obtained.

As the simplest example, as shown in FIGS. 3A and 3B, consider two point images separated by a distance D and having parallel polarization directions among the point images separated into a plurality of points. The dotted line curve in the figure shows the original intensity distribution (point spread distribution) of each point image, and the solid line curve shows the point spread distribution obtained as a synthesis of two point images. White and black of the point images represent phase differences, and in FIG. 3A, the upper and lower dotted lines also indicate that the phase difference between the two point images is shifted by (m+1)λ/2. In FIG. 3(b), the phase difference between the two point images is zero or an integer multiple of the wavelength. In terms of geometrical optics, the point image distribution is uniform regardless of the phase difference given to the two point images. However, the point images actually have a finite spread due to diffraction and aberration, and interference occurs according to the phase difference between the point image distributions.

Specifically, in FIG. 3A, a drop occurs between point images due to cancellation, whereas in the example of FIG. 3B, a smooth point image distribution is obtained due to reinforcement. As a result, there is a difference between the point spread functions obtained by synthesizing the two point images between FIG. 3(a) and FIG. 3(b). In this case, the apparent beam separation width in consideration of interference is smaller in FIG. 3B than in FIG. 3A.

In general, the polarization relationship between the point images immediately after being separated by the birefringent optical element is orthogonal to each other, as indicated by the double-headed arrows in FIG. 2(d). A sum of intensity distributions without consideration of the phase may be used. If the separation width of each point image is sufficiently larger than the spread due to diffraction or aberration of the point image, the influence of interference is considered to be minor.

However, in a multi-layer, ie multi-point separation type optical low-pass filter, the distance between point images becomes close in the process of point image separation, so the influence of interference is likely to appear. Especially in a multi-point separation type optical low-pass filter that separates point images into 8 points or more, more significantly 16 points or more, the influence between the point images becomes complicated, so we considered the interference due to the phase difference between the point images. It is necessary to select the configuration of the optical low-pass filter so as to obtain an appropriate point spread distribution.

To explain the effect of interference in a multi-point separation type optical low-pass filter, consider an optical low-pass filter that performs 16-point separation with a four-layer configuration, as in FIG. FIG. 4 shows the separation process of point images when four layers of birefringent optical elements are stacked in the optical axis direction. The birefringent optical element is hereinafter referred to as LPF. In addition, in FIG. 4, the LPFs of the first to fourth layers arranged in order from the light incident side are denoted as LPF1, LPF2, LPF3, and LPF4, respectively.

FIGS. 4A to 4D show the state of point image separation when light rays pass through LPF1, LPF2, LPF3, and LPF4 in this order. Solid circles represent point images of ordinary rays, and dotted circles represent point images of extraordinary rays. An arrow pointing from one point image to another point image indicates the direction of beam separation. As shown in FIG. 4, light beams incident on the optical low-pass filter (incident light beams) are split in the respective light beam separation directions at angles of 225°, 0°, 135°, and 270° with the x-axis in each of LPF1 to LPF4. separated into As a result, the incident light beam is separated four times in total, and 16 point images are finally formed as shown in FIG. 4(d).

4(a) to 4(d), beside each LPF, the phase difference of each point image is indicated by shading as FIGS. 4(e) to 4(h). Here, it is assumed that the phase difference Δ between the ordinary ray and the extraordinary ray generated in each LPF is the same.

As shown in FIG. 4(e), the incident ray passes through the LPF 1, so that the point image of the extraordinary ray is given a phase difference Δ with respect to the ordinary ray. Then, when each light beam passes through LPF 2, the phase difference distribution (0 to 2Δ) shown in FIG. 4(f) is obtained. distribution (0 to 3Δ) is obtained. Finally, each ray passes through the LPF 4, forming a point image distribution of the ordinary ray and the extraordinary ray as shown in FIG. 4(h). . Therefore, interference between point images corresponding to the phase difference distribution occurs in the point image distribution of 8 points at the time of transmission through the LPF 3, and the point images are superimposed to form a point image corresponding to 16 point separation.

Focusing on the phase difference distribution in the LPF 3 where the influence of interference should be considered, the point image with the minimum phase difference (0) and the point image with the maximum phase difference (3Δ) are adjacent to each other near the center. Three point images with phase difference Δ and phase difference 2Δ are symmetrically arranged. The phase difference distribution between the eight point images of the ordinary ray and the eight point images of the extraordinary ray in the LPF 4 shown in FIG. Due to the added jump, the two do not have the same interference distribution.

Next, FIG. 5 shows an example (comparative example) of an optical low-pass filter that performs 16-point separation with a four-layer structure similar to that of FIG. 4, but with a different structure. In FIG. 5, the LPFs of the first to fourth layers are denoted as LPF1', LPF2', LPF3', and LPF4', respectively. The beam splitting directions at LPF1' to LPF4' shown in FIG. 5 are different from the beam splitting directions at LPF1 to LPF4 shown in FIG. As shown in FIGS. 5(a) to 5(d), the incident light rays entering the optical low-pass filters form angles of 45°, 0°, 315°, and 90° with the x-axis in LPF1′ to LPF4′, respectively. is separated in each ray separation direction. As a result, the incident light beam is separated four times in total, and 16 point images are finally formed as shown in FIG. 5(d).

In terms of geometrical optics, the point spread distribution obtained by the LPF4' of FIG. 5 is the same as the point spread distribution obtained by the LPF4 of FIG. Similar to FIG. 4, the phase difference of each point image is shown as FIGS. 5(e) to 5(h) next to each LPF in FIG. The phase difference distribution obtained by LPF3' of FIG. 5 is different from the phase difference distribution obtained by LPF3 of FIG. The phase difference of the leftmost point image is 0, which is the minimum. In addition, three point images with a phase difference Δ and three point images with a phase difference of 2Δ are formed, and the rightmost point image has the maximum phase difference of 3Δ. Thus, a phase difference distribution is obtained in which the phase difference increases from left to right in the figure. As can be seen by comparing FIG. 4(g) and FIG. 5(g), the phase difference distribution obtained by LPF3′ in FIG. high quality).

In order to explain the symmetry of the phase difference distribution of the point image, the configuration of LPF1 to LPF4 in FIG. 4 and the configuration of LPF1' to LPF4' in FIG. 5 are represented by vectors. FIG. 4(i) shows the overlapping of LPF1 to LPF4 with the direction of the vector as the ray separation direction and the length of the vector as the ray separation width, and FIG. shown in

Double-line arrows indicate synthetic vectors obtained by adding vectors in LPF1 to LPF3 and LPF1' to LPF3', which are layers in which the phase difference should be considered. The direction of the composite vector can be considered as the gradient direction of the phase difference distribution of the entire point image, and the length can be considered as the amount of deviation of the phase difference distribution. In FIG. 4(i), there is little deviation in the direction of the vector, and as a result the resultant vector cancels out and becomes short. It can be considered that this corresponds to the substantially uniform phase difference distribution formed by the LPF 4 shown in FIG. 4(h).

On the other hand, in the comparative example shown in FIG. 5(i), since the phase difference distribution is biased, the vectors are largely biased in the x direction, and as a result, the combined vector extends long in the x direction. Comparing the composite vector shown in FIG. 5(i) with the phase difference distribution obtained by the LPF 4′ shown in FIG. It can be seen that it corresponds to the presence of Therefore, the configuration of FIG. 4 has higher symmetry of the phase difference distribution than the configuration of FIG. It can be considered that the high symmetry of the phase difference distribution facilitates homogenization of the point spread distribution by wavelength averaging, which will be described later.

Further, as shown in FIGS. 4(h) and 5(h), the length of the synthesized vector is the position of the point image with the minimum phase difference and the phase difference is equal to the distance (also called ray separation width) Ds between the position of the maximum point image. This is because, with the point image having the minimum phase difference as the starting point, the arrival position of the point image having the maximum phase difference coincides with the end point of the composite vector. Therefore, the distance Ds can be used as an evaluation value for evaluating the symmetry of the phase difference in the point spread. In order to obtain a uniform point spread, it is desirable to design the azimuths of the optical axes of LPF1 to LPF4 and the beam separation width so that the distance Ds is sufficiently small.

Specifically, in an optical low-pass filter composed of N layers (N≧3) of LPFs, the light beam separation width of the LPFs from the first layer to the (N-1)th layer arranged in order from the light incident side Let Da be the total value. Let Ds [μm] be the distance between the light beam with the maximum phase difference and the light beam with the minimum phase difference among the light beams separated by the LPF from the first layer to the (N−1)th layer. At this time, it is desirable that the beam separation width Ds satisfies the following conditional expression (1) with respect to Da.

Ds ≤ 0.5 Da (1)
By satisfying this condition, it is possible to obtain an optical low-pass filter that has high azimuth symmetry even when the interference effect between point images is taken into consideration, and that is closer to the geometrical optics design value.

Furthermore, as a practical guideline for using this optical low-pass filter in the optical system of an imaging device, Ds is the diffraction spread of the d-line (587 nm) at F5.6, which is a width of 4 μm (1.22λdF=4.010 μm). If it is smaller, it is considered that the influence of interference between point images is significant. Considering manufacturing errors and the like, when Ds is 4.2 μm or less (when the condition Ds≦4.2 μm is satisfied), it is particularly desirable to satisfy conditional expression (1). In fact, as the pixel pitch of the image pickup device continues to shrink, when trying to obtain a low-pass filter effect with a smaller light beam separation width and a brighter F number, even if Ds is about 2 μm, a point image It is thought that the influence on the deformation of the point spread distribution due to the interference between them increases.

An imaging device using an optical low-pass filter generally operates in a broad wavelength band such as a visible light band. Since the phase difference of each LPF has wavelength dispersion, the phase difference Δ changes depending on the wavelength, and the distribution of point images also changes accordingly. However, in an actual captured image obtained by the imaging device, the point image distribution is averaged by the sensitivity of the image sensor and the bandwidth of the color filter, resulting in a point image distribution with high symmetry.

FIG. 6 shows, as an example, the point spread distribution when the wavelength is averaged from 460 nm to 640 nm in the configuration shown in Table 1. In FIG. As a result of the mutual interference of the intensity distributions due to the spread of the point images, the point image distributions are symmetrical even when the wavelength averaging is considered. FIG. 7 shows the point spread distribution in the configuration shown in Table 2 as a comparative example. In this comparative example, the asymmetry of the point spread distribution remains even when the wavelength averaging is considered.

The difference between the point spread distributions in FIGS. 6 and 7 is due to the phase difference distribution between the point spreads. The phase difference distribution of the point images in FIG. 7 is highly biased, and asymmetry remains even when changes due to wavelength dispersion are considered. On the other hand, in the point image distribution of FIG. 6, the phase difference distribution of the point image is highly symmetrical, and a uniform point image can be obtained by averaging the wavelengths. By satisfying the conditional expression (1), it is possible to distribute the light beam splitting directions in each LPF so as not to be biased, thereby obtaining a uniform point spread distribution.
Furthermore, it is more preferable to satisfy the following conditional expression (1a).

Ds≦0.4 Da (1a)
In order for the wavelengths to be sufficiently averaged in the wavelength range in which the optical low-pass filter of this embodiment is used (use wavelength range), it is desirable that the chromatic dispersion of the phase difference of each LPF is large to some extent. If the wavelength dispersion is small, the wavelength may not be sufficiently averaged when the point spread distribution is biased when a light beam of a single wavelength is incident.

Specifically, the shortest wavelength in the operating wavelength range of the optical low-pass filter is λmin, and the longest wavelength is λmax. In at least one of the LPFs from the 1st layer to the (N-1)th layer, when the phase differences occurring with respect to the shortest wavelength λmin and the longest wavelength λmax are respectively δmin and δmax, the following conditional expression: It is desirable to satisfy (2).

0.5≦δmin/λmin−δmax/λmax (2)
Further, it is desirable to satisfy the following conditional expression (3), where δs is the phase difference generated in at least one LPF among the N layers of LPFs.

0.5≦δs/λmin−δs/λmax (3)
By satisfying conditional expression (2) or (3), a sufficient wavelength averaging effect can be obtained.

The usable wavelength range is determined by the specifications of the imaging device in which the optical low-pass filter is used. is set in consideration of If it is a visible light wavelength range, a wavelength range with an arbitrary wavelength width corresponding to a specific color range of approximately 100 nm within the range of approximately 400 nm to 700 nm is selected. As long as it is a wavelength range set by the imaging device, it may be an infrared wavelength range or an ultraviolet wavelength range other than visible light wavelengths.

Considering the wavelength range to be considered when the optical low-pass filter functions as the wavelength range to be used, it is desirable to satisfy the conditional expression (2) in order to obtain the effect of this embodiment. Furthermore, it is more desirable that not only one of the LPFs from the 1st layer to the (N-1)th layer but also the other LPFs satisfy the conditional expression (2).

Furthermore, it is more preferable to satisfy the following conditional expressions (2a) and (2b) in order.

0.5≦δmin/λmin−δmax/λmax≦3.0 (2a)
0.6≦δmin/λmin−δmax/λmax≦2.0 (2b)
Moreover, it is more preferable to satisfy the following conditional expressions (3a) and (3b) in order.

0.5≦δs/λmin−δs/λmax≦10.0 (3a)
0.55≦δs/λmin−δs/λmax≦8.5 (3b)
In addition, in the description so far, the configuration for performing point image separation without using a wave plate was shown, but a wave plate (λ/4 plate) is arranged between two adjacent LPF layers, and these LPFs mutually It may be configured to perform point image separation in an orthogonal direction. In this case, the wave plate does not act to separate the point image and is used for polarization state conversion, and does not substantially contribute to the composite vector. Similar considerations apply to composite vectors.

However, increasing the number of layers that do not contribute to point image separation in a multi-point separation optical low-pass filter in which interference between point images is likely to occur leads to an increase in material cost and thickness. Therefore, a first LPF (first optically anisotropic layer, first anisotropic optical element) that performs point image separation in the x direction (first direction) or the y direction (second direction) and , and a second LPF (second optically anisotropic layer, second anisotropic optical element) that performs point image separation in a direction of 45° or 135° with respect to the x or y direction. good. In this case, in order to effectively suppress false colors that tend to occur in diagonal components, it is desirable that the light beam separation width of the first LPF is smaller than the light beam separation width of the second LPF.

Further, in the configuration example shown in FIG. 4, four LPF1 to LPF4 having the same beam separation width are used, but the beam separation width of each LPF may be different from each other. As shown in conditional expression (1), the sufficiently small beam separation width Ds means that the deviation of the combined vector in the beam separation direction is small regardless of the azimuth. Therefore, even if the light beam separation widths of LPF1 to LPF4 are different from each other, the same effect as the configuration of FIG. 4 can be obtained.

Even if interference between point images is taken into account, the basic design value of the optical low-pass filter may be calculated from geometrical optics conditions. More preferably, a point spread distribution considering the effects of diffraction spread and interference is obtained in advance, and the strength of the LPF characteristic is adjusted based on the results. At this time, satisfying the conditional expressions (1) and (2) and calculating the design value from the geometrical optics conditions are not contradictory conditions. In the geometrical optics design value, there are two degrees of freedom for the LPF beam splitting direction (for example, +45° azimuth separation and 225° azimuth separation are geometrically equivalent). The design value and conditional expression (1) can be treated independently. That is, after performing geometrical optics design, the ray separation width of each LPF can be set so as to satisfy the conditional expression (1) within the above degrees of freedom.

The same applies to the phase difference related to conditional expression (2). For example, the method of determining the separation width of the LPF includes two parameters, the tilt angle θ and the thickness d. Since these parameters are independent of each other, the phase difference can be arbitrarily adjusted with the same beam separation width. Therefore, the beam separation width and the phase difference can be controlled independently of each other. For example, the inclination angle .theta. of about 45.degree. However, if the thickness d is sufficiently thick, the tilt angle θ can be set to a value larger than 45°. Even if the separation width is the same, the latter can increase the phase difference.

Regarding the conditional expression (1), when the LPFs of N layers are configured to have the same beam separation width, as a simpler configuration method, the angle formed by the beam separation directions of the two LPFs adjacent to each other is may be set to about 135°. This makes it possible to realize an optical low-pass filter that satisfies conditional expression (1). However, since the setting of the beam separation width and the beam separation direction is determined by the design conditions, it is not necessary to strictly set the above angle to 135°. Select from the range.

So far, the conditions for obtaining a point spread distribution close to the design value that has been uniformed by averaging the wavelengths have mainly been described when considering the interference between the point spreads. On the other hand, in order to obtain a highly symmetric point image distribution at a single wavelength, not only the phase difference between the point images described above, but also the phase jump that occurs in the point image separation and the (N-1)th layer It is necessary to consider how to superimpose the formed point spread distributions. Furthermore, when the conditional expression (1) is satisfied, since the condition of the optical axis orientation of the LPF is determined to some extent, the light beam separation direction of the LPF of the Nth layer, which is the final layer, and the phase difference caused by the LPF of each layer Δ should be used to adjust the symmetry of the point spread distribution.

When the light beam separation widths of the LPFs of the N layers are all equal, the optical anisotropy of the second to (N−1) layers with respect to the wavelength λa longer than the shortest wavelength λmin and shorter than the longest wavelength λmax It is desirable that the phase difference Δnd of the LPF of at least one (at least one other than the first layer and the Nth layer) of the magnetic layers satisfies the following equation (4). m is a natural number.
Δnd=m×λa (4)
On the other hand, when the above-described first and second LPFs are alternately stacked and their beam separation widths are different from each other, at least one layer of LPF excluding the first layer and the Nth layer with respect to the wavelength λa It is desirable that the phase difference Δnd of satisfies the following equation (5).
Δnd=(m+0.5)/2×λa (5)
Thus, even with a point spread at a specific wavelength, a highly symmetrical point spread can be obtained by controlling the phase difference of the LPF of each layer and the light beam separation direction.

Various materials can be used as the material of the LPF of each layer. For example, transparent uniaxially anisotropic crystal materials such as quartz, sapphire, and lithium niobate can be used, and these materials may be used in combination in an N-layer LPF. In addition to the crystal material, the LPF may be configured using liquid crystal, or the LPF may be configured using structural birefringence.

Specific examples (numerical examples) of the optical low-pass filter described above will be described below.

An optical low-pass filter of Example 1 (numerical example 1) of the present invention will be described. Table 1 shows the configuration of the optical low-pass filter of this embodiment. As shown in FIG. 1, the optical low-pass filter 10 of this embodiment is constructed by stacking LPFs 1 to 4 made of quartz in this order along the optical axis direction. As shown in Table 1, the beam splitting directions of LPFs 1-4 are 225°, 0°, 135°, and 270° with respect to the x-direction, respectively. Also, the light beam separation widths of LPFs 1 to 4 are all the same, 2.10 μm.

The phase difference distribution of point images in the LPF 3 of the third layer is as shown in the third row of FIG. The distance Ds between the point image with the maximum phase difference and the point image with the minimum phase difference is 0.870 μm, which corresponds to 0.138 Da, so that conditional expression (1) is satisfied.

In addition, the difference between the values obtained by dividing the phase difference generated in the LPF of each layer by 460 nm and 640 nm at the shortest wavelength of 460 nm and the longest wavelength of 640 nm in the wavelength range used is 1.118, so conditional expression (2) is satisfied. .

FIG. 6 shows the point image intensity distribution of the optical low-pass filter of this embodiment. This point image intensity distribution shows the distribution at the focal position when a light flux corresponding to F2.0 is made incident on the optical low-pass filter. The wavelength is the result of averaging in the range from 460 nm to 640 nm. Although different from the 16-point separation, it can be seen that a point spread distribution with less unevenness is obtained for each direction.

8 shows the spatial frequency characteristic (MTF) of the point spread intensity distribution of FIG. The four lines in FIG. 8 indicate the MTF in horizontal (x), vertical (y), +45° (P), and -45° (N) directions, respectively. The MTF of this example exhibits a symmetrical distribution in all directions.

Since the optical low-pass filter of this embodiment has a Nyquist frequency Fn of about 110/mm, it is possible to improve the sense of resolution in the horizontal and vertical directions while appropriately suppressing color moire occurring in the diagonal direction.

On the other hand, Table 2 shows the configuration of the optical low-pass filter of Comparative Example 1. In this comparative example, the beam separation width and phase difference are the same as in Example 1, but the beam separation direction is different. The distance Ds is 5.070 μm, which corresponds to 0.805 Da, and does not satisfy conditional expression (1).

The point image intensity distribution of Comparative Example 1 is shown in FIG. 7, and the MTF graph is shown in FIG. The point spread distribution of Comparative Example 1 is partially biased near the center. This results in higher MTFs in the horizontal (x) and -45° (N) directions. This is because the distribution shape of the point spread projected in the horizontal direction (x) and the -45° direction (N) is concentrated in the center compared to the design value. In addition, the drop in MTF in the high frequency range of 200 lp/mm or more is not sufficient overall. As a result, color moiré may increase in a specific direction, or high-frequency folding moiré may occur.

According to this embodiment, the point spread of the multi-point separation type optical low-pass filter can be made uniform, and an optical low-pass filter with a highly symmetrical point spread and close to the design value can be obtained.

An optical low-pass filter of Example 2 (numerical example 2) of the present invention will be described. Table 3 shows the configuration of the optical low-pass filter of this embodiment. In this embodiment, the light separation width of LPFs 1 to 4 is as small as 1.50 μm as compared with the first embodiment. The light beam splitting directions of LPFs 1 to 4 are 45°, 180°, 315°, and 90° with respect to the x direction, respectively. The beam separation widths of LPFs 1 to 4 are all equal, 1.75 μm.

The distance Ds is 0.725 μm, which corresponds to 0.138 Da, and thus satisfies conditional expression (1). In addition, the difference between the values obtained by dividing the phase difference generated in the LPF of each layer by 460 nm and 640 nm at the shortest wavelength of 460 nm and the longest wavelength of 640 nm in the wavelength range used is 0.798, so conditional expression (2) is satisfied. .

FIG. 10 shows the point image intensity distribution of the optical low-pass filter of this embodiment. This point image intensity distribution shows the distribution at the focal position when a light flux corresponding to F2.0 is made incident on the optical low-pass filter. Since the point images are closer to each other than in Example 1, the point image intensity distribution has a shape close to a uniform Gaussian distribution due to their interference.

11 shows the spatial frequency characteristic (MTF) of the point image of FIG. The meaning of MTF indicated by four lines in FIG. 11 is the same as in the first embodiment. The MTF of this example exhibits a symmetrical distribution in all directions. In the optical low-pass filter of the present embodiment, the Nyquist frequency Fn is about 150/mm, so that it is possible to improve the sense of resolution in the horizontal and vertical directions while appropriately suppressing color moire that occurs in the diagonal direction. .

An optical low-pass filter of Example 3 (numerical example 3) of the present invention will be described. Table 4 shows the configuration of the optical low-pass filter of this embodiment. In this embodiment, the light separation width of LPFs 1 to 4 is as large as 3.03 μm compared to the first embodiment. Also, the light beam splitting directions of LPFs 1 to 4 are 135°, 0°, 225°, and 90° with respect to the x direction. Further, the LPFs 1 to 4 all have the same beam separation width of 3.03 μm.

The distance Ds is 1.255 μm, which corresponds to 0.138 Da, and thus satisfies conditional expression (1). In addition, the difference between the values obtained by dividing the phase difference generated in the LPF of each layer by 460 nm and 640 nm at the shortest wavelength of 460 nm and the longest wavelength of 640 nm in the wavelength range used is 1.613, so conditional expression (2) is satisfied. .

FIG. 12 shows the point image intensity distribution of the optical low-pass filter of this embodiment. This point image intensity distribution shows the distribution at the focal position when a light beam corresponding to F4.0 is made incident on the optical low-pass filter. Although the point spread distribution is slightly biased, a highly symmetrical and uniform distribution is obtained.

13 shows the spatial frequency characteristic (MTF) of the point image of FIG. 12. In FIG. The meaning of MTF indicated by four lines in FIG. 13 is the same as in the first embodiment. The MTF of this embodiment exhibits a symmetrical distribution in all directions, and the MTF is uniformly suppressed up to the high frequency range. In the optical low-pass filter of the present embodiment, the Nyquist frequency Fn is about 80/mm, so that it is possible to improve the sense of resolution in the horizontal and vertical directions while appropriately suppressing color moire that occurs in the diagonal direction. .

An optical low-pass filter of Example 4 (numerical example 4) of the present invention will be described. FIG. 14 and Table 5 show the configuration of the optical low-pass filter of this example. As shown in FIG. 14, the optical low-pass filter 40 of this embodiment is constructed by stacking five layers of LPFs 31 to 35 each made of quartz in this order in the optical axis direction. The beam splitting directions of the LPFs 31 to 35 are 180°, 45°, 90°, 315°, and 90° with respect to the x direction, respectively, and the beam splitting directions of the LPF 31 and the LPF 35 are the same. Also, the beam separation width (1.32 μm) of LPF 31 and LPF 35 is half the beam separation width (2.64 μm) of the other LPFs 32-34.

Since the low-pass filter 40 of this embodiment has five layers of LPFs 31 to 35, it is originally configured to separate the incident light beam into 32 points. However, since eight of them overlap at the same position, the optical low-pass filter is divided into 24 points. Therefore, although the geometrical optics point spread is not exactly the same in the x direction and the y direction, the beam separation width is the same in the x direction and the y direction.

In this example, Ds is 1.714 μm, which corresponds to 0.186 Da, so that conditional expression (1) is satisfied. Also. The value of conditional expression (2) is 1.404 for LPFs 31, 32 and 34, and 0.701 for LPFs 33 and 35, thus satisfying conditional expression (2).

FIG. 15 shows the point image intensity distribution of the optical low-pass filter 40 of this embodiment. This point image intensity distribution shows the distribution at the focal position when a light flux corresponding to F2.0 is made incident on the optical low-pass filter. Although the point spread distribution is slightly biased, a highly symmetrical and uniform distribution is obtained.

FIG. 16 shows the spatial frequency characteristic (MTF) of the point image of FIG. The meaning of MTF indicated by four lines in the figure is the same as in the first embodiment. As described above, the point spread in this embodiment does not have the same point spread in the two directions in terms of geometrical optics because the number of beam splitting times differs between the x direction and the y direction. However, in a region strongly affected by the interference between point images, the MTFs in the four directions exhibit a relatively symmetrical distribution. Therefore, according to this embodiment, an optical low-pass filter having a highly symmetrical and uniform point spread distribution can be obtained.

An optical low-pass filter of Example 5 (numerical example 5) of the present invention will be described. FIG. 17 and Table 6 show the configuration of the optical low-pass filter 50 of this embodiment. The optical low-pass filter 50 of this embodiment includes second and fourth LPFs 42 and 44 that perform point image separation with a beam separation width Db in the y and x directions, respectively, and a beam separation width Dc (Dc> Db) has first and third LPFs 41 and 43 for performing point image separation. The first to fourth LPFs 41, 42, 43, 44 are alternately stacked in this order in the optical axis direction.

The light beam splitting directions of the first to fourth LPFs 41, 42, 43 and 44 are 45°, 180°, 315° and 90° with respect to the x direction, respectively. The material of the second and fourth LPFs 42, 44 is crystal, and the material of the first and third LPFs 41, 43 is lithium niobate ( _NbLiO3 ). The beam separation width of the second and fourth LPFs 42 and 44 is 1.73 μm, and the beam separation width of the first and third LPFs 41 and 43 is 3.48 μm.

Since the distance Ds is 3.191 μm, which corresponds to 0.367 Da, conditional expression (1) is satisfied. Further, the difference in the value obtained by dividing the phase difference generated in at least one layer of LPF by 460 nm and 640 nm at the shortest wavelength of 460 nm and the longest wavelength of 640 nm in the used wavelength range is about 1.038 for the second and fourth LPFs 42 and 44. Thus, the conditional expression (2) is satisfied.

FIG. 18 shows the point image intensity distribution of the optical low-pass filter 50 of this embodiment. This point image intensity distribution shows the distribution at the focal position when a light beam corresponding to F4.0 is made incident on the optical low-pass filter. In the optical low-pass filter of this embodiment, by making the horizontal and vertical ray separation widths and the diagonal ray separation widths different, it is possible to achieve both suppression of diagonal color moire and resolution in the horizontal and vertical directions. It is configured. A uniform distribution of the MTF is obtained in each of the horizontal and vertical directions and the two diagonal directions.

FIG. 19 shows the spatial frequency characteristic (MTF) of the point image of FIG. Of the MTFs indicated by the four lines in the figure, the MTFs for horizontal and vertical components show a tendency to gently decrease over high frequencies. The MTF in the diagonal direction drops around 140 lp/mm, and the configuration is such that horizontal and vertical resolution can be maintained up to high frequencies while suppressing false colors occurring in this vicinity. Folding occurs in the diagonal component at 220 lp/mm, but it is suppressed to a low level compared to the conventional four-point separation or the like.

Embodiment 6 FIG. 20 shows the configuration of an imaging apparatus 100 that is Embodiment 6 of the present invention. The imaging apparatus 100 has a configuration in which the optical low-pass filter 81 of Examples 1 to 5 described above is arranged between the imaging optical system 83 and the imaging device 82 . An optical image is formed on the imaging device 82 by the light flux that has passed through the imaging optical system 83 and further passed through the optical low-pass filter 81 . The imaging device 82 photoelectrically converts the optical image to generate an image signal.

The imaging element 82 has a Bayer array of R, G, and B pixels, and color moire is likely to occur particularly in a diagonal (diagonal) spatial high-frequency region. In addition, there is concern about luminance moiré caused by aliasing in a high-frequency region. To solve these problems, the optical low-pass filters 81 of Examples 1 to 5 are used to provide good color moire and luminance moire suppression effects in each of the diagonal, horizontal, and vertical directions. can be done. The configuration of the optical low-pass filter 81 may be appropriately selected according to the pixel pitch of the imaging device 82 and the configuration of the imaging device 100 .

Each embodiment described above is merely a representative example, and various modifications and changes can be made to each embodiment in carrying out the present invention.

For example, different materials, such as quartz and lithium niobate, may be combined for each layer of LPF. Also, the number of layers may be six or more depending on the tolerance of thickness and cost. As the number of layers increases, the overall point spread is smoothed, but the effect of interference also increases. Therefore, by satisfying the above-mentioned conditional expressions, an optical low-pass filter with a highly symmetrical and uniform point spread can be obtained. can get.

In addition, although the strength of the optical low-pass filter is proportional to the beam separation width, it is also possible to control the cutoff spatial frequency by scaling the entire beam separation width. You can shift the overall cutoff frequency including horizontal and diagonal directions. If you want to control the horizontal direction and the oblique direction independently, you can control the oblique direction separation and the horizontal separation of the point image separately. desirable.

Table 7 shows the values of each conditional expression in each example. For Examples 4 and 5, numerical values are described separately for each material of the birefringent optical element. In addition, regarding conditional expression (3), the respective values of the phase difference δs for each wavelength are shown. As shown in Table 7, conditional expression (3) is satisfied in any of the examples.

1, 2, 3, 4 Optically anisotropic layer 6 Incident ray 7 Ordinary ray 8 Extraordinary ray 10, 40, 50 Optical low-pass filter

Claims (14)

An optical low pass filter having a plurality of optically anisotropic layers each separating an incident light beam into a plurality of light beams,
Each of the plurality of optically anisotropic layers is the first layer to the Nth layer ( 4 ≤ N ≤ 5) in order from the light incident side, and the total width of light beam separation by the first layer to the (N-1)th layer The value is Da [μm], and the distance between the ray having the maximum phase difference and the ray having the minimum phase difference with respect to the incident light beam among the light beams separated by the first to (N−1)th layers is Ds [μm],
Ds≤0.50 Da
satisfies the following conditions,
An optical low-pass filter, wherein the angle formed by the light beam splitting directions of all two adjacent optically anisotropic layers among the first to N-th layers is 120° or more and 150° or less . λmin and λmax are the shortest and longest wavelengths in the wavelength range used, and the phase difference that occurs with respect to the shortest wavelength λmin and the longest wavelength λmax in at least one of the first to (N-1)-th layers is When δmin and δmax respectively,
0.5≦δmin/λmin−δmax/λmax
2. The optical low-pass filter according to claim 1, wherein the following condition is satisfied. When the shortest wavelength and the longest wavelength in the wavelength range used are λmin and λmax, respectively, and the phase difference occurring in at least one of the first to Nth layers is δs,
0.5≦δs/λmin−δs/λmax
3. The optical low-pass filter according to claim 1, wherein the following condition is satisfied. 4. The optical low-pass filter according to any one of claims 1 to 3 , wherein the first to N-th layers all have the same beam separation width. In at least one of the second layer to the (N-1)th layer, the phase difference generated for a wavelength λa [nm] longer than the shortest wavelength λmin and shorter than the longest wavelength in the wavelength range used is Δnd [nm] , where m is a natural number,
Δnd=m×λa
5. The optical low-pass filter according to claim 4 , wherein the following condition is satisfied. The first to Nth layers include first and second optically anisotropic layers, and the first optically anisotropic layer is in the first direction or in the second direction orthogonal to the first direction. direction, and wherein said two optically anisotropic layers separate said incident light beam in directions at 45° or 135° to said first or second direction. 4. The optical low-pass filter according to any one of 1 to 3 . 7. The optical low-pass filter according to claim 6 , wherein the ray separation width of said first optically anisotropic layer is smaller than the ray separation width of said second optically anisotropic layer. In the first to Nth layers, the first and second optically anisotropic layers having different light beam separation widths are alternately laminated,
In at least one layer excluding the first layer and the Nth layer, the phase difference that occurs with respect to the wavelength λa [nm] longer than the shortest wavelength λmin and shorter than the longest wavelength in the wavelength range used is Δnd [nm], a natural number is m,
Δnd=(m+0.5)/2×λa
8. The optical low-pass filter according to claim 6 , wherein the following condition is satisfied. Each of the first to Nth layers is composed of at least one of an optically anisotropic layer made of crystal and an optically anisotropic layer made of lithium niobate. 9. The optical low-pass filter according to any one of items 1 to 8 . Ds≤4.2
10. The optical low-pass filter according to any one of claims 1 to 9 , wherein the following condition is satisfied. Ds≤0.4 Da
11. The optical low-pass filter according to any one of claims 1 to 10 , wherein the following condition is satisfied. λmin and λmax are the shortest and longest wavelengths in the wavelength range used, and the phase difference that occurs with respect to the shortest wavelength λmin and the longest wavelength λmax in at least one of the first to (N-1)-th layers is When δmin and δmax respectively,
0.5≦δmin/λmin−δmax/λmax≦3.0
12. The optical low-pass filter according to any one of claims 1 to 11 , wherein the following condition is satisfied. When the shortest wavelength and the longest wavelength in the wavelength range used are λmin and λmax, respectively, and the phase difference occurring in at least one of the first to Nth layers is δs,
0.5≦δs/λmin−δs/λmax≦10.0
13. The optical low-pass filter according to any one of claims 1 to 12 , wherein the following condition is satisfied. an optical low-pass filter according to any one of claims 1 to 13 ;
and an imaging device that photoelectrically converts an optical image formed by light that has passed through the optical low-pass filter.

Images