JP2017063246A - Imaging device and photometric sensor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an imaging device which enhances metering low brightness limit and the resolution of a face detection image without sacrifice of color detection or light source detection performance.SOLUTION: In a photometric sensor 26, 26C is a pixel array, and a light receiver for photoelectric conversion is placed. The pixel array has a repetition arrangement unit where an arrangement pattern of 8 pixel period by blue transmission filters B1 and B2, green transmission filters G1-G4, a red transmission filter R, and a near-infrared filter IR is 1 period. The pixel arrangement ratio is such that the pixels of blue sensitivity are 2 pixels out of 8 pixels, the pixels of green sensitivity are 4 pixels out of 8 pixels, and the pixel of both red sensitivity and infrared sensitivity is 1 pixel out of 8 pixels.SELECTED DRAWING: Figure 2

Description

  The present invention relates to an imaging device and a photometric sensor.

  In general, a color sensor used for imaging or photometry has pixels having spectral sensitivities of three primary colors of B (blue), G (green), and R (red). FIG. 6A shows an arrangement example of pixels having spectral sensitivities of three primary colors called a so-called Bayer arrangement. Two G pixels are diagonally arranged in four 2 × 2 pixels, and B and R are one pixel at a time. Is to be placed.

  In such B, G, and R primary color sensors, there are cases where it is difficult to distinguish between the color of the light source and the color of the subject, and it is difficult to specify the type of the light source. As a countermeasure, a photometric sensor in which an IR (near infrared) pixel having a spectral sensitivity peak in a longer wavelength region than the R pixel is added may be used. A long wavelength region of about 700 nm or more is included in sunlight and the like, but is hardly included in fluorescent lamps and mercury lamps. Therefore, it is highly effective when used for determining the type of such a light source. A color sensor having a four-color configuration in which IR pixels are added to the three primary colors B, G, and R is described in Patent Document 1.

US Patent No. 6211521

In the color sensor having a four-color configuration described in Patent Document 1, IR pixels are provided for the purpose of eliminating an infrared cut filter provided externally to the sensor.
However, it is sufficient for the IR pixel in the color sensor of the four-color configuration used to determine the type of light source described above to have sensitivity in the wavelength range of about 650 nm to 800 nm as the spectral sensitivity range, and light in a longer wavelength range is incident than necessary. Even if it is harmful to photometry and color detection, an external infrared cut filter is used. In this case, the cut-off wavelength of the infrared cut filter is moderate from the middle of 700 nm to the latter half. FIG. 6B shows an example of the spectral sensitivity of the four colors and the cutoff wavelength of the infrared cut filter. FIG. 6C shows an example of a four-color Bayer array in which one of the G pixels in the three-color Bayer array shown in FIG.

  Actual color filters used in color sensors cannot achieve ideal spectral characteristics, and G pixels often have some sensitivity near 700 nm. Using an external infrared cut filter, which is often configured not to transmit more than 700nm, and using such a 4-color photometric sensor together, try to make the spectral sensitivity characteristics as close as possible. Then, the photometric sensor has to make the ratio of R considerably low when generating the photometric brightness value due to the sensitivity of the B or G pixel near 700 nm.

  In such a case, the photometry is performed using only B and G pixels in the four-color Bayer arrangement, and only about half of the light receiving area can be used for photometry, which is disadvantageous for low luminance. In addition, since G pixels are reduced as compared with the normal Bayer, there is a problem that the resolution of the luminance image is disadvantageous, such as performing face detection.

In order to solve the above problems, a photometric sensor according to the present invention is:
It has sensitivity to incident light from the visible light region to the near infrared light region,
A plurality of four types of pixels each having a spectral transmission characteristic with respect to a wavelength from the first to the fourth,
The spectral transmission characteristic of the fourth pixel is a transmission characteristic of a part of the spectral transmission characteristic of the third pixel,
The first pixel has a pixel arrangement ratio of 2 pixels in 8 pixels, the second pixel has 4 pixels in 8 pixels, and the third and fourth pixels have a pixel arrangement ratio of 1 pixel in 8 pixels. To do.

  According to the photometric sensor according to the present invention, the photometric image sensor having IR pixels for detecting the light source does not sacrifice the color detection or the light source detection performance, and the photometric low luminance limit and the resolution of the face detection image are reduced. Improvement can be expected.

Cross section of camera and interchangeable lens Diagram showing configuration example of photometric sensor The figure showing the example of composition of a signal processing means Camera control block diagram The figure showing the example of pixel arrangement of the metering sensor A diagram showing a conventional pixel arrangement of a photometric sensor and an example of spectral characteristics

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. FIG. 1 is a cross-sectional view of a camera, an interchangeable lens, and a flash device according to an embodiment of the present invention.

[Example 1]
FIG. 1 is a sectional view mainly showing the arrangement of optical members, sensors, etc. in a camera embodying the present invention.
In this figure, a configuration of a so-called single-lens reflex camera capable of exchanging lenses is shown. In FIG. 1, 1 is a camera body and 2 is an interchangeable lens.

  In the camera body 1, 10 is a mechanical shutter, 11 is an optical low-pass filter, 12 is a first infrared cut filter having a cutoff wavelength of about 680 nm to 700 nm, and 13 is a storage type photoelectric conversion element in an area such as a CMOS or CCD. It is an imaging device.

  14 is a semi-transparent main mirror, 15 is a first reflecting mirror, and both the main mirror 14 and the first reflecting mirror 15 jump up to the top during photographing. 16 is a paraxial imaging plane conjugate with the imaging element surface 13 by the first reflecting mirror 15, 17 is a second reflecting mirror, 18 is a stop having two openings, 19 is a secondary imaging lens, 20 Is a focus detection sensor. The focus detection sensor 20 is composed of, for example, an area-storing photoelectric conversion element using a CMOS process, and has a light receiving pixel portion that is divided into a plurality of portions corresponding to the two openings of the diaphragm 18. Further, in addition to the light receiving pixel portion, a signal accumulation portion, a signal processing peripheral circuit, and the like are formed as an integrated circuit on the same chip.

  The configuration from the first reflecting mirror 15 to the focus detection sensor 20 enables focus detection by an image shift method at each part in the photographing screen. 21 is a focusing plate having diffusivity, 22 is a pentaprism, 23 is an eyepiece lens, 24 is a photometric lens, 25 is a second infrared cut filter having a cutoff wavelength in the middle of 700 nm to the latter half, and 26 is a subject's object. This is a photometric sensor for obtaining information about luminance. The focus plate 21, the pentaprism 22 and the eyepiece lens 23 constitute a finder optical system. Of the light beam reflected by the main mirror 13 and diffused by the focus plate 21, a part outside the optical axis is incident on the photometric sensor 26.

Reference numeral 27 denotes a mount portion for attaching the photographing lens, 28 denotes a contact portion for performing information communication with the photographing lens, and 29 denotes a connection portion to which the flash device is attached.
In the interchangeable lens 2, reference numerals 30a to 30e denote optical lenses constituting the photographing lens, reference numeral 31 denotes a diaphragm, reference numeral 32 denotes a contact part for performing information communication with the camera body, and reference numeral 33 denotes a mount part to be attached to the camera.

The photometric sensor 26 will be described with reference to FIG.
The photometric sensor 26 is composed of, for example, an area-storing photoelectric conversion element using a CMOS process, and FIG. 2A is a block diagram illustrating an internal configuration example of the photometric sensor 26. In the photometric sensor 26, reference numeral 26A denotes a function setting circuit for setting functions such as an operation clock, accumulation control, and AD conversion control inside the sensor in accordance with data transmitted from a signal processing means 41 described later. Reference numeral 26B denotes an operation clock generation circuit inside the sensor.

  Reference numeral 26C denotes a pixel array, for example, where a light receiving unit for photoelectric conversion of tens of thousands to hundreds of thousands of pixels is arranged. In the pixel array 26C, an array pattern of 8 pixel periods by blue transmission filters B1 and B2, green transmission filters G1 to G4, red transmission filter R, and near-infrared transmission filter IR as illustrated in FIG. Thus, color information can be obtained together with the luminance information of the incident image. An accumulation control and pixel scanning circuit 26D controls the accumulation of the pixel array 26C and the pixel scanning at the time of reading. Reference numeral 26E denotes a read control circuit for sequentially reading analog signals accumulated in the respective pixels of the pixel array.

  The analog signal accumulated by each pixel output from the readout control circuit 26E is input to the 26F AD conversion circuit and converted into digital data. Reference numeral 26G denotes an AD conversion gain control circuit that adjusts the conversion gain of the AD conversion circuit 26F. The signal data stored in each pixel converted to digital data is output to the output circuit 26H and output to the signal processing means 41 described later. In the output circuit 26H, parallel-serial conversion, conversion to a differential signal, and the like are performed as necessary.

FIG. 3 shows a configuration example of the signal processing means 41.
An output signal for each pixel of the photometric sensor 26 is stored in the memory unit 43 through the input circuit 42. In the input circuit 42, serial-parallel conversion is performed as necessary, correction of the characteristics of the light receiving optical system of the photometric sensor 26, dark level correction, and the like are performed. The output signal for each pixel of the photometric sensor 26 stored in the memory unit 43 is input to the block integration unit 44. The block integration unit 44 calculates an addition integral value for each color of the output signal for each pixel of the photometric sensor 26 for each block.

  As an example, assuming that the total number of effective pixels of the photometric sensor 26 is 96000 pixels of horizontal 400 pixels × vertical 240 pixels, 960 pixels of horizontal 40 pixels × vertical 24 pixels are unit blocks, and the total effective pixel area is 10 × 10 × 100. Divide into blocks. Since the pixel array of the photometric sensor 26 has an array pattern with a period of 8 pixels illustrated in FIG. 2B, addition integration of pixel signals for each of the eight colors included in each divided block is performed. The added integral value for each block is output to the photometric value generation unit 45 and the light source information generation unit 46.

The photometric value generation unit 45 calculates the photometric value Y (i) based on the added integral value for each of the eight colors for each block.
Y (i) = Kyb × ΣB (i) + Kyg × ΣG (i) + Kyr × R (i)
However, i = 1 to 100, Kyb × 2 + Kyg × 4 + Kyr = 1, ΣB (i) = B1 (i) + B2 (i), ΣG (i) = G1 (i) + G2 (i) + G3 (i) + G4 (i )

The light source information generation unit 46 calculates light source color information C (i) based on the addition integral value for each of the eight colors for each block.
C (i) = {Kcb1 × ΣB (i) + Kcg1 × ΣG (i) + Kcr1 × R (i) + Kcir1 × IR (i)} ÷ {Kcb2 × ΣB (i) + Kcg2 × ΣG (i) + Kcr2 × R (i ) + Kcir2 × IR (i)}
However, i = 1 to 100, ΣB (i) = B1 (i) + B2 (i), ΣG (i) = G1 (i) + G2 (i) + G3 (i) + G4 (i)

  In the IR correction processing unit 47, B, G, R having a spectral sensitivity center in the visible light region with respect to each pixel signal output of the array pattern of the 8-pixel cycle illustrated in FIG. 2B stored in the memory unit 43. The pixel signal output is corrected with the IR pixel output. This is to reduce the influence of the cutoff characteristic of the second infrared cut filter 25 that is closer to a longer wavelength so that the output sensitivity of the IR pixel can be guaranteed.

B1s (j) = B1 (j) −Ksb × IR (j)
B2s (j) = B2 (j) −Ksb × IR (j)
G1s (j) = G1 (j) −Ksg × IR (j)
G2s (j) = G2 (j) −Ksg × IR (j)
G3s (j) = G3 (j) −Ksg × IR (j)
G4s (j) = G4 (j) −Ksg × IR (j)
Rs (j) = R (j) −Ksr × IR (j)
However, j = 1 to 12000

  Note that 12000 is a value obtained by dividing all effective pixels 96000 by 8 of the pixel arrangement period. The correction coefficients Ksb, Ksg, and Ksr may be appropriately determined as necessary, and may be set to 0 if the influence of the long wavelength near the near infrared is slight.

  The B, G, and R pixel signal outputs after the correction calculation are output to the interpolation processing unit 48, and a demosaic process is performed to interpolate the pixel values at all pixel positions from the pixel values of the mosaic arrangement pattern for each color. The three primary color information for all effective pixel regions of horizontal 400 pixels × vertical 240 pixels obtained by demosaic processing is assumed to be b (m, n), g (m, n), and r (m, n). Here, m = 1 to 400 and n = 1 to 240.

  The demosaiced signals b (m, n), g (m, n), and r (m, n) are output to the conversion processing unit 49 and subjected to matrix conversion using predetermined coefficients (M11 to M33). The luminance image information Br (m, n) and biaxial color difference image information Cx (m, n) and Cy (m, n) are converted and output.

  Each calculation information based on the output signal of the photometric sensor 26 calculated by the signal processing means 41 is used for camera control in the configuration shown in FIG. Luminance image information Br (m, n) output from the signal processing means 41 and biaxial color difference image information Cx (m, n) and Cy (m, n) are input to the subject position information detection means 61. The subject position information detection means 61 detects a person's face from the luminance image information by a template matching method, and outputs information related to the detected face position and size. In addition, the subject position information detecting means 61 outputs the movement information of the subject position based on the color information using the color difference image information.

  Reference numeral 62 denotes an exposure calculation unit which receives the photometric value Y (i) calculated by the signal processing unit 41, the face detection information detected by the subject position information detection unit 61, and the movement information of the subject position. The exposure calculation means 62 determines the main subject area based on the face detection information and the subject position movement information, and increases the weighting of the main subject area with respect to the photometric value Y (i), thereby providing optimum exposure control information for photographing. (Shutter speed, aperture value, shooting sensitivity) are calculated. When the exposure control information is calculated, it is output to the exposure control means 63, and on the basis of this, the mechanical shutter 10 is driven, the aperture 31 is driven, and the sensitivity of the image sensor 13 is set.

  The light source color information C (i) calculated by the signal processing means 41 is input to the focus detection means 64. Further, the focus detection means 64 also receives the output signal of the focus detection sensor 20, the face detection information detected by the subject position information detection means 61, and the subject position movement information. The focus detection unit 64 determines a subject area to be focused in accordance with the face detection information and the movement information of the subject position, and obtains image shift information of the corresponding area from the focus detection sensor 20. The focus lens drive amount for bringing the subject into focus is calculated from the image shift information. At this time, the detection error due to the light source color is corrected by the light source color information C (i). The calculated driving amount of the focus lens is output to the focus adjustment control means 65, and the focus adjustment of the interchangeable lens 2 is performed.

  In the array pattern of the 8-pixel cycle of the photometric sensor described in the present embodiment, the G (green) pixel having a higher ratio at the time of photometric value generation compared to the photometric sensor of the conventional array pattern shown in FIG. Since the pixel ratio of the B (blue) pixel can be increased, the low luminance limit is advantageous, and the G (green) pixel is arranged in the same manner as the normal Bayer in FIG. There is an effect that a sufficient information resolution is maintained. This is the end of the description of the first embodiment.

[Example 2]
In the first embodiment, the photometric sensor is an array pattern having an 8-pixel cycle shown in FIG. 2B. However, the array pattern is not limited to this, and the 8-pixel cycle shown in FIG. 5A is used. The same effect can be obtained by the arrangement pattern and the arrangement pattern having a period of 16 pixels shown in FIG.

  The present invention can be applied to an apparatus for taking a still image by combining a flash device with a digital camera, a film camera, or the like.

1 camera body, 2 interchangeable lens, 11 first infrared cut filter,
12 image sensor, 20 focus detection sensor, 21 focus plate,
25 Second infrared cut filter, 26 Photometric sensor, 41 Signal processing means

Claims (3)

It has sensitivity to incident light from the visible light region to the near infrared light region,
A plurality of four types of pixels each having a spectral transmission characteristic with respect to a wavelength from the first to the fourth,
The spectral transmission characteristic of the fourth pixel is a transmission characteristic of a part of the spectral transmission characteristic of the third pixel,
The first pixel has a pixel arrangement ratio of 2 pixels in 8 pixels, the second pixel has 4 pixels in 8 pixels, and the third pixel and the fourth pixel both have 1 pixel in 8 pixels. Features photometric sensor. The first pixel has a main sensitivity in a blue region of visible light, the second pixel has a main sensitivity in a green region of visible light, and the third pixel has a main sensitivity in a red region of visible light. The photometric sensor according to claim 1, wherein the fourth pixel has a main sensitivity in a long wavelength region of the spectral sensitivity characteristics of the third pixel. 3. The photometric sensor according to claim 1, wherein the fourth pixel has sensitivity mainly on a longer wavelength side than 650 nm. 4.