JP4773767B2 - FOCUS DETECTION DEVICE, OPTICAL DEVICE, AND FOCUS DETECTION METHOD - Google Patents

Description

  The present invention relates to a focus detection apparatus and method for performing focus detection by a so-called phase difference detection method based on photoelectric conversion results of a pair of object images obtained from a pair of light reception units among a plurality of pairs of light reception units. .

  As a focus detection device used for an optical apparatus such as a single-lens reflex camera, there is one using a phase difference detection method. In this method, a pair of subject images formed by light beams that have passed through different pupil regions in the photographing lens are photoelectrically converted by a pair of sensor arrays, and the relative displacement (phase difference) of the output (image signal) is obtained. In this method, the defocus amount of the photographing lens is detected from the phase difference. In this method, the pair of sensor arrays extract only a luminance distribution in a specific area or a specific direction of the subject space, and therefore it is impossible to detect the defocus amount for a subject that does not have a luminance distribution in that area.

  Therefore, Patent Documents 1 and 2 prepare a plurality of pairs of sensor arrays and corresponding focus detection optical systems, extract luminance distributions in a plurality of areas and directions of a subject, and perform focus detection for more subjects. A method is disclosed that enables

  Incidentally, as disclosed in Patent Documents 1 and 2, when a plurality of focus detection results (defocus amounts) are obtained, it is necessary to select one focus detection result to be used for actual focus adjustment from the results. is there.

In the focus detection device disclosed in Patent Document 1, the reliability of the focus detection result is obtained from the defocus amount and the contrast of the subject, and the focus detection result is selected based on the reliability. Further, in the focus detection device disclosed in Patent Document 2, the focus detection speed is improved by selecting a sensor array for which the accumulation operation has been completed first from a plurality of pairs of sensor arrays.
Japanese Patent Laid-Open No. 11-84222 (paragraphs 0050 to 0071, FIGS. 12 to 16) JP-A-2-64516 (page 9, upper left column, line 2 to page 13, lower right column, line 8, FIG. 7)

  However, in the focus detection devices disclosed in Patent Documents 1 and 2, since the reliability of the focus detection result is determined only by the defocus amount and the contrast of the subject, the degree of coincidence between the pair of subject images is completely considered. Absent. If this degree of coincidence is not taken into account, when the ghost light enters only one of the pair of subject images and the subject image collapses, the focus detection result deviates from the actual focus state. There is a problem that the shifted focus detection result is selected.

  Further, the focus detection device disclosed in Patent Document 2 does not perform a process of comparing the focus detection results of a plurality of pairs of sensor arrays. For this reason, it is not possible to select a focus detection result with better accuracy.

  As described above, the conventional focus detection apparatus cannot select appropriate information with high accuracy from a plurality of focus detection results, and as a result, there is a problem that the focus detection accuracy is lowered.

  The present invention provides a focus detection device and a focus detection method capable of appropriately selecting a focus detection field for generating focus detection information used for focus adjustment from a plurality of focus detection fields and performing highly accurate focus detection. The purpose is to do.

A focus detection apparatus according to an aspect of the present invention includes a plurality of pairs of light receiving units corresponding to a plurality of focus detection fields, and a sensor that photoelectrically converts a pair of object images formed on each pair of light receiving units; Arithmetic means for generating focus detection information used for focus adjustment based on signals from a pair of light receiving units corresponding to the specific focus detection field among the plurality of focus detection fields; Then, the calculation means calculates the degree of correlation when the degree of coincidence between the pair of object images calculated for each focus detection field of view and the relative positional relationship between the pair of object images is shifted from the focused state by a predetermined number of light receiving pixels. The specific focus detection visual field is selected based on the evaluation value using the amount of change, the sharpness of the pair of object images, and the contrast ratio of the pair of object images as parameters .
The evaluation value is expressed by the following equation, where the degree of coincidence is U, the amount of change in the correlation amount is ΔV, the sharpness is SH, and the contrast ratio is PBD .
Evaluation value = U / (ΔV × SH × PBD)

A focus detection method according to another aspect of the present invention includes a plurality of pairs of light receiving units corresponding to a plurality of focus detection fields, and photoelectrically converts a pair of object images formed on each pair of light receiving units. A first step of obtaining a signal from the sensor; a second step of generating focus detection information used for focus adjustment based on signals from a pair of light receiving units corresponding to a specific focus detection field among a plurality of focus detection fields; Have In the second step, the degree of coincidence between the pair of object images calculated for each focus detection field and the correlation when the relative positional relationship between the pair of object images is shifted from the focused state by a predetermined number of light receiving pixels. The specific focus detection visual field is selected based on the evaluation values using the amount of change, the sharpness of the pair of object images, and the contrast ratio of the pair of object images as parameters .
The evaluation value is expressed by the following equation, where the degree of coincidence is U, the amount of change in the correlation amount is ΔV, the sharpness is SH, and the contrast ratio is PBD.
Evaluation value = U / (ΔV × SH × PBD)

According to the present invention, among the plurality of focus detection field, it is possible to select the focus detection field of view higher precision suitable focus detection information is obtained. Therefore, the focus detection accuracy can be improved. If focus adjustment control is performed using such focus detection information, higher focusing accuracy can be obtained.

  Embodiments of the present invention will be described below with reference to the drawings.

  A focus detection method that is Embodiment 1 of the present invention will be described with reference to FIGS. First, FIG. 1 shows the configuration of a camera as an optical apparatus equipped with the focus detection apparatus of this embodiment.

  The camera of the present embodiment is a single plate type digital color camera using an image pickup element which is a photoelectric conversion element such as a CCD sensor or a CMOS sensor. An image signal representing a moving image or a still image can be obtained by driving the image sensor continuously or once. The image sensor is an area sensor of a type that converts light received for each pixel into an electric signal, accumulates electric charge according to the amount of received light, and reads out the electric charge.

  In FIG. 1, 1 is a photographing optical system, and 2 is a main mirror that divides a subject light beam from the photographing optical system. A part of the main mirror 2 is a half mirror, which transmits a part of the subject luminous flux and reflects the remaining subject luminous flux upward. Reference numeral 3 reflects the subject luminous flux transmitted through the main mirror 2 downward. A focus detection device 20 that detects the focus state of the photographing optical system 1 is disposed below the sub mirror 3. The reflected light from the sub mirror 3 is guided to the focus detection device 20. Reference numeral 4 denotes an image sensor that receives a subject light beam formed by the photographing optical system 1 and converts it into an image signal.

  The image sensor 4 of the present embodiment is packaged and is a CMOS process compatible sensor (CMOS sensor) which is one of amplification type solid-state image sensors. One of the features of the CMOS sensor is that a MOS transistor in the area sensor unit and peripheral circuits such as an image sensor driving circuit, an AD conversion circuit, and an image processing circuit can be formed in the same process. As a result, the number of masks and process steps can be greatly reduced as compared with the CCD sensor. In addition, there is a feature that random access to an arbitrary pixel is possible. As a result, it is easy to read out the charge accumulation signal thinned out for display, and real-time display can be performed at a high display rate. The image sensor 4 uses this feature to perform a display image output operation and a high-definition image output operation.

  The focus detection device 20 performs focus detection by a phase difference detection method. In the phase difference detection method, the relative positional relationship (phase difference) between two subject images formed by light beams that have passed through a pair of different pupil regions of the photographic optical system 1 is detected, and this is used to determine the focus state of the photographic optical system 1. Used as information to represent. A technique for detecting a focus state of a photographing optical system based on a phase difference between two subject images obtained from a pair of different pupil regions of the photographing optical system is well known as disclosed in Japanese Patent Laid-Open No. 52-138924. Technology.

  The main mirror 2 is supported by a camera body (not shown) by a shaft portion 2a, and is rotatable with respect to the camera body. The sub mirror 3 is supported by a shaft 3 a on the holding member of the main mirror 2 and is rotatable with respect to the main mirror 2. Therefore, when the main mirror 2 rotates about the shaft portion 2a and the sub mirror 3 rotates about the shaft portion 3a, the main mirror 2 is at a position inclined by 45 degrees with respect to the optical axis of the photographing optical system 1. Is disposed at a position inclined downward by about 45 degrees (hereinafter, this state is referred to as a mirror-down state). Further, by storing both the main mirror 2 and the sub mirror 3 upward, the main mirror 2 and the sub mirror 3 are disposed in a state of being retracted from the optical path of the subject light beam (hereinafter, this state is referred to as a mirror-up state).

  In the mirror-down state, the subject luminous flux from the photographing optical system 1 is divided into two parts, an upper finder optical system and a lower focus detection device. On the other hand, in the mirror-up state, all the subject luminous flux from the photographing optical system 1 is guided to the image sensor 4.

  An optical low-pass filter 22 that limits the cut-off frequency of the photographic optical system 1 so that a spatial frequency component higher than necessary in the object image is not transmitted onto the image sensor 4 in the optical path from the photographic optical system 1 to the image sensor 4. Is provided. The optical low-pass filter 22 is also formed with an infrared cut filter.

  On the light incident surface side of the optical low-pass filter 22, a shutter unit 21 that restricts the exposure time of the subject light beam incident on the image sensor 4 is disposed. A front curtain 21a and a rear curtain 21b each composed of a plurality of shutter blades travel in the short side direction of the image sensor 4, and the shutter time is controlled by the travel interval between the front curtain 21a and the rear curtain 21b.

  Reference numeral 5 denotes a light shielding member (hereinafter referred to as a cover) that restricts a subject light beam reflected by the sub mirror 3. The cover 5 has an opening that transmits only the light beam necessary for focus detection out of the subject light beam. The aperture is disposed in the vicinity of the imaging plane of the photographing optical system 1, guides a necessary subject light beam to a focus detection optical system described below, and shields an unnecessary subject light beam. Since the upper surface of the cover 5 is exposed on the inner wall of the camera body, a light shielding line is provided so that the reflected light does not reach the image sensor 4.

  Reference numeral 6 denotes a light shielding sheet that further restricts the light beam that has passed through the opening of the cover 5, and 7 denotes a field lens that projects a pupil of a focus detection optical system, which will be described later, onto a pupil of the photographing optical system. Reference numeral 8 denotes a holding member for holding each component of the focus detection device 20, and 9 denotes a plate for attaching the holding member 8 to the camera body.

  Reference numeral 11 denotes an IR-CUT filter. In the IR-CUT filter 11, a multilayer deposited film that reflects infrared component light is formed on the glass surface.

  Reference numeral 12 denotes a folding mirror. In the folding mirror 12, an aluminum vapor deposition film is formed on the glass surface, and the mirror 12 reflects light having a wavelength of 400 to 800 nm with substantially the same reflectance. In order to house the focus detection device 20 in a limited space below the camera, the folding mirror 12 folds the subject light flux to reduce the size of the focus detection device 20.

  Reference numeral 13 denotes a stop of the re-imaging optical system. A pair of apertures is formed in the diaphragm 13 to limit the light beam incident on the pair of lens portions of the re-imaging lens 14.

Reference numeral 14 denotes a re-imaging lens. A pair of lenses is provided corresponding to the pair of apertures of the diaphragm 13, and light beams from different pupil regions of the photographing optical system 1 are imaged on the focus detection sensor 17, respectively.
Reference numeral 16 denotes a sensor holder that holds the focus detection sensor 17, and 15 denotes an abutting member that is interposed between the sensor holder 16 and the holding member 8 and constitutes an inclination adjustment mechanism of the focus detection sensor 17.

  The sensor holder 16 is a member that holds the focus detection sensor 17. The focus detection sensor 17 and the sensor holder are fixed with an instantaneous adhesive or a welding agent.

  Reference numeral 40 denotes an arithmetic circuit that calculates an evaluation value, which will be described later, based on a signal from the focus detection sensor 17 and generates focus detection information based on the evaluation value. The arithmetic circuit 40 calculates the driving direction and driving amount of the focus lens (not shown) included in the photographing optical system 1 to the in-focus position based on the focus detection information, and the focus is controlled through a lens control circuit (not shown). Control adjustment behavior. That is, the arithmetic circuit 40 also functions as a control circuit that performs focus adjustment control. However, an arithmetic circuit that performs focus detection calculation and a control circuit that performs focus adjustment control may be provided separately in the camera.

  Cover 5, light shielding sheet 6, field lens 7, holding member 8, plate 9, eccentric cam 10, IR-CUT 11, folding mirror 12, diaphragm 13, re-imaging lens 14, contact member 15, sensor holder 16, focus detection The sensor 17 and the arithmetic circuit 40 constitute the focus detection device 20.

  Here, the IR-CUT filter 11 is not disposed perpendicular to the optical axis after being folded by the folding mirror 12 but is tilted by a predetermined angle. This is to prevent the occurrence of a ghost caused by the light reflected by the light receiving surface of the focus detection sensor 17 being reflected by the surface of the IR-CUT filter 11 and reaching the light receiving surface of the focus detection sensor 17 again. . By tilting the IR-CUT filter 11 by a predetermined angle, the light returned from the focus detection sensor 17 is reflected by the surface of the IR-CUT filter 11 and reaches a place off the light receiving surface of the focus detection sensor 17. Not observed as a ghost.

  FIG. 2 shows an exploded view of the focus detection device 20 of this embodiment. The cover 5 has an opening that transmits only the light beam necessary for focus detection out of the subject light beam, and the opening is disposed in the vicinity of the imaging surface of the photographing optical system 1. As a result, only the subject luminous flux necessary for focus detection is guided to the focus detection optical system, and unnecessary subject luminous flux is shielded.

  The cover 5 is formed with two holes 5a. The holding member 8 fixes the cover 5 by inserting the protrusions 8e into these holes 5a. Since the upper surface of the cover 5 is exposed on the inner wall of the camera body, a light shielding line is provided so that the reflected light does not reach the image sensor 4.

  The light shielding sheet 6 further restricts the light beam transmitted through the opening of the cover 5. The field lens 7 projects the diaphragm 13 onto the pupil of the photographing optical system. The holding member 8 holds each component constituting the focus detection device 20.

  The plate 9 is used for attaching the holding member 8 to the camera body. The eccentric cam 10 is rotatably held by the shaft portion 8 a of the holding member 8.

  The holding member 8 includes the cover 5, the light shielding sheet 6, the field lens 7, the eccentric cam 10, the IR-CUT filter 11, the folding mirror 12, the diaphragm 13, the re-imaging lens 14, the contact member 15, the sensor holder 16, A focus detection sensor 17 and separators 18 and 19 are held.

  The holding member 8 is attached to the camera body via the plate 9. When the camera body and the holding member 8 are made of different materials, the linear expansion coefficient is often different. In this case, the expansion / contraction rate differs between the camera body and the holding member 8 due to temperature changes, and the holding member 8 may be distorted due to deformation of the camera body. Since each component of the focus detection optical system is held by the holding member 8, the distortion of the holding member 8 is directly connected to an error in the focus detection result. Therefore, in this embodiment, the holding member 8 is not directly fixed to the camera body, but is fixed via the plate 9 made of a material having a linear expansion coefficient close to that of the holding member 8. Thereby, when the linear expansion coefficient differs between the camera body and the plate 9, the deformation is absorbed by the plate 9, and the holding member 8 is not distorted.

  The plate 9 is preferably made of a material having a large Young's modulus in order to make it difficult to transmit the deformation of the camera body to the holding member 8. For this reason, what was manufactured by bending metal plates, such as stainless steel and iron, by press work is used.

  Furthermore, in order to maintain the strength of the plate 9, it is desirable that the shape be connected in an annular shape. For this reason, the plate 9 is formed in an annularly connected shape surrounding the holding member 8 while being L-shaped as a whole. As a result, the strength sufficient to absorb the deformation of the camera body can be maintained, and the distortion of the holding member 8 can be prevented.

  The plate 9 is formed with U-shaped bearing portions 9 a and 9 b having contact portions in the short side direction of the image sensor 4. The bearings 9a and 9b have a width in the front-rear direction of the camera (subject-image plane direction) set to a dimension that allows the shafts 8a and 8b to be engaged. In a state where the shaft portions 8a and 8b are engaged with the bearing portions 9a and 9b, the position of the holding member 8 in the front-rear direction of the camera with respect to the plate 9 is uniquely determined. The U-shaped hole of the bearing portion 9a is longer than the U-shaped hole of the bearing 9b. For this reason, when the holding member 8 is arranged at the position of the design value with respect to the plate 9, the shaft portion 8a and the bearing portion 9a do not contact in the short side direction of the image sensor 4, but the shaft portion 8b and the bearing portion 9b are It comes into contact. That is, the shaft portion 8b and the bearing portion 9b are in contact with each other in the front-rear direction of the camera and the short side direction of the image pickup device 4 to uniquely determine the position, and the shaft portion 8a and the bearing portion 9a are arranged in the short side direction of the image pickup device 4. Is only movable.

  The eccentric cam 10 held by the shaft portion 8 a of the holding member 8 is in contact with the plate 9 on the outer periphery thereof to determine the height position of the shaft portion 8 a with respect to the plate 9. Thereby, as described above, the position of the shaft portion 8 a that can move in the short side direction of the image sensor 4 is uniquely determined by the rotation angle of the eccentric cam 10.

  The holding member 8 is rotatable with respect to the plate 9 about the shaft portions 8a and 8b. Further, the holding member 8 can be rotated so that the optical axis of the photographing optical system 1 is the central axis by changing the height position of the shaft portion 8 a by the rotation of the eccentric cam 10.

  The focus detection device 20 can adjust the tilt in two directions by rotating the shafts 8 a and 8 b as the center axis and adjusting the height position of the shaft portion 8 a by rotating the eccentric cam 10. By adjusting the tilt in these two directions, so-called pupil adjustment that corrects the deviation between the optical axis of the focus detection optical system and the optical axis of the photographing optical system 1 can be performed. After the pupil adjustment is completed, an instantaneous adhesive or welding agent is poured into the contact surfaces of the holding member 8, the plate 9, and the eccentric cam 10, and these are fixed.

  If the cover 5 is attached in a state where the bearing portions 9a and 9b of the plate 9 and the shaft portions 8a and 8b of the holding member 8 are in contact, the cover 5 prevents the plate 9 from coming off. For this reason, the plate 9 cannot be detached from the holding member 8 even if it is not bonded.

  As described above, the IR-CUT filter 11 is formed by forming a multilayer deposited film that reflects infrared light on the glass surface. The focus detection sensor 17 is sensitive to light having a longer wavelength than visible light. On the other hand, since the optical low-pass filter 22 including an infrared cut filter is disposed on the light incident surface side of the image sensor 4, light having a wavelength longer than that of visible light is cut and does not reach the image sensor 4. . When the spectral characteristics of the light received by the focus detection sensor 17 and the image sensor 4 are different, the in-focus positions are changed. In this case, when the focus lens of the photographing optical system 1 is driven to the in-focus position obtained by the focus detection device 20, a phenomenon that the image sensor 4 is out of focus occurs. For this reason, the IR-CUT filter 11 has the same spectral characteristics of light reaching the image sensor 4 and the focus detection sensor 17 and plays the role of preventing the above-described problems.

  As described above, the folding mirror 12 has an aluminum vapor deposition film formed on the glass surface, and reflects light having a wavelength of 400 to 800 nm with substantially the same reflectance.

  A pair of apertures is formed in the diaphragm 13 of the re-imaging optical system to limit the light beam incident on the pair of lens portions of the re-imaging lens 14. The diaphragm 13 is projected onto the pupil of the photographing optical system 1 by the field lens 7. Light beams from a pair of different pupil regions in the pupil of the photographing optical system 1 are transmitted through the projected pair of apertures of the diaphragm 13. As the material of the diaphragm 13, it is preferable to use a metal plate such as phosphor bronze or a light shielding resin sheet. In the case of using a phosphor bronze plate, first, the die is punched out, and the punched phosphor bronze plate is black-plated to produce the aperture 13. In the case of the light shielding resin sheet, it can be manufactured by punching with a mold.

The re-imaging lens 14 includes a pair of lenses corresponding to the pair of apertures of the diaphragm 13, and images light beams from different pupil regions of the photographing optical system 1 on the focus detection sensor 17. The re-imaging lens 14 includes two positioning dowels, and the dowel engages a reference hole provided in the holding member 8 and a reference hole provided in the diaphragm 13, whereby the re-imaging lens 14 is held by the holding member. It is positioned with respect to 8 and the diaphragm 1 3. The position of the re-imaging lens 14 in the optical axis direction is also determined by contacting the holding member 8 with the stop 13 sandwiched by the receiving surface of the holding member 8.

The diaphragm 13 and the re-imaging lens 14 are fixed to the holding member 8 with an adhesive. The re-imaging lens 14 is manufactured by injection molding with a transparent resin. The transparent resin that is the material of the re-imaging lens 14 and the resin that is the material of the holding member 8 have different linear expansion coefficients and different expansion / contraction ratios due to temperature changes. For this reason, if the holding member 8 and the re-imaging lens 4 are firmly bonded, the re-imaging lens 14 may be distorted when a temperature change occurs. Since the deformation of the re-imaging lens 14 greatly affects the focus detection result, the focus detection accuracy may be lowered. For this reason, when the diaphragm 13 and the re-imaging lens 14 are bonded to the holding member 8, a soft adhesive having a Young's modulus of 10 kgf / mm ² or less is used to prevent the re-imaging lens 14 from being deformed under temperature. It is out.

Sensor holder 16 holds the focus detection sensor 17, the contact member 1 5 constitutes a tilt adjusting mechanism of the focus detection sensor 17 is interposed between the sensor holder 16 and the holding member 8. The contact member 15 is formed with an opening 15 a for transmitting the subject light flux and a pair of spherical portions 15 b for contacting the holding member 8. Since the opening 15a does not have a role of limiting the subject light flux, the opening 15a is formed as an opening larger than the effective diameter of the subject light flux. The spherical portion 15b is formed at both ends of the contact member 15, and has a shape obtained by cutting out a part of the spherical surface. That is, two spheres are provided at both ends of the abutting member 15 and abut against the holding member 8.

  The focus detection sensor 17 and the sensor holder 16 are fixed with an instantaneous adhesive or a welding agent. The sensor holder 16 has an opening 16a for transmitting the subject light flux. Since the opening 16a does not have a role of limiting the subject light flux, the opening 16a is formed as an opening larger than the effective diameter of the subject light flux.

  Next, a 5-axis adjustment mechanism of the re-imaging lens 14 and the focus detection sensor 16 will be described with reference to FIG. FIG. 3 shows an AA cross section in FIG. The re-imaging lens 14 includes two positioning dowels 14a, and the dowels 14a engage with positioning holes 8f provided in the holding member 8, whereby the position of the re-imaging lens 14 with respect to the holding member 8 is determined. Further, the re-imaging lens 14 is in contact with the holding member 8 with the stop 13 sandwiched by the re-imaging lens receiving surface of the holding member 8, so that the position in the optical axis direction is also determined.

The contact member 15 is formed with a pair of spherical portions 15b. The spherical surface portions 15b are provided at both ends of the contact member 15, and each has a shape obtained by cutting out a part of the spherical surface. That is, two spheres are provided at both ends of the abutting member 15 and abut against the holding member 8.
On the other hand, the holding member 8 is formed with a pair of spherical contact portions 8c having a predetermined radius of curvature. The curvature radius at this time is the same as the curvature radius of the spherical surface portion 15 b of the contact member 15. The contact member 15 can rotate in the directions of 30a and 30b with respect to the holding member 8 by sliding the contact portion 8c of the holding member 8 and the spherical surface portion 15b of the contact member 15. When rotating in the direction of 30a, the contact member 15 rotates about the center of curvature of the spherical surface portion 15b. On the other hand, when rotating in the direction of 30b, the abutting member 15 moves along the abutting surface 8c of the holding member 8, so that the center of curvature of the spherical surface of the abutting surface 8c rotates about the center of rotation.

The sensor holder 16 holds the focus detection sensor 17. The focus detection sensor 17 is a ceramic package. 1 7 b is a laminated type ceramic package, 1 7 c is a cover glass, the chip 1 7 a is placed in the ceramic package 1 7 b, the lid is covered with the cover glass 1 7 c, and the periphery is fixed with an adhesive And a package is comprised by sealing.

The focus detection sensor 17 and the sensor holder 16 are fixed by an instantaneous adhesive or the like at a portion not covered with the cover glass 1 7 c on the light incident surface side of the ceramic package 1 7 b. Reference numeral 23 denotes an adhesion portion. In order to secure the bonding location 23, the ceramic package 17b is set larger than the cover glass 17c. Thereby, since the adhesion | attachment location 23 moves away from the light beam transmission area | region of the cover glass 17c, it can prevent that the light beam transmission area | region of the cover glass 17c becomes dirty by the whitening etc. by an adhesive agent.

  The contact member 15 and the sensor holder 16 are in contact with each other, can rotate about the optical axis, and can move on a plane perpendicular to the optical axis.

The re-imaging lens 14 is fixed to the holding member 8. On the other hand, the focus detection sensor 17 is integrated with the sensor holder 16 and can rotate with respect to the contact member 15 about the optical axis and move on a plane perpendicular to the optical axis. Further, the focus detection sensor 17, the sensor holder 16, and the contact member 15 are integrated and can rotate in the directions of 30 a and 30 b with respect to the holding member 8.

  By providing the rotation mechanism as described above and a shift mechanism on the plane having the optical axis as a normal line, a five-axis adjustment mechanism for the re-imaging lens 14 and the focus detection sensor 17 can be realized. After the 5-axis adjustment is completed, an instantaneous adhesive, a mold welding agent, or the like is poured into the sliding surface, and the holding member 8, the contact member 15, and the sensor holder 16 are fixed.

In FIG. 2, reference numerals 18 and 19 denote separators that prevent mixing of the subject luminous flux in the central visual field and the subject luminous flux in the peripheral visual field. If the light beam transmitted through the central lens portion of the field lens 7 passes through the peripheral visual field opening of the stop 13 and reaches the focus detection sensor 17, it becomes ghost light and causes a detection error. Therefore, in the optical path between the field lens 7 and the IR-CUT filter 11, separators 18 and 19 are interposed in the gap between the central visual field light beam and the peripheral visual field light beam to prevent these light beams from being mixed. .
FIG. 4 shows the layout of the focus detection field of view on the shooting screen. Reference numeral 24 denotes a photographing range by the image sensor 4. There are 21 lines of focus detection fields (hereinafter referred to as focus detection lines) each having a line shape. Since a pair of subject images is formed on the light receiving surface of the focus detection sensor 17 for one line, 42 subject images are formed on the light receiving surface of the focus detection sensor 17 by the re-imaging lens 14. At the center of the shooting screen, there are two focus detection lines L1 and L2 that are long in the vertical direction (hereinafter referred to as vertical eyes), and two focus detection lines that are long in the horizontal direction (hereinafter referred to as horizontal eyes) L3 and L4. is there. The vertical eye has a correlation direction of the vertical direction (short-side direction of the image sensor 4), and detects the focus state of the photographing optical system 1 from the luminance distribution in the vertical direction. On the other hand, for the horizontal eye, the correlation direction is the horizontal direction (long side direction of the image sensor 4), and the focus state of the photographing optical system 1 is detected from the luminance distribution in the horizontal direction.

  The focus detection lines L1 and L2 arranged in two lines in the vertical direction are arranged so as to be shifted by a minute amount in the vertical direction. This shift amount is half the pitch of the pixels arranged in the correlation direction. Variation in detection is reduced by calculating the detection result by taking into account both of the two focus detection results obtained from the focus detection lines L1 and L2 that are arranged with a shift of half the pixel pitch. As for the focus detection lines L3 and L4 arranged in two horizontal directions, the detection variation is reduced by shifting the focus detection lines L3 and L4 by a half amount of the pixel pitch similarly to the focus detection lines L1 and L2.

  Further, at the center of the screen, there is one focus detection line L19 having a long base line length (this is shown in FIG. 6) and performing focus detection using a light flux of F2.8 in the horizontal direction. The focus detection line L19 has a long base line length compared to the focus detection lines L3 and L4, and thus the amount of image movement on the light receiving surface of the focus detection sensor 17 is large. For this reason, detection with higher accuracy than the focus detection lines L3 and L4 can be realized. However, since the focus detection line L19 uses F2.8 light flux, it operates only when a bright lens of F2.8 or higher is attached. Further, since the amount of image movement is large, the ability to detect a large defocus is inferior to the focus detection lines L3 and L4. As described above, when a focus detection line for F2.8 (hereinafter referred to as F2.8) L19 is provided, when a bright lens of F2.8 or higher, which requires high detection accuracy, is attached, it responds accordingly. In addition, highly accurate detection can be realized.

  In the focus detection apparatus 20 of the present embodiment, the luminance distribution in both the vertical and horizontal directions can be detected by arranging the focus detection lines in the vertical and horizontal cross form at the center of the screen. By obtaining the luminance distribution in both the vertical and horizontal directions, it is not necessary to forcibly perform focus detection on a subject with little luminance distribution, so that detection variation can be reduced. As a result, it is possible to increase the types of subjects for which focus detection is possible and to improve detection accuracy.

  Next, focus detection lines arranged in the vertical direction of the screen will be described. Horizontal eyes L5 and L11 are arranged at the upper part of the screen, and the focus state of the photographing optical system 1 is detected from the luminance distribution in the horizontal direction of the subject image located at the upper part of the screen. Further, horizontal eyes L6 and L12 are arranged at the lower part, and the focus state of the photographing optical system 1 is detected from the luminance distribution in the horizontal direction of the subject image located at the lower part of the screen. Further, the F2.8 horizontal eye L20 is arranged at the same position of the focus detection line L5, and the F2.8 horizontal eye L21 is arranged at the same position as the horizontal eye L6. Thereby, when a bright lens of F2.8 or higher that requires high detection accuracy is attached, high-precision focus detection can be realized.

  Focus detection is performed so that the focus detection lines L7, L9 are in contact with the left ends of the focus detection lines L3, L4, L5, L6, L11, L12, and the right ends of the focus detection lines L3, L4, L5, L6, L11, L12. Lines L8 and L10 are arranged side by side in the vertical direction. The focus detection lines L7, L9, L14, and L16 are vertical eyes, and detect the focus state of the photographing optical system 1 from the luminance distribution in the vertical direction.

  Focus detection lines L13 and L15 are arranged on the left side of the focus detection lines L7 and L9, and focus detection lines L14 and L16 are arranged outside the focus detection lines L8 and 10 in the vertical direction. The focus detection lines L13, L14, L15, and L16 are also vertical eyes, and detect the focus state of the photographing optical system 1 from the luminance distribution in the vertical direction.

  Focus detection lines L17 and L18 are arranged on the outermost side in the horizontal direction of the screen. The focus detection lines L17 and L18 have the center in the vertical direction on the optical axis, and detect the focus detection state of the photographing optical system 1 from the vertical luminance distribution of the subject image positioned in the horizontal direction of the photographing screen.

  FIG. 5 shows the re-imaging lens 14 as viewed from the light exit surface side. A plurality of pairs of lens portions for re-imaging a pair of subject images are formed on the light exit surface side. Each lens part is a spherical lens, and has a convex shape in the light exit direction.

The lens portions 14-1A and 14-1B are for re-imaging the vertical focus detection light beam at the center of the screen , and correspond to the focus detection lines L1, L2, L7, L8, L9, and L10 in FIG. is doing. The light beams of the focus detection lines L1, L2, L7, L8, L9, and L10 are re-imaged by the lens portions 14-1A and 14-1B, respectively, and are paired in the vertical direction on the light receiving surface of the focus detection sensor 17. Form a subject image.

Lens portion 14-2a, 14-2b is for reimaging focus detection light fluxes courses in the middle of the screen, and corresponds to a focus detection line L3, L4, L5, L6, L11, L12. The luminous fluxes corresponding to the focus detection lines L3, L4, L5, L6, L11, and L12 are re-imaged by the lens portions 14-2A and 14-2B, respectively, and are arranged in the lateral direction on the light receiving surface of the focus detection sensor 17. A pair of subject images are formed .
The lens units 14-3A and 14-3B are for re-imaging the F2.8 focus detection light beam and correspond to the focus detection lines L19, L20, and L21. Since the base length of the F2.8 eye is long, the distance between the lens portions 14-3A and 14-3B is longer than that of the lens portions 14-1A and 14-1B, and the lens portions 14-3A and 14-3B are in the lens portion 14. -1A and 14-1B are arranged outside.

  The lens portions 14-4A and 14-4B correspond to the focus detection lines L13 and L15, and the lens portions 14-5A and 14-5B correspond to the focus detection lines L14 and L16. Further, the lens portions 14-6A and 14-6B arranged outside these lens portions correspond to the focus detection line L17, and the lens portions 14-7A and 14-7B correspond to the focus detection line L18. The light beams of these focus detection lines are also re-imaged by the corresponding lens units, and form a pair of subject images arranged in the vertical direction on the light receiving surface of the focus detection sensor 17.

  FIG. 6 shows the chip (light receiving surface) of the focus detection sensor 17 as viewed from the light incident surface side. A plurality of pairs of light receiving units are arranged corresponding to the respective imaging positions of the pair of subject images re-imaged by the re-imaging lens 14 described in FIG.

  The light flux on the focus detection line L1 is re-imaged on the light receiving portions L1A and L1B in FIG. 6 by the lens portions 14-1A and 14-1B in FIG. Further, the light flux of the focus detection line L2 is re-imaged on the light receiving portions L2A and L2B by the lens portions 14-1A and 14-1B. Similarly, the light fluxes of the focus detection lines L7, L8, L9, and L10 are re-imaged on the light receiving portions L7A, L8A, L9A, and L10A by the lens portion 14-1A, and the light receiving portions L7B, L8B by the lens portion 14-1B. , L9B, L10B.

The same applies to the horizontal eye. The light beams of the focus detection lines L3, L4 , L5, L6, L11, and L12 are re-imaged on the light receiving portions L3A, L4A , L5A, L6A, L11A, and L12A by the lens portion 14-2A. Further, the image is re-imaged on the light receiving portions L3B, L4B , L5B, L6B, L11B, and L12B by the lens portion 14-2B.

  The same applies to the F2.8 side. The light fluxes of the focus detection lines L19, L20, and L21 are re-imaged on the light receiving portions L19A, L20A, and L21A by the lens portion 14-3A, and re-imaged on the light receiving portions L19B, L20B, and L21B by the lens portion 14-3B. Is done.

  The same applies to the vertical eyes around the screen. The light fluxes on the focus detection lines L13 and L15 are re-imaged on the light receiving portions L13A and L15A by the lens portion 14-4A, and re-imaged on the light receiving portions L13B and L15B by the lens portion 14-4B. The light fluxes on the focus detection lines L14 and L16 are re-imaged on the light receiving portions L14A and L16A by the lens portion 14-5A, and re-imaged on the light receiving portions L14B and L16B by the lens portion 14-5B.

  The light beam of the focus detection line L17 is re-imaged on the light receiving unit L17A by the lens unit 14-6A, and re-imaged on the light receiving unit L17B by the lens unit 14-6B.

  The light beam of the focus detection line L18 is re-imaged on the light receiving unit L18A by the lens unit 14-7A, and re-imaged on the light receiving unit L18B by the lens unit 14-7B.

  Each light receiving unit of the focus detection sensor 17 is configured by a pixel array, and an object image formed on the focus detection line by the imaging optical system 1 is between the output signal waveforms of each pair of light receiving units. In response to this, a relatively laterally shifted state is observed. The shift direction of the output signal waveform is reversed between the front pin and the rear pin, and the principle of focus detection is to detect this phase difference (shift amount) including the direction using a method such as correlation calculation.

  7A and 7B are diagrams showing output signal waveforms of the focus detection sensor 17. The horizontal axis represents the pixel arrangement, and the vertical axis represents the output value. FIG. 7A shows an output signal waveform when the object image is not in focus, and FIG. 7B shows an output signal waveform when the object image is in focus. By detecting a phase difference using a known method using correlation calculation, for example, a method disclosed in Japanese Patent Publication No. 5-88445, a defocus amount including information on the defocus direction can be obtained. If the defocus amount as the obtained focus detection result (focus adjustment information) is converted into an amount (and direction) to drive the focus lens of the imaging optical system 1, automatic focus adjustment is possible. According to this phase difference detection method, the amount and direction in which the focus lens should be driven can be known before the focus lens is driven. Therefore, normally, only one lens drive to the in-focus position is required, and extremely fast focus adjustment is possible. Is possible.

  As described with reference to FIGS. 4, 5 and 6, at the center of the screen, three focus detection results by the vertical eye (L1, L2), the horizontal eye (L3, L4), and the F2.8 eye (L19) are displayed. Obtainable. However, it is very important how to derive the final detection result, that is, the focus detection result actually used for focus adjustment, from these three detection results existing in the same visual field region at the center of the screen. If the guidance is poor, there is a risk that detection variations will increase rather than improving detection accuracy.

Therefore, in the focus detection device 20 of the present embodiment, for each focus detection line, 1) the degree of coincidence between a pair of subject images, 2) the number of edges of the subject image, 3) the sharpness of the subject image, and 4) the subject image. The S level (SELECT LEVEL), which is an evaluation value with the light / dark ratio as a parameter, is calculated. Based on the value of the S level for each focus detection line, in the center of the screen, the final detection result among the vertical eye (L1, L2), the horizontal eye (L3, L4), and the F2.8 eye (L19) A focus detection line for obtaining a focus detection result to be determined is determined. Note that determining the focus detection line also means determining a pair of light receiving units on the focus detection sensor 17 and finally determining a focus detection result used for focus adjustment. As a result, the focus detection line can be selected in consideration of the above four elements (parameters), so that an appropriate focus detection line with little focus detection variation can be selected. In the following description, a focus detection line that obtains a focus detection result to be a final detection result is referred to as a specific focus detection line (specific focus detection visual field) .

  First, 1) the degree of coincidence between a pair of subject images will be described with reference to FIG. When a pair of subject images are photoelectrically converted by the pair of light receiving portions of the focus detection sensor 17 described with reference to FIG. 6, a pair of image signals is obtained. Hereinafter, this pair of image signals is referred to as an A image and a B image. Here, it is assumed that an A image and a B image as shown in FIG. 8 are obtained. FIG. 8 shows a case where ghost light or the like is incident on either the A image or the B image and the two images do not have the same waveform for easy understanding of the explanation of the degree of coincidence. Since the A image and the B image have different shapes, a portion that does not match even in a focused state occurs. The area of this non-matching part is defined as the matching degree.

  When the number of pixel columns constituting the light receiving unit is N and the outputs of the i-th pixel of the A image and B image at the time of focusing are a [i] and b [i], the matching degree U can be expressed as follows.

As shown in FIG. 8 and Expression (1), by defining the sum of absolute values of pixel output differences ( differences between a pair of object images) between the A image and the B image as the degree of coincidence U, The degree of coincidence can be accurately expressed. When the coincidence between the A image and the B image is low, the coincidence U increases, and when the coincidence between the A image and the B image is high, the coincidence U decreases.

  By determining the specific focus detection line in consideration of the degree of coincidence U defined in this way, it becomes possible to easily select a focus detection line having a good AB image coincidence and high reliability. Further, the AB image differs depending on the incidence of ghost light or the like, making it difficult to select a focus detection line with low reliability. Thereby, the focus detection accuracy can be improved.

Next, 2) the number of edges of the subject image will be described with reference to FIGS. 9A and 9B. In the focus detection apparatus 20 of the present embodiment, the correlation change amount calculated in the correlation calculation is used as a parameter indicating the number of edges. FIG. 9A is an A image and a B image in a focused state. On the other hand, FIG. 9B shows the A and B images when the relative positional relationship between the A and B images is shifted by one pixel from the focused state ( shifted by a predetermined number of pixels) . The correlation change amount is obtained by calculating the correlation amount in each state and calculating the change amount. Assuming that the correlation amount in FIG. 9A is V1, the correlation amount in FIG. 9B is V2, and the correlation change amount is ΔV, each is obtained as follows.

  The correlation change amount ΔV is an area of a non-matching portion that is generated by shifting from the focused state to the shifted state from the state where the edges of the A and B images are shifted by one pixel from the focused state. For this reason, when the number of edges of the A image and the B image increases, the area of the non-matching portion caused by the deviation also increases. From this, it is understood that the correlation change amount ΔV is appropriate as a parameter representing the number of edges. Note that the shift amount may be larger than one pixel.

  In the correlation calculation, the shift amount at which the correlation change amount becomes zero is calculated, and the defocus amount of the photographing optical system 1 is detected. Therefore, the larger the correlation change amount, the smaller the shift amount calculation error and the detection variation can be suppressed. For this reason, it is reasonable to use the correlation change amount ΔV when determining the line to be selected as the final result.

  By determining the specific focus detection line in consideration of the thus defined correlation change amount ΔV, it is easy to select a focus detection line corresponding to a more reliable subject with a large number of edges and a large amount of information. can do. Thereby, the focus detection accuracy can be improved.

Next, 3) the sharpness of the subject image will be described with reference to FIG. The sharpness referred to in the present embodiment represents whether the value changes sharply or gradually when the A image and B image (signal values) change from a certain bottom value to a peak value. . In this embodiment, the sharpness is a ratio between the primary contrast evaluation value C1 and the secondary contrast evaluation value C2 obtained from the A image and the B image. The primary contrast evaluation value is a sum of absolute values of output differences of adjacent pixels in the AB image, and the secondary contrast evaluation value is a square sum of output differences of adjacent pixels in the AB image. When the number of pixel columns constituting the light receiving unit of the focus detection sensor 17 is N, and the i-th pixel output of the B image is a [i], b [i], the primary contrast evaluation value C1, The secondary contrast evaluation value C2 can be expressed as follows.

  Since the primary contrast evaluation value C1 is the absolute sum of the adjacent pixel output differences, the primary contrast evaluation value C1 obtained from the edge portion of the waveform can be obtained even if the gradation change of the edge portion is steep or gentle. If the peak value and the bottom value of the edge part are the same, they are the same. For example, the primary contrast evaluation value C1 obtained from a subject in which half is white, half is black, and white and black are in contact with the boundary line, and one end is white, the other end is black, and from the white end to the black end is gentle. It is the same as the primary contrast evaluation value C1 obtained from a subject whose color changes.

  On the other hand, since the secondary contrast evaluation value C2 is the sum of squares of the adjacent pixel output differences, the value obtained from the edge portion where the gradation change is steep is larger than that obtained from the edge portion where the gradation change is gentle. Become. This is also clear from equation (6). For this reason, it is appropriate to estimate the magnitude of the gradation change of the edge portion based on the secondary contrast evaluation value C2.

  When the sharpness of the subject image is SH, the sharpness SH can be expressed as follows as a ratio of the primary contrast evaluation value C1 and the secondary contrast evaluation value C2.

Here, the reason why the sharpness SH is not expressed by the secondary contrast evaluation value C2 and is divided by the primary contrast evaluation value C1 is to eliminate the influence of the number of edges. For example, C2_1 is a secondary contrast evaluation value obtained from a subject in which half is white, half is black, and white and black are bordered by a boundary line, both ends are black, the center is white, and white and black are two boundary lines. Let C2_2 be the secondary contrast evaluation value obtained from the subject that is in contact with. Sharpness indicates whether the value changes sharply or gradually when the AB image changes from a certain bottom value to a peak value, so the number of edges is one. Whether it is two or two, it should be the same value. Now, the following relationship holds between the secondary contrast evaluation value C2_1 of the edge chart and the secondary contrast evaluation value C2_2 of the single bar chart.

  As shown in Expression (8), the secondary contrast evaluation value C2 increases as the number of edges increases. Therefore, it is inappropriate to express the sharpness only by the secondary contrast evaluation value C2. Therefore, in order to eliminate the influence of the number of edges, normalization is performed by dividing by the primary contrast evaluation value C1. C1_1 is a primary contrast evaluation value obtained from a subject in which half is white, half is black, and white and black meet at the boundary line, both ends are black, the center is white, and white and black meet at two boundary lines When the primary contrast evaluation value obtained from such a subject is C1_2, the following relationship is established.

  In addition, the sharpness obtained from a subject in which half is white, half is black, and white and black are in contact with the boundary line is SH_1, both ends are black, the center is white, and white and black are in contact with two boundary lines. When the sharpness obtained from such a subject is SH_2, it can be expressed as follows.

  Substituting Equation (8) and Equation (9) into Equation (11), SH_2 can be expressed as follows.

  From equation (12), sharpness SH_1 obtained from a subject in which half is white, half is black, and white and black are in contact with the border, and two borders are black at both ends, white at the center, and white and black It can be seen that the sharpness SH_2 obtained from the subject in contact with is the same. From the same consideration, it can be seen that the value of the sharpness SH does not change no matter how many edges are added.

  From the above, the sharpness of the subject image can be accurately expressed by setting the ratio between the primary contrast evaluation value and the secondary contrast evaluation value as sharpness. Then, by determining the specific focus detection line in consideration of sharpness, it becomes possible to make it difficult to select a focus detection line corresponding to a subject or the like whose color changes gently. Thereby, the focus detection accuracy can be improved.

  Next, 4) the contrast ratio of the subject image will be described with reference to FIG. The contrast ratio in the present embodiment is a parameter that indicates whether or not the density of the subject image is clear. Specifically, it indicates how much the height from the bottom value to the peak value of the subject image (signal value) is larger than the height from the dark value to the peak value of the sensor output. Yes.

  The dark value, bottom value, and peak value of the A image in the focused state are DARK_A, BOTTOM_A, and PEAK_A, respectively, and the dark value, bottom value, and peak value of the B image are DARK_B, BOTTOM_B, and PEAK_B. In this case, the light / dark ratio PBD_A obtained from the A image and the light / dark ratio PBD_B obtained from the B image can be expressed as follows.

  Of the light / dark ratios obtained from the A and B images as described above, the larger one is defined as the light / dark ratio PBD.

  According to the light / dark ratio PBD shown in Expression (15), the height from the bottom value to the peak value of the subject image becomes larger than the height from the dark value to the peak value of the sensor output. Can be expressed. As can be seen from FIG. 11 and equation (15), the light / dark ratio PBD is a number between 0 and 1. When the light / dark ratio PBD is close to 1, it means that the height from the bottom value to the peak value of the subject image is substantially the same as the height from the dark value to the peak value of the sensor output. In this case, it is determined that the subject image is clear.

  On the contrary, when the contrast ratio is close to 0, it means that the height from the bottom value to the peak value is very small compared to the height from the dark value to the peak value. In this case, it is determined that the subject is not clear.

  As described above, by using the light / dark ratio PBD calculated based on the dark value, the bottom value, and the peak value of the subject image, it can be accurately determined whether or not the subject image is clear. By determining the specific focus detection line in consideration of the light / dark ratio PBD, it is possible to make it difficult to select a focus detection line corresponding to a subject with many variations and an indistinct shade. Thereby, the focus detection accuracy can be improved.

  In the focus detection apparatus of this embodiment, the S level for selecting a specific focus detection line is calculated based on the above four parameters. Here, 1) As the degree of coincidence U is smaller, the reliability is higher. 2) The larger the value of the correlation change amount ΔV, the higher the reliability. 3) The greater the value of the sharpness SH, the higher the reliability. 4) As the value of the light / dark ratio PBD is larger, the reliability is higher. Therefore, the S level is defined as follows.

  Here, if the degree of coincidence U is used as a numerator and the product of the correlation change amount ΔV, the sharpness SH and the light / dark ratio PBD is used as the denominator, if it is determined that the reliability is high from each parameter, the S level is reduced. Defined. The S level defined in this way is calculated for each focus detection line, and the focus detection line with the smallest S level is set as the specific focus detection line, and the focus detection result is selected as the final result.

  For focus detection lines composed of two lines such as the vertical eye (L1, L2) and the horizontal eye (L3, L4), take the average of the two lines or adopt the larger or smaller one Then, confirm the S level. The specific focus detection line can be selected in consideration of all of the degree of coincidence, the number of edges, sharpness, and contrast ratio, so it is possible to select a focus detection line with high reliability and small variation as the specific focus detection line It becomes. By selecting a specific focus detection line based on such S level, it is possible to improve focus detection accuracy.

  Next, how much the focus detection variation becomes smaller by selecting a line based on the S level will be described with reference to FIGS.

  FIG. 14 is a graph showing how much focus detection variation occurs when a focus detection line is selected in consideration of only the sharpness SH. The horizontal axis indicates 3σ indicating variation when focus detection is performed a plurality of times, and the vertical axis indicates the value of sharpness SH when focus detection is performed on various subjects. Here, focus detection was performed for nine types of subjects, and the relationship between variation 3σ and sharpness SH was plotted. The reason why a plurality of plots exist for the same subject is because a plurality of data are obtained by changing the luminance.

Here, nine types of subjects will be described. “Monochrome” of “Monochrome E1” means that the subject has a white part and a black part, and “E1” means that the white part and the black part are in contact with one edge as a boundary line. is there. That is, “monochrome E1” means that the subject is such that the white portion and the black portion are in contact with each other at one boundary line.

  “Monochrome” of “Monochrome E2” means that the subject has a white part and a black part, and “E2” means that the white part and the black part are in contact with each other with two edges as a boundary line. It is. That is, “monochrome E2” means that the subject is such that the white part and the black part are in contact with each other at two boundary lines.

  Similarly, “Monochrome E2_2” is a subject in which a white portion and a black portion are in contact with each other at two boundary lines, but is given an alias because the pattern is different from “Monochrome E2”.

  “Monochrome” of “Monochrome E4” means that the subject has a white portion and a black portion, and “E4” means that the white portion and the black portion are in contact with each other with two edges as a boundary line. It is. That is, “black and white E4” means that the subject is such that the white portion and the black portion are in contact with each other at two boundary lines.

  “Monochrome” of “Monochrome E0” means that the subject has a white part and a black part, and “E0” means that the boundary between the white part and the black part is not an edge, but from the white part to the black part. It means that it is changing gently. That is, “black and white E0” means that the subject is gently changing from white to black.

  “White gray” of “white gray E1” means that the subject has a white portion and a gray portion, and “E1” means that the white portion and the gray portion are in contact with one edge as a boundary line. It means that. That is, “white gray E1” means that the subject is such that the white portion and the gray portion are in contact with each other at one boundary line.

  Similarly, “white gray E1_2” is an object in which the white portion and the gray portion are in contact with each other at one boundary, but is given an alias because the pattern is different from “white gray E1”.

  “White gray” of “white gray E2” means that the subject has a white portion and a gray portion, and “E2” means that the white portion and the gray portion are in contact with each other with two edges as a boundary line. It means that. That is, “white gray E2” means that the subject is such that the white portion and the gray portion are in contact with each other at two boundary lines.

  A “black line scar” is a subject with many fine line marks on the entire black surface. Basically, it is uniformly black, but it has a fine pattern with low contrast due to reflection of flaws.

  When data is plotted with the variation 3σ on the horizontal axis and a parameter on the vertical axis, if a certain parameter is appropriate for estimating the variation 3σ, the parameter and the variation 3σ have a positive correlation. . For this reason, the plot data is distributed on a straight line rising upward. If such a parameter can be found, it is possible to select the one having the smaller variation 3σ by selecting the focus detection line having the smaller parameter as the specific focus detection line.

  Looking at FIG. 14, the distribution is roughly upward, but the spread is quite large. Looking at the trends for each subject, subjects with only one edge, such as “black and white E1”, “white gray E1”, and “white gray E1_2”, have a relatively small sharpness SH. 3σ is large. This is because the parameter of sharpness SH is not affected by the number of edges. The sharpness SH does not take this into consideration even though the variation is small because the amount of information when performing the correlation calculation is larger when the number of edges is larger. For this reason, “black and white E2” and “black and white E4” have only one edge such as “black and white E1”, “white gray E1”, and “white gray E1_2” having large variations 3σ even though the variation 3σ is smaller. It is easier to select subjects that will not be selected. From the result of FIG. 14, it is understood that a focus detection line with small focus detection result variation 3σ cannot be selected only by considering the sharpness SH.

  FIG. 15 is a graph showing how much focus detection variation occurs when a line is selected in consideration of only the contrast ratio PBD. Similarly to FIG. 14, the horizontal axis indicates the variation 3σ when the focus is detected a plurality of times, and the vertical axis indicates the value of the light / dark ratio PBD when the focus of various subjects is detected.

  As can be seen from FIG. 15, the light / dark ratio is an object having neither a black part nor a white part, such as “white gray E1”, “white gray E1_2”, “white gray E2”, and “black scratches”. For a subject having both a black portion and a white portion, the value is almost 1 for all of the subjects. This means that the parameter of the light / dark ratio is how large the difference between the sensor output values of the black part and the white part is compared to the difference between the dark value and the sensor output value of the white part. Because it is only. For this reason, if the subject has a black portion and a white portion, it is natural that the contrast ratio is almost 1. For this reason, the light / dark ratio PBD and the focus detection variation 3σ have no correlation, and it is difficult to estimate the focus detection variation 3σ only from the light / dark ratio PBD.

  FIG. 16 is a graph showing how much focus detection variation occurs when a focus detection line is selected considering only the degree of coincidence U and the correlation change amount ΔV. Similarly to FIG. 14, the horizontal axis indicates the variation 3σ when the focus is detected a plurality of times, and the vertical axis indicates the value of the degree of coincidence U / correlation change ΔV when the focus is detected for various subjects.

  As can be seen from FIG. 16, although the distribution is roughly upward, the spread is quite large. Looking at the trends for each subject, “black and white E0” and “black flaws” have a large focus detection variation 3σ even though the degree of coincidence U / correlation change ΔV is relatively small. As shown in FIG. 12, the degree of coincidence U / correlation change ΔV for “monochrome E0” is very small depending on the positional relationship, and the value of coincidence U as a parameter is This is because it becomes very small.

  Further, the degree of coincidence U / correlation change ΔV of “black line scar” is small due to the effect of the peak / bottom processing function mounted on the focus detection sensor 17. The peak / bottom processing function stretches the subject image so that the bottom value of the subject image (image signal) becomes the minimum value of the sensor output, so that the peak value becomes the maximum value of the sensor output. However, it is a function that uses the full gradation of sensor output. Thereby, the gradation of a subject with low brightness can be increased. “Black line scratches” have very low brightness because there are many fine line scratches on a black background. Such a subject with very low brightness has a very large variation in focus detection because even if it is stretched by the peak / bottom processing, a subject having almost no gradation is forcibly stretched.

  Furthermore, many fine flaws are image signals composed of high-frequency components, but since high-frequency components that are higher than the Nyquist frequency determined from the pixel aperture cannot be detected by the sensor, they become aliasing components and cause detection errors. . Nevertheless, the “black flaws” are stretched by innumerable flaws by the peak / bottom processing, resulting in a waveform in which a lot of bright and dark edges exist as shown in FIG. Then, the correlation change amount ΔV becomes a very large value, and as a result, the matching degree U / correlation change amount ΔV becomes very small.

Such a dirty “black line scar” has a large focus detection variation 3σ due to the influence of the peak / bottom processing and the high frequency component higher than the Nyquist frequency, although the degree of coincidence U / correlation change ΔV is small. It will be lost.

  For the reasons described above, the value of the degree of coincidence U / correlation change ΔV becomes small when “monochrome E0” or “black flaws” with large focus detection variation. From the result of FIG. 16, it is understood that it is difficult to select a focus detection line having a small focus detection result variation 3σ only by considering the degree of coincidence U / correlation change ΔV.

  Therefore, with reference to FIG. 17, it will be described that if line selection is performed according to the S level of the present embodiment, focus detection variation can be suppressed. The horizontal axis of FIG. 17 indicates the variation 3σ when the focus is detected a plurality of times, and the vertical axis indicates the value of the S level when the focus of various subjects is detected.

  As can be seen from the results of FIG. 17, if the focus detection line is selected according to the S level, a positive correlation is established between the focus detection variation 3σ and the S level, and the plot data is on a straight line rising upward. Distributed. Compared to the case where the sharpness SH and the contrast ratio PBD described above are used alone as the focus detection line selection parameters and the case where the degree of coincidence U / correlation change ΔV is used as the focus detection line selection parameters, the spread of the distribution is considerably reduced. ing. That is, it is possible to select a focus detection line with a small focus detection variation 3σ by selecting a focus detection line with a smaller S level in this embodiment as a specific focus detection line (final detection result).

In the focus detection apparatus 20 of the present embodiment, as described in the description of FIGS. 4, 5, and 6 , the vertical eye (L 1, L 2) and the horizontal eye (L 3, L 4) And there are three focus detection lines of F2.8th (L19). Then, for each of the three focus detection lines (in other words, the three focus detection results), the S level (the matching level U, the number of edges (correlation change ΔV), the sharpness SH, and the light / dark ratio PBD) are used as parameters. (SELECT LEVEL) is calculated. Further, among the three focus detection lines (focus detection results), the one having the smallest value of the S level is selected as a specific focus detection line (final detection result). Thereby, it is possible to select an appropriate focus detection line with less focus detection variation.

  As described above, according to the present embodiment, the specific focus detection line is determined based on the S level using all of the degree of coincidence U, the number of edges (correlation change ΔV), the sharpness SH, and the light / dark ratio PBD as parameters. Thus, an appropriate result with high accuracy can be selected from a plurality of focus detection results for substantially the same part of the subject, and focus detection accuracy can be improved.

  Hereinafter, the outline of the focus detection operation, focus adjustment, and operation flow in the arithmetic circuit 40 in the focus detection apparatus 20 of the present embodiment will be briefly described with reference to the flowchart shown in FIG. This operation is executed in accordance with a computer program stored in a memory provided inside or outside the arithmetic circuit 40.

  When a first stroke operation (so-called half-pressing operation) of the release switch of the camera is detected in the main shooting flow (not shown) of the camera, the arithmetic circuit 40 operates the focus detection device 20 (START).

  In step (abbreviated as “S” in the figure) 201, the focus is received in response to a manual operation (or line-of-sight detection result) of the photographer from among 21 focus detection lines provided in the screen, or according to a predetermined algorithm. Specify the detection line.

  Next, in step 201, it is determined whether or not the designated focus detection line is a focus detection line (L1 to L4 and L19) included in the center of the screen. If it is not the focus detection line at the center of the screen, the process proceeds to step 203, and the focus detection result (defocus amount) based on the photoelectric conversion signals from the pair of light receiving units on the sensor 17 corresponding to the designated focus detection line. Is calculated. Then, the process proceeds to Step 207. On the other hand, if the designated focus detection line is the focus detection line at the center of the screen, the process proceeds to step 204.

  In step 204, the focus detection result (defocus amount) in each focus detection line is calculated based on the photoelectric conversion signal from each pair of light receiving units on the sensor 17 corresponding to each focus detection line included in the center of the screen. calculate.

  Next, in step 205, based on the photoelectric conversion signal from each pair of light receiving units on the sensor 17 corresponding to each focus detection line included in the center of the screen, the matching degree U, the number of edges (correlation change amount ΔV), The sharpness SH and the light / dark ratio PBD are calculated. Further, the S level is calculated from the values of these four parameters.

  Next, in step 206, the focus detection line corresponding to the smallest value among the S level values of each focus detection line is determined (selected) as the specific focus detection line.

  In step 207, the driving amount (including the driving direction) of the focus lens necessary for obtaining the in-focus state is calculated from the defocus amount in the specific focus detection line. Specifically, the number of drive pulses of an actuator (for example, a vibration type motor) that drives the focus lens is calculated.

  Next, in step 208, the focus lens is driven by the calculated drive amount. Thereby, in a normal case, a focused state can be obtained. Thereafter, a predetermined focus determination is performed, and if the focus is obtained, the process returns to the main shooting flow of the camera and waits for the second stroke operation (so-called full press operation) of the release switch for starting the shooting operation. . If the in-focus state has not been obtained, the process returns to step 202 and the focus detection and focus adjustment operations are repeated.

  Here, first, focus detection results for the three focus detection lines in the center of the screen are calculated, then specific focus detection lines are determined (calculation and comparison of S level values), and among the above three focus detection results. The focus detection information at the specific focus detection line determined from the above is selected as the final detection result. However, in the present invention, the specific focus detection line may be determined by other procedures. For example, first, a specific focus detection line may be determined from the three focus detection lines in the center of the screen, and then a focus detection result may be obtained for the determined specific focus detection line.

In the first embodiment, a specific focus detection line (final detection) is selected from a plurality of focus detection lines (focus detection results) in consideration of four parameters of the degree of coincidence U, the number of edges (correlation change amount ΔV), sharpness SH, and contrast ratio PBD. Result). In the present embodiment, a specific focus detection line (final detection result) is derived from a plurality of focus detection lines (focus detection results) in consideration of only the matching degree U, the number of edges (ΔV), and the sharpness SH.

  The configuration of the focus detection apparatus of the present embodiment is the same as that of the first embodiment. As described with reference to FIGS. 4, 5, and 6, three focus detection results are obtained from the vertical eye (L1, L2), the horizontal eye (L3, L4), and the F2.8 eye (L19) at the center of the screen. It is done. In the present embodiment, the specific focus detection line (final detection result) is obtained based on the value of S level (SELECT LEVEL) _2 using the degree of coincidence U obtained from a pair of subject images, the number of edges (ΔV), and the sharpness SH as parameters. decide. As described above, by selecting the specific focus detection line in consideration of the above three elements (parameters), it is possible to select an appropriate focus detection line with little focus detection variation. The degree of coincidence U, the number of edges (ΔV), and the sharpness SH are as described in the first embodiment.

  In this embodiment, the S level_2 is calculated based on the three parameters of the degree of coincidence U, the number of edges (ΔV), and the sharpness SH. 1) The smaller the value of the degree of coincidence U, the higher the reliability. 2) The larger the value of the correlation change amount ΔV, the higher the reliability. 3) The greater the value of the sharpness SH, the higher the reliability. Therefore, the S level is defined as follows.

  In the above formula (17), when the degree of coincidence is the numerator and the product of the correlation change amount ΔV and the sharpness SH is used as the denominator, if it is determined that the reliability is high from each parameter, the value of the S level_2 becomes small. It is defined to be. The S level_2 defined as described above is calculated for each focus detection line, and the focus detection result of the focus detection line (specific focus detection line) having the smallest value of the S level_2 is selected as the final detection result.

  How much the focus detection variation becomes smaller by selecting the specific focus detection line based on the S level_2 calculated in this way will be described with reference to FIG. In FIG. 18, the horizontal axis indicates the variation 3σ when the focus is detected a plurality of times, and the vertical axis indicates the value of S level_2 when the focus is detected for various subjects.

  As can be seen from FIG. 18, if the focus detection line is selected according to the S level_2, a positive correlation is established between the focus detection variation 3σ and the S level_2, and the plot data is straight up. Distributed on the line. The spread of the distribution is also considerably smaller than the examples shown in FIGS. That is, by selecting the focus detection line having the smallest S level_2 in this embodiment as the specific focus detection line (final detection result), it is possible to select a focus detection line having a small focus detection variation 3σ.

  Therefore, according to the present embodiment, an appropriate result with high accuracy can be selected from a plurality of focus detection results for substantially the same part of the subject, and focus detection accuracy can be improved.

  In this embodiment, unlike the first and second embodiments, a plurality of focus detection lines (focus detection results) are considered in consideration of only three parameters of the degree of coincidence U, the number of edges (correlation change ΔV), and the light / dark ratio PBD. A specific focus detection line (final detection result) is derived from

  The configuration of the focus detection apparatus of the present embodiment is the same as that of the first embodiment. As described with reference to FIGS. 4, 5, and 6, three focus detection results are obtained from the vertical eye (L1, L2), the horizontal eye (L3, L4), and the F2.8 eye (L19) at the center of the screen. It is done. In the present embodiment, the specific focus detection line (final) is determined based on the value of S level (SELECT LEVEL) _3 using the degree of coincidence U obtained from a pair of subject images, the number of edges (correlation change ΔV), and the light / dark ratio PBD as parameters. Detection result). As described above, by selecting the specific focus detection line in consideration of the above three elements (parameters), it is possible to select an appropriate focus detection line with little focus detection variation. The degree of coincidence U, the number of edges (ΔV), and the light / dark ratio PBD are as described in the first embodiment.

  In this embodiment, the S level_3 is calculated based on the three parameters of the degree of coincidence U, the number of edges (correlation change amount ΔV), and the light / dark ratio PBD. 1) The smaller the value of the degree of coincidence U, the higher the reliability. 2) The larger the value of the correlation change amount ΔV, the higher the reliability. 4) As the value of the light / dark ratio PBD is larger, the reliability is higher. Therefore, S level_3 is defined as follows.

  In Equation (18), when the degree of coincidence U is used as the numerator and the product of the correlation change amount ΔV and the light / dark ratio PBD is taken as the denominator, if it is determined that the reliability is high from each parameter, the S level — 3 is reduced. It is defined as follows. The S level_3 defined in this way is calculated for each focus detection line, and the focus detection result of the focus detection line (specific focus detection line) having the smallest value of the S level_3 is selected as the final detection result.

  The extent to which the focus detection variation is reduced by selecting the specific focus detection line based on the S level_3 calculated in this way will be described with reference to FIGS. 16 and 19. FIG. In FIG. 19, the horizontal axis indicates the variation 3σ when the focus is detected a plurality of times, and the vertical axis indicates the value of S level_3 when the focus is detected for various subjects.

  Comparing FIG. 16 and FIG. 19 showing the value of the degree of coincidence U / correlation change ΔV, the plot data in both cases has a roughly upward distribution. However, since FIG. 19 considers the contrast ratio of the subject image in FIG. 16, the S levels of “white gray E1”, “white gray E1_2”, “white gray E2”, and “black scratched line”. The value of _3 is large. Accordingly, it is possible to make it difficult to select “white gray E1”, “white gray E1_2”, “white gray E2”, and “black scratch line”, which have large focus detection variations compared to FIG.

  That is, by selecting the focus detection line having the smallest S level_3 in the present embodiment as the specific focus detection line (final detection result), it is possible to select a focus detection line having a small focus detection variation 3σ.

  Therefore, according to the present embodiment, an appropriate result with high accuracy can be selected from a plurality of focus detection results for substantially the same part of the subject, and focus detection accuracy can be improved.

  In Examples 1 to 3, the S level (SELECT LEVEL) is calculated in the three focus detection lines of the vertical eye (L1, L2), the horizontal eye (L3, L4), and the F2.8 eye (L19) at the center of the screen. The final detection result was chosen. On the other hand, in the present embodiment, the S level (SELECT LEVEL) is calculated for all focus detection lines in the imaging range 24 shown in FIG. 4, and the final detection result is selected.

  The configuration of the focus detection apparatus of the present embodiment is the same as that of the first to third embodiments. As described in the description of FIGS. 4, 5, and 6, there are a total of 21 focus detection lines in the imaging range 24, and 21 focus detection results are obtained. The 21 focus detection lines are the central vertical eye (L1, L2), the central horizontal eye (L3, L4), the central F2.8 eye (L19), the four upper and lower horizontal eyes (L5, L6, L11, L12), 2 There are two upper and lower F2.8 eyes (L20, L21), four vertical eyes (L7 to L10), and six peripheral vertical eyes (L13 to L18).

  In this embodiment, based on the magnitude of the S level (SELECT LEVEL) taking into account the degree of coincidence U obtained from a pair of subject images, the number of edges (correlation change ΔV), sharpness SH and light / dark ratio PBD. A specific focus detection line (final detection result) is selected from the focus detection line (focus detection result). As a result, the specific focus detection line can be selected in consideration of the above four elements (parameters), so that an appropriate focus detection line with little focus detection variation can be selected.

  The degree of coincidence U, the number of edges (ΔV), the sharpness SH, and the contrast ratio PBD are as described in the first embodiment. The S level calculation method is the same as the S level calculation method of the first embodiment. For this reason, the distribution of the S level and its effect in the present embodiment are also as described in the first embodiment with reference to FIG.

  When the focus detection line is selected according to the S level of this embodiment, a positive correlation is established between the focus detection variation 3σ and the S level, and the plot data is distributed on a straight line that rises to the right. Yes. The spread of the distribution is also considerably smaller than the examples shown in FIGS. That is, by selecting the focus detection line with the smallest S level as the specific focus detection line, it is possible to select the focus detection line with the small focus detection variation 3σ. As a result, it is possible to select a focus detection line that can obtain an appropriate focus detection result with high accuracy from a plurality of focus detection lines existing on the entire screen, and it is possible to improve focus detection accuracy.

  Note that the operation of the arithmetic circuit 40 in this embodiment is performed by deleting steps 201 to 203 in the flowchart shown in FIG. 20 and performing focus detection on all 21 focus detection lines in steps 204 and 205. The result and the S level value may be calculated. In step 206, the focus detection line having the smallest S level value may be selected as the specific focus detection line from the 21 focus detection lines.

  Further, even when the selection target of the specific focus detection line is the all-focus detection line as described above, the S level is calculated using only three parameters as described in the second and third embodiments, and the S level is calculated. It may be used to select a specific focus detection line.

  In addition, the calculation method of the degree of coincidence U, the number of edges (ΔV), the sharpness SH, and the light / dark ratio PBD described in the above embodiments is merely an example, and is obtained by another calculation method having the same meaning as each parameter. A value may be used. In addition, the calculation method described in the embodiment is only an example for the S level, and may be calculated using another calculation method.

  Further, in each of the above embodiments, a single-lens reflex digital camera has been described. However, the present invention can also be applied to a case where the focus detection device is mounted on another optical device such as a lens-integrated camera or an interchangeable lens. .

1 is a cross-sectional view of a camera equipped with a focus detection apparatus that is Embodiment 1 of the present invention. 1 is an exploded perspective view of a focus detection apparatus according to Embodiment 1. FIG. AA sectional drawing in FIG. FIG. 3 is a diagram illustrating a layout of a focus detection field of view in the focus detection apparatus according to the first embodiment. FIG. 3 is a diagram illustrating a lens arrangement of a re-imaging lens in the focus detection apparatus according to the first embodiment. FIG. 3 is a diagram illustrating arrangement of light receiving units on a focus detection sensor in the focus detection apparatus according to the first embodiment. FIG. 3 is a diagram illustrating an output signal waveform from a focus detection sensor according to the first embodiment. FIG. 3 is a diagram illustrating an output signal waveform from a focus detection sensor according to the first embodiment. FIG. 6 is a diagram for explaining a degree of coincidence between a pair of subject images in the first embodiment. FIG. 6 is a diagram illustrating a correlation change amount of a subject image in the first embodiment. FIG. 6 is a diagram illustrating a correlation change amount of a subject image in the first embodiment. FIG. 3 is a diagram for explaining sharpness of a subject image in Embodiment 1. FIG. 6 is a diagram for explaining the contrast ratio of a pair of subject images in the first embodiment. The figure which shows the waveform of the image progression of the to-be-photographed object from which the brightness changes smoothly from a white part to a black part. The figure which shows the waveform of the image signal of the to-be-photographed object which has many line scars on a black background. The figure which shows the relationship between focus detection variation and sharpness. The figure which shows the relationship between focus detection variation and a light / dark ratio. The figure which shows the relationship between a focus detection variation and the value of a coincidence degree / correlation change amount of an image. FIG. 6 is a diagram illustrating a relationship between focus detection variation and S level in the first embodiment. The figure which shows the relationship between the focus detection variation in Example 2 of this invention, and S level. The figure which shows the relationship between the focus detection variation in Example 3 of this invention, and S level. 3 is a flowchart showing the operation of the focus detection apparatus in Embodiment 1.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Image pick-up optical system 2 Main mirror 3 Sub mirror 4 Image pick-up element 7 Field lens 8 Holding member 10 Eccentric cam 11 IR-CUT filter 12 Folding mirror 13 Aperture 14 Re-imaging lens 14-1A-14-7B Lens part of a re-imaging lens DESCRIPTION OF SYMBOLS 16 Sensor holder 17 Focus detection sensor 20 Focus detection apparatus 21 Shutter 22 Optical low-pass filter 24 Shooting range L1-L21 Focus detection line L1A-L21B Light-receiving part on a focus detection sensor

Claims (7)

A sensor having a plurality of pairs of light receiving units corresponding to a plurality of focus detection fields, and photoelectrically converting a pair of object images formed on each pair of light receiving units;
Calculation means for generating focus detection information used for focus adjustment based on signals from a pair of light receiving units corresponding to a specific focus detection field among the plurality of focus detection fields,
The calculation means calculates the degree of coincidence between the pair of object images and the relative positional relationship between the pair of object images calculated for each focus detection field of view and a correlation when the predetermined number of light receiving pixels are shifted from the focused state. The specific focus detection field of view is selected based on an evaluation value using the amount of change in amount, the sharpness of the pair of object images, and the contrast ratio of the pair of object images as parameters ,
The evaluation value is expressed by the following equation when the degree of coincidence is U, the amount of change in the correlation amount is ΔV, the sharpness is SH, and the contrast ratio is PBD. .
Evaluation value = U / (ΔV × SH × PBD) Sharpness of the subject image, the focus detection apparatus according to claim 1, characterized in that the ratio of the absolute sum and the square sum of the outputs of neighboring pixels in the light receiving portion. The light / dark ratio of the subject image is expressed by the following equation, where PEAK is the peak value of the pixel output of the light receiving unit, BOTTOM is the bottom value of the pixel output, and DARK is the dark value. Item 4. The focus detection apparatus according to Item 1 .
Light / dark ratio = (PEAK-BOTTOM) / (PEAK-DARK) The degree of coincidence, the focus detecting apparatus according to claim 1, characterized in that the sum of the absolute values of the differences of the pair of object images. It said calculating means, among the plurality of focus detection field, according to claims 1, characterized in that performing an operation on each of a plurality of the focus detection field in any one of the 4 comprise the same viewing area Focus detection device. An optical apparatus characterized by claim comprising a focus detecting device according to claim 1 to any one of 5, comprises a control means for performing focus adjustment control based on the focus detection information. A first step of obtaining a signal from a sensor having a plurality of pairs of light receiving units corresponding to a plurality of focus detection fields, and photoelectrically converting a pair of object images formed on each pair of light receiving units;
A second step of generating focus detection information used for focus adjustment based on signals from a pair of light receiving units corresponding to a specific focus detection field among the plurality of focus detection fields;
In the second step, when the degree of coincidence of the pair of object images and the relative positional relationship between the pair of object images calculated for each focus detection field of view are shifted from the focused state by a predetermined number of light receiving pixels. The specific focus detection field of view is selected based on the evaluation value using the amount of change in the correlation amount, the sharpness of the pair of object images, and the contrast ratio of the pair of object images as parameters ,
The focus detection method is characterized in that the evaluation value is expressed by the following equation when the degree of coincidence is U, the amount of change in the correlation amount is ΔV, the sharpness is SH, and the contrast ratio is PBD. .
Evaluation value = U / (ΔV × SH × PBD)