JP2010016743A - Distance measuring apparatus, distance measuring method, distance measuring program, or imaging device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform a distance measurement with high precision rate even when an object distance is spatially varied. <P>SOLUTION: In a distance measuring apparatus having a blurring correlation amount operating unit (103) for operating correlation amount of blurring in each of process target areas (k1, k2) including distance measuring target pixel among a plurality of images in which blurring imaged by different imaging parameters is different, and an object distance-deciding unit (109) for calculating an object distance in every process target area from correlation amount of blurring operated in every process target area, the blurring correlation amount operating unit (103) comprises a pixel correlation amount operating unit (106) for operating the correlation amount of blurring of a pixel unit in every pixel within the process target area, a weighting control unit (107) for setting a weight coefficient of the correlation amount of blurring of the pixel unit in every pixel within the process target area, and a weighting average-operating unit (108) for operating a value of the weighting average as the correlation amount of blurring of the process target area by performing weighting average of the correlation amount of blurring of the pixel unit based on the weight coefficient. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a technique for measuring the distance of a subject using image information.

  2. Description of the Related Art Conventionally, a depth from defocus (DFD) method has been proposed as a system that can obtain distance information of a scene to be captured at an imaging apparatus such as a digital camera at a low cost. In the DFD method, a plurality of images with different blurs are captured by an imaging device such as a CCD (charge coupled device) or a CMOS (complementary metal oxide semiconductor) by controlling imaging parameters such as a photographing lens. Further, a blur correlation amount between the images is calculated for each predetermined target pixel. By referring to a reference table that defines the relationship between the blur correlation amount and the subject distance, the subject distance corresponding to the blur correlation amount is obtained, and distance measurement is performed on the target pixel (see Non-Patent Document 1).

  Here, the blur correlation amount indicates the degree of blur correlation between the images, and is a value correlated with the dispersion of the point spread function (PSF) of the optical system in each image. For example, the blur correlation amount is represented by a difference in dispersion of the point spread function (PSF) of each image. PSF is a function representing the spread (that is, how to blur) a light ray when an ideal point image passes through the optical system.

In the DFD method, it is theoretically possible to perform distance measurement in units of one pixel. In order to stabilize the distance measurement accuracy, it is necessary to calculate the subject distance for the distance measurement target pixel in consideration of the blur correlation amount of the pixel area (also referred to as the DFD kernel) in the vicinity of the distance measurement target pixel. .
M. Subbarao and G. Surya, "Depth from Defocus: A Spatial Domain Approach," International Journal of Computer Vision, Vol. 13, No. 3, pp. 271-294, 1994

  However, when the subject distance varies spatially in the DFD kernel, both the long-distance pixels and the short-distance pixels are taken into account. May be reduced. Specifically, when the boundary between a distant object and a near object exists in the DFD kernel, the ranging accuracy may be low.

  An object of the present invention is to perform distance measurement with high accuracy even when the subject distance fluctuates in space.

  A distance measuring apparatus according to an aspect of the present invention provides a blur correlation amount for each processing target region including a distance measurement target pixel that is a target of distance measurement between a plurality of images with different blurs captured with different shooting parameters. A blur correlation amount calculation unit that calculates a subject distance determination unit that calculates a subject distance for each processing target region from a blur correlation amount calculated for each processing target region, The blur correlation amount calculation unit includes a pixel correlation amount calculation unit that calculates a correlation amount of blur in units of pixels for each pixel in the processing target region, and a blur in units of pixels for each pixel in the processing target region. A weight control unit that sets a weighting coefficient of the correlation amount, and performs a weighted average of the blur correlation amount of the pixel unit based on the weighting factor, and uses the value of the weighted average as a blur correlation amount of the processing target region Weighted average calculation unit to calculate , Characterized in that it comprises a.

  A distance measuring method according to another aspect of the present invention provides a correlation of blur for each processing target area including a distance measuring target pixel to be a distance measuring target between a plurality of images having different blurs shot with different shooting parameters. A blur correlation amount calculating step for calculating an amount; and a subject distance determining step for calculating a subject distance for each processing target region from a blur correlation amount calculated for each processing target region. The blur correlation amount calculating step includes: a pixel correlation amount calculating step for calculating a blur correlation amount for each pixel in the processing target region; and a pixel unit for each pixel in the processing target region. A weighting control step for setting a weighting coefficient for a blur correlation amount, and performing a weighted average of the blur correlation amount for each pixel based on the weighting factor, and calculating the weighted average value as a blur correlation of the processing target region Characterized in that it and a weighted average calculation step of calculating a.

  A distance measuring program according to still another aspect of the present invention provides a blur detection method for each processing target area including a distance measurement target pixel to be a distance measurement object between a plurality of images having different blurs shot with different shooting parameters. A distance measurement program comprising: a blur correlation amount calculation procedure for calculating a correlation amount; and a subject distance determination procedure for calculating a subject distance for each processing target region from the blur correlation amount calculated for each processing target region. The blur correlation amount calculation procedure includes a pixel correlation amount calculation procedure for calculating a blur correlation amount in pixel units for each pixel in the processing target region, and a pixel unit for each pixel in the processing target region. A weighting control procedure for setting a weighting coefficient of a blur correlation amount, and performing a weighted average of the blur correlation amount in pixel units based on the weighting factor, and calculating the weighted average value as a blur correlation of the processing target region amount Characterized in that it comprises a weighted average calculation procedure for calculating by.

  According to these aspects, the blur correlation amount in pixel units at each pixel in the processing target area is calculated. Further, the weighted average value of the blur correlation amount in pixel units is calculated as the blur correlation amount of the processing target region, and the subject distance can be calculated for each processing target region from the weighted average value. Thus, even if there is a variation in subject distance within the processing target area, the calculated subject distance corresponds to a weighted average obtained within the processing target area. Therefore, setting the processing target area maintains the stability of the ranging accuracy, and further reduces the influence of the change in the subject distance in the processing target area on the ranging accuracy.

  According to the present invention, it is possible to reduce the influence of the change in the subject distance in the processing target region on the distance measurement accuracy of the distance measurement target pixel while maintaining the stability of the distance measurement accuracy.

[First embodiment]
A distance measuring apparatus according to the first embodiment will be described with reference to FIG. The first embodiment will be described assuming that the distance measuring device is mounted on an imaging device (imaging electronic device) such as a digital camera or a digital video camera. However, the present invention is applicable without being limited to this.

  FIG. 1 shows an imaging device equipped with a distance measuring device according to the first embodiment. The imaging apparatus includes an optical system 10, an optical system control unit 20, an image sensor 101, a luminance signal control unit 102, a blur correlation amount calculation unit 103, a subject distance determination unit 109, an LUT storage unit 110, and a memory 111. The luminance signal control unit 102, the blur correlation amount calculation unit 103, and the subject distance determination unit 109 (or all of them) may be configured by a logic circuit, or from a CPU (Central Processing Unit) and a memory that stores a calculation program. It may be configured. The distance measuring device includes a blur correlation amount calculation unit 103, a subject distance determination unit 109, an LUT storage unit 110, and the like.

  The blur correlation amount calculation unit 103 includes a buffer 104, a kernel position control unit 105, a pixel unit blur correlation amount calculation unit 106, a weight control unit 107, and a weighted average blur correlation amount calculation unit 108. Thereafter, the “kernel position control unit” is the kernel control unit, the “subject distance determination unit” is the distance determination unit, the “pixel unit blur correlation amount calculation unit” is the pixel correlation amount calculation unit, and the “weighted average blur correlation amount calculation”. Part "is called a weighted average calculation part.

  The optical system 10 includes a plurality of lens groups 301, a diaphragm 303, and the like, and functions as a photographing lens that forms an image of light taken from outside. In FIG. 1, the optical system 10 is shown as a single focus lens, but the configuration of the optical system 10 is not limited to that shown in FIG. 1. For example, the optical system 10 may constitute a zoom lens. For example, the number of lens groups 301 may be any number, and the position of the diaphragm 303 can be changed as appropriate. The focal position of the optical system 10 is variable by the optical system control unit 20.

  The optical system control unit 20 functions as a shooting parameter control unit that changes shooting parameters (camera parameters). The shooting parameters include the focal position of the optical system 10 (or the position of the focus lens), the aperture diameter of the diaphragm 303, the focal length of the optical system 10, and the like. Hereinafter, in the present embodiment, a case where the imaging parameter is the focal position of the optical system 10 will be described.

  The optical system control unit 20 includes a lens driving mechanism that moves one or more lens groups 301 to change the focal position of the optical system 10. The lens group may be composed of a single lens. Hereinafter, one or more lens groups 301 that move to change the focal position of the optical system 10 are referred to as focus lenses. The lens driving mechanism includes a motor (or actuator) such as an electric motor or an ultrasonic motor. The lens driving mechanism may move all the lens groups 301 at once by the all-group movement method. Alternatively, the lens driving mechanism may move one lens group by a front focusing method, a back focusing method, or the like. The optical system control unit 20 includes a diaphragm control mechanism for adjusting the aperture diameter of the diaphragm 303. The aperture control mechanism includes a motor (or actuator) such as an electric motor.

  The position of the focus lens (hereinafter referred to as the focus lens position) is detected by the lens position detection unit 401. The detected focus lens position may be used by the optical system control unit 20, the blur correlation amount calculation unit 103, the distance determination unit 109, and the like as necessary. The lens position detection unit 401 includes a sensor that detects the amount of movement of the lens group that moves to change the focal position. The aperture diameter of the diaphragm 303 is detected by the aperture diameter detector 403. The detected aperture diameter may be used by the optical system control unit 20, the blur correlation amount calculation unit 103, the distance determination unit 109, and the like as necessary. The aperture diameter detection unit 403 may detect an operation amount (rotation amount or the like) of the motor of the aperture control mechanism as an index of the aperture diameter.

  A light beam that passes through the optical system 10 and forms an image is converted into an electrical signal on the image sensor 101. The image sensor 101 is composed of a CCD (charge coupled device) or a CMOS (complementary metal oxide semiconductor). This electric signal is sent out as image data from the image sensor 101.

  The luminance signal control unit 102 converts image data captured by the image sensor 101 into a digital signal. This digital signal is called a luminance signal. The luminance signal control unit 102 performs processing such as image magnification correction processing and luminance distribution normalization processing on the image data converted into the luminance signal. The image magnification correction process is a process for correcting different magnifications between different images.

  The buffer 104 holds the image data converted into the luminance signal by the luminance signal control unit 102. Here, the buffer 104 is a memory having a memory space for holding at least two pieces of image data. The buffer 104 holds a luminance signal converted from two pieces of image data taken from the same position with respect to the same subject. The two pieces of image data are taken and acquired with different shooting parameters and have different blurring methods. In the present embodiment, the shooting parameter is the focal position of the optical system 10, and two pieces of image data having different focal positions are photographed by changing the focus lens position. Therefore, the optical system control unit 20 moves the focus lens (at least one lens group) and changes the focal position of the optical system 10.

  The kernel control unit 105 sets, as a DFD kernel, a pixel group that forms a distance measurement target pixel position to be a distance measurement target and a neighboring area in the plurality of images held in the buffer 104. This DFD kernel is a target of distance measurement processing by the distance measuring apparatus, and the DFD kernel may be referred to as a processing target area. The DFD kernel contains a predetermined number of pixels and is set at a predetermined position in the image. The predetermined position is, for example, near the center of the image, but is not particularly limited. A plurality of DFD kernels may be set. Although the shape of the DFD kernel is not particularly limited, in the present embodiment, the DFD kernel is set in a rectangular area having the distance measurement target pixel as a central pixel (or a pixel near the center). Although the case where the predetermined number of pixels of the DFD kernel is 100 (10 × 10) will be described, the predetermined number of pixels is not limited to this. The distance measurement target pixel may be set at any position in the DFD kernel, but in this embodiment, the distance measurement target pixel is arranged at the center position (or near the center) of the DFD kernel.

  The pixel correlation amount calculation unit 106 (pixel unit blur correlation amount calculation unit) extracts the pixel areas set as the DFD kernel from the two pieces of image data held in the buffer 104, respectively. Subsequently, the pixel correlation amount calculation unit 106 calculates a blur correlation amount in units of pixels for each pixel in the DFD kernel.

  The weight control unit 107 sets weighting coefficients corresponding to the blur correlation amounts for each pixel. A weight coefficient is input from the weight control unit 107 to the pixel correlation amount calculation unit 106. The pixel correlation amount calculation unit 106 outputs a value obtained by multiplying the blur correlation amount for each pixel by a weighting coefficient to the weighted average calculation unit 108. Note that a weighting coefficient may be input from the weighting control unit 107 to the weighted average calculation unit 108, and the weighted average calculation unit 108 may multiply the blur correlation amount for each pixel by the weighting coefficient.

  The weighted average calculation unit 108 outputs the sum of the blur correlation amount for each pixel in the weighted DFD kernel as the blur correlation amount of the DFD kernel. That is, the weighted average calculation unit 108 calculates the value of the weighted average (or weighted addition) of the blur correlation amount of each pixel in the DFD kernel.

  The distance determination unit 109 (subject distance determination unit) calculates the subject distance from the weighted average value of the blur correlation amount in units of pixels input from the weighted average calculation unit 108 based on a lookup table (LUT), and stores it in the memory 111. Store. The LUT storage unit 110 includes a memory such as a RAM (Random Access Memory) and a ROM (Read Only Memory), and stores a lookup table. When the distance determination unit 109 stores the subject distance information in the memory 111, the kernel control unit 105 determines whether to change the distance measurement point and select a new DFD kernel from the buffer 104.

  As shown in FIG. 2, the look-up table defines the relationship between the subject distance U and the blur correlation amount σ. The blur correlation amount σ has a linear relationship with the reciprocal of the subject distance U (see Non-Patent Document 1). If there is no corresponding blur correlation amount σ in the lookup table, the subject distance U can be obtained by interpolation. The look-up table may be provided for each state of the optical system 10 (for example, aperture diameter D or focal length f). This is because the relationship between the subject distance and the blur correlation amount depends on the state of the optical system 10 and is represented by an individual relational expression depending on the state of the optical system 10. In this case, the distance determination unit 109 may select a lookup table to be used according to the state of the optical system 10. For example, when the aperture diameter D of the diaphragm 303 and the focal length f of the optical system 10 are variable, a lookup table to be used may be selected according to the aperture diameter D and the focal distance f of the diaphragm.

  The case where the focal length f is variable is when the optical system 10 is an interchangeable lens or when the optical system 10 is a zoom lens. In this case, a focal length detection unit 405 is provided, and the focal length detection unit 405 detects the focal length f. When the optical system 10 is an interchangeable lens, the focal length detection unit 405 can read the focal length f stored in the memory of the interchangeable lens. When the optical system 10 is a zoom lens, the focal length detection unit 405 can detect the operation amount (rotation amount, etc.) of the motor of the optical system control unit 20 that moves the lens group for zooming as an index of the focal length f. The distance determination unit 109 can specify a lookup table to be used from the detected focal length f.

  In addition, since there is a one-to-one correspondence between the focus position of the focus lens and the subject distance, the focus lens focus position at which the optical system 10 is focused is used instead of the subject distance in the lookup table. Good. That is, the look-up table may define a relationship between the focus lens focus position (focus lens focus position) and the blur correlation amount.

  Next, specific processing in the blur correlation amount calculation unit 103 in the present embodiment will be described.

  First, luminance signals acquired at different focal positions are held in the buffer 104 as two pieces of image data G1 and G2 having different blurs. Subsequently, the kernel control unit 105 selects a distance measurement target pixel for measuring the subject distance from the photographed scene. Next, the kernel control unit 105 sets the luminance value (pixel value) of the M × M rectangular area in the vicinity of the distance measurement target pixel as the DFD kernel in the images G1 and G2. Here, the luminance values in the DFD kernel in each of the images G1 and G2 are g1 (u, v), g2 (u, v) {u, v: 1,2,3. Indicated as (cu, cv). Equation (1) shows an arithmetic expression for the blur correlation amount G (u, v) for each pixel at an arbitrary pixel position (u, v) in the DFD kernel.

Here, C is a constant derived from an experimental value, and is a mathematical symbol representing the second derivative (Laplacian) of the luminance value.

  The correlation amount of blur for each pixel is calculated by calculating the difference between the luminance values at the predetermined pixels and the second derivative in two images having different blurs as shown in Expression (1), and calculating the difference between the two images. It is calculated by dividing by the average value of the second derivative at the pixel. The blur correlation amount indicates the degree of blur correlation in units of pixels between images.

  The output Gk of the blur correlation amount calculated by the weighted average calculation unit 108 for the DFD kernel is given by Expression (2).

  The blur correlation amount Gk calculated for the DFD kernel is given as a weighted average value of the blur correlation amount G (u, v) in pixel units. The weighted average value is calculated from the sum of products of the weight coefficient W (u, v) and the blur correlation amount G (u, v). The weight coefficient W (u, v) of the blur correlation amount G (u, v) for each unit pixel is given by the product of Wi (u, v) and Wg (u, v). Here, Wi (u, v) and Wg (u, v) are respectively a weighting factor according to the distance difference between the pixel to be measured and a weighting factor according to the luminance difference from the ranged pixel. Show. Also, σi and σg indicate the standard deviation of normal distribution. The standard deviations σi and σg are appropriately given by experimental data. Further, the sum of the weight coefficients is normalized to 1.

For example, as shown in FIG. 3A, the weighting coefficient Wi (u, v) is the inter-coordinate distance between the pixel (u, v) for calculating the blur correlation amount and the distance measurement target pixel (cu, cv) [ (cu-u) ² + (cv-v) ² ] ^1/2 may be given as a normal distribution function. By giving a weighting coefficient according to such a probability density function, a higher weight can be set for a blur correlation amount calculated for a pixel closer to the distance measurement target pixel. For example, as shown in FIG. 3B, the weighting coefficient Wg (u, v) is a difference G (cu, cv) between the pixel for calculating the blur correlation amount and the blur correlation amount in pixel units related to the distance measurement target pixel. ) -G (u, v) may be given as a normal distribution function. The difference G (cu, cv) −G (u, v) is an indicator of the luminance difference between the pixel for calculating the blur correlation amount and the pixel to be measured. By giving a weighting coefficient according to such a probability density function, a pixel having a smaller luminance difference from the distance measurement target pixel can be set to have a higher weight of the calculated blur correlation amount. Using such a weighting factor W (u, v), the influence of the blur correlation amount of a pixel that is likely to have a small difference in subject distance from the distance measurement target pixel becomes large in subject distance calculation.

Note that the weight coefficient Wi (u, v) is not a normal distribution function, but is a distance [(cu-u) ² + (cv-v) ² ] ^1/2 between (cu, cv) and (u, v). Any function that becomes smaller as it increases can be used. The weight coefficient Wg (u, v) can be used as long as it is a function that becomes smaller as G (cu, cv) −G (u, v) increases, besides the normal distribution function. Further, as the weight coefficient Wg (u, v), a normal distribution function related to the luminance difference may be calculated more directly as in the following equation (3). In the expression (3), in two images G1 and G2, the luminance difference between the pixel for calculating the blur correlation amount and the pixel for distance measurement is summed and averaged g1 (cu, cv) -g1 (u, v Use the normal distribution function for) + g2 (cu, cv) -g2 (u, v).

  Next, the procedure for calculating the subject distance in the distance measuring apparatus according to the present embodiment will be described with reference to FIGS.

  In the flowchart of FIG. 4, in step S1, the optical system control unit 20 moves the focus lens of the optical system 10 to a predetermined first focus lens position L1. The focal position of the optical system 10 is the first focal position corresponding to the first focus lens position L1. Thus, the focal position of the optical system 10 as the imaging parameter is set to the first focal position. FIG. 5A shows a state in which the focus lens of the optical system 10 has moved to a first focus lens position L1 that has been determined in advance.

  In step S2, the first image G1 is acquired at the first focal position (first imaging parameter) of the optical system 10.

  In step S <b> 3, the luminance signal control unit 102 performs processing such as image magnification correction processing and luminance distribution normalization processing on the acquired luminance signal of the first image G <b> 1 and stores it in the buffer 104 as image data. . The image magnification correction process is a process for correcting a different magnification between the first and second images.

  In step S4, the kernel control unit 105 sets the coordinate position (cu, cv) of the distance measurement target pixel. The kernel control unit 105 reads a pixel area including the distance measurement target pixel from the image G1 held in the buffer 104 as a DFD kernel (processing target area).

  In step S5, the optical system control unit 20 moves the focus lens of the optical system 10 to a predetermined second focus lens position L2. The focal position of the optical system 10 is the second focal position corresponding to the second focus lens position L2. Thus, the focal position of the optical system 10 as the imaging parameter is set to the second focal position. FIG. 5B shows a state in which the focus lens of the optical system 10 has moved to a predetermined second focus lens position L2.

  In step S6, the second image G2 is acquired at the second focal position (second imaging parameter) of the optical system 10.

  In step S <b> 7, the luminance signal control unit 102 performs processing such as image magnification correction processing and luminance distribution normalization processing on the acquired luminance signal of the second image G <b> 2 and stores it in the buffer 104 as image data. .

  In step S8, the kernel control unit 105 sets the coordinate position (cu, cv) of the distance measurement target pixel as in S4. The kernel control unit 105 reads a pixel area including a distance measurement target pixel from the image G2 held in the buffer 104 as a DFD kernel (processing target area). The region of the DFD kernel of the image G2 corresponds to the region of the DFD kernel of the image G1. The position of the DFD kernel is basically at the same position in the images G1 and G2 with respect to the image center.

  In step S9, the pixel correlation amount calculation unit 106 calculates a blur correlation amount for each pixel in the DFD kernel.

  In step S <b> 10, the pixel correlation amount calculation unit 106 weights the blur correlation amount for each pixel calculated in step S <b> 9 using a weighting factor defined in advance as a normal distribution function for the DFD kernel. The pixel correlation amount calculation unit 106 calculates the product of the blur correlation amount and the weight coefficient for each pixel as the weighted blur correlation amount.

  In step S11, the weighted average calculation unit 108 obtains the sum of the weighted blur correlation amounts calculated in S10, and calculates the weighted average value of the blur correlation amount G (u, v) for each pixel in the DFD kernel. .

  In step S12, the distance determination unit 109 refers to the lookup table in the LUT storage unit 110 and corresponds to the weighted average value of the blur correlation amount G (u, v) in units of pixels (that is, the blur correlation amount of the DFD kernel). The subject distance to be calculated is calculated.

  In step S <b> 13, the distance determination unit 109 stores the subject distance related to the current distance measurement target pixel in the memory 111.

  In step S14, it is determined whether or not there are other distance measurement target pixels for the image data stored in the buffer. If there are no other pixels to be measured, the process ends.

  If there are other distance measurement target pixels in step S15, the distance measurement target pixels are updated, and then the processes after S4 and S8 are repeated.

  The subject distance calculated in this way is used not only for obtaining the focus position of the focus lens in automatic focus control but also for dimming control during flash photography. In addition, a distance map representing the difference in subject distance in the image can be created from the subject distance obtained for each distance measurement target pixel.

  Next, functions and effects of the first embodiment will be described.

  The weighted average value of the blur correlation amount in pixel units is calculated as the blur correlation amount of the DFD kernel, and the subject distance can be calculated for each DFD kernel from the weighted average value. Thus, even if there is a variation in subject distance in the DFD kernel, the calculated subject distance corresponds to a weighted average obtained in the DFD kernel. Therefore, setting the DFD kernel maintains the stability of the distance measurement accuracy, while reducing the influence on the distance measurement accuracy of the subject distance variation in the DFD kernel.

  Further, when calculating the weighted average value of the pixel unit blur correlation amount in the DFD kernel, a weighting coefficient corresponding to the inter-pixel distance and the luminance difference is used with reference to the distance measurement target pixel. The weighting coefficient is given by a function that decreases as the distance between the pixels to be measured and the pixel for which the pixel unit blur correlation amount is calculated and the luminance difference increases, for example, a normal distribution function. Thereby, the blur correlation amount calculated from a pixel that is close to the distance measurement target pixel position and is likely to have a small subject distance difference from the distance measurement target pixel is emphasized in the subject distance calculation. Therefore, it is possible to calculate the distance information of the subject imaged at the distance measurement target pixel position with high accuracy.

[Modification of First Embodiment]
In the first embodiment, the case where the imaging parameter is the focal position of the optical system 10 has been described. However, the imaging parameter may be the aperture diameter D of the diaphragm 303 or the focal length f of the optical system 10.

  When the photographing parameter is the aperture diameter D, in step S1 of FIG. 4, the optical system control unit 20 sets the aperture diameter D in advance by the aperture control mechanism instead of setting the focus lens position. Set to D1. In step S2, the first image G1 is acquired at the first aperture diameter D1 (first imaging parameter) of the optical system 10. In step S5, the optical system control unit 20 sets the aperture diameter D to a predetermined second aperture diameter D2 by the diaphragm control mechanism. In step S6, the second image G2 is acquired at the second aperture diameter D1 (second imaging parameter) of the optical system 10. The other steps in FIG. 4 are the same as those in the first embodiment.

  When the optical system 10 is a zoom lens and the shooting parameter is the focal length f of the optical system 10, in step S1, the optical system control unit 20 sets the focal length f to a predetermined first focal length f1. To do. In step S2, the first image G1 is acquired at the first focal length f1 (first imaging parameter) of the optical system 10. In step S5, the optical system control unit 20 sets the focal length f to a predetermined second focal length f2. In step S6, the second image G2 is acquired at the second focal length f2 (second imaging parameter) of the optical system 10. The other steps in FIG. 4 are the same as those in the first embodiment.

  Also in this modified example, the stability of the distance measurement accuracy by using the DFD kernel is maintained, and the influence on the distance measurement accuracy due to the variation of the subject distance in the DFD kernel is reduced. For this reason, it is possible to calculate the distance information of the subject imaged at the distance measurement target pixel position with high accuracy.

[Second Embodiment]
With reference to FIG. 6 and FIG. 7, the distance measuring device or the imaging device according to the second embodiment will be described. In the configuration of the distance measuring apparatus or the imaging apparatus as shown in FIG. 6, the description of the same parts as in the first embodiment is omitted, and only the differences in structure will be mainly described.

  In the present embodiment, the configuration of the weight control unit 107 is different from that of the first embodiment. In the present embodiment, the weight control unit 107 includes an area dividing unit 113, a distance measurement target pixel attribute determining unit 114 (also referred to as an attribute determining unit), and an effective region setting unit 115.

  The area dividing unit 113 divides the DFD kernel area into a plurality of areas. The area dividing unit 113 includes a position coordinate determining unit that determines the position coordinates of the main subject area (predetermined pixel area) as a predetermined feature amount from the image information of the shooting scene, and holds the position coordinate information. The area dividing unit 113 divides the DFD kernel into a main subject area and a non-main subject area based on the position coordinates of the main subject area. Labels (identification symbols) are assigned to both divided regions in the DFD kernel. Here, the main subject area indicates an approximate area where the main subject exists, and the non-main subject area is an area other than the main subject area. The main subject area and the non-main subject area are collectively referred to as divided areas.

  A position coordinate determination unit that determines the position coordinates of the main subject area as a predetermined pixel area from the image information of the shooting scene can use a known face recognition technique. In this embodiment, the case of the face recognition technology that holds the main subject area as a rectangular area will be described for the sake of simplification. However, the present embodiment will be described regardless of the shape of the main subject area. Is applicable.

  The attribute determination unit 114 determines whether the distance measurement target pixel is included in the main subject region or the non-main subject region. The effective area setting unit 115 sets an area including a distance measurement target pixel as an effective area in the DFD kernel. The effective area setting unit 115 sets a weighting factor W (u, v) to 0 in a divided area that does not include a distance measurement target pixel and includes a distance measurement target pixel in the main subject area and the non-main subject area. The area weighting factor W (u, v) is set to 1 / N. Here, N is the number of pixels in the divided area including the distance measurement target pixel in the DFD kernel. The effective area setting unit 115 excludes an area having a weighting factor W (u, v) of 0, that is, an area that does not include a distance measurement target pixel from the DFD kernel. Thereby, the pixel correlation amount calculation unit 106 calculates the blur correlation amount G (u, v) for each unit pixel only for the pixels constituting the region including the distance measurement target pixel.

  A specific example of effective area designation in the DFD kernel will be described with reference to FIGS. 7 (a) and 7 (b). FIG. 7A shows an enlarged view of image data obtained by capturing a person and a background as a subject and a DFD kernel corresponding to a distance measurement target pixel. Examples of distance measurement target pixels are kg1 and kg2, and the corresponding DFD kernels are k1 and k2. Further, the face area F of a person imaged as a subject is recognized in advance as a main subject area by using known face recognition means as main subject recognition means. The recognition result is held as the image coordinates of the rectangular face area F.

  The subject distance calculation procedure in this case is shown in the flowchart of FIG. This calculation procedure is obtained by adding DFD kernel effective area designation (S21, S22) before pixel unit blur correlation amount calculation (S9) in the flowchart of the first embodiment of FIG.

  The procedure (subroutine) for specifying the effective area in steps S21 and S22 of FIG. 8 is shown in the flowchart of FIG.

  In step S31, the area dividing unit 113 compares the image coordinates of the DFD kernel and the image coordinates of the face recognition area F, and assigns a label for displaying the pixels of the main subject area to the pixels whose both image coordinates match. A label for displaying a pixel in the non-main subject area is assigned to a pixel in which the image coordinates of the DFD kernel and the image coordinates of the face recognition area F do not match. Thereby, based on the position coordinates of the main subject area, the DFD kernel area is divided into a main subject area and a non-main subject area.

  In step S32, the attribute determination unit 114 determines whether the distance measurement target pixel kg belongs to the main subject region or the non-main subject region based on the coordinate position. In the example of FIG. 7A, it is determined that the distance measurement target pixel Kg1 belongs to a non-main subject, and the distance measurement target pixel kg2 belongs to a main subject.

  In step S33, the effective area setting unit 115 sets the weighting factor of the divided area to which the distance measurement target pixel kg does not belong to zero. When the number of pixels in the divided area to which the distance measurement target pixel kg belongs is N, the effective area setting unit 115 sets the weight coefficient of the blur correlation amount in pixel units calculated in this divided area to 1 / N. Thereby, the average value of the blur correlation amount in the divided area to which the distance measurement target pixel belongs can be calculated.

  When the number of pixels in the area to which the distance measurement target pixel Kg1 belongs is N1, the weight coefficient table corresponding to the DFD kernel k1 is wt1 in FIG. 7B. When the number of pixels in the area to which the distance measurement target pixel kg2 belongs is N2, the weight coefficient table corresponding to the DFD kernel k2 is wt2 in FIG. 7B.

  In step S34, pixels with a weighting factor of 0 are excluded from the DFD kernel.

  In addition, after calculating the blur correlation amount of all pixels of the DFD kernel without excluding pixels, the weighted average is obtained by multiplying the weight coefficient W (u, v) by the blur correlation amount G (u, v) in pixel units. A value may be calculated. In this case, since the weight coefficient W (u, v) of the area not including the distance measurement target pixel is 0, the weighted average value is calculated in consideration of only the pixels of the area including the distance measurement target pixel.

  In addition, by setting the weighting factor to a value other than 0, it is possible to specify a divided area that does not include a distance measurement target pixel and exclude it from the DFD kernel. This is because both areas can be specified if the weighting coefficients of the area including the distance measurement target pixel and the area not including the distance measurement target pixel are different values. Furthermore, the weighting coefficient in the divided area to which the distance measurement target pixel belongs can be set using the above-described normal distribution function.

  Next, the operation and effect of the second embodiment will be described.

  When calculating the average blur correlation amount at each pixel position in the DFD kernel, the area of the DFD kernel is divided into a main subject area and a non-main subject area. By using the position coordinates of the main subject area (predetermined pixel area), the DFD kernel area can be clearly divided. Since different weighting coefficients are assigned to the area including the distance measurement target pixel and the area not including the distance measurement target pixel, the area not including the distance measurement target pixel can be specified simultaneously with the setting of the weighting coefficient. The divided area to which the pixel to be measured does not belong is excluded from the DFD kernel, or the weight coefficient is set to 0. As a result, it is possible to calculate the average value of the blur correlation amount using only the pixels in the divided area where an image of the distance measurement target pixel position and an image with a small subject distance difference are formed. Therefore, the subject distance of the distance measurement target pixel can be calculated with high accuracy.

[Third embodiment]
With reference to FIGS. 10 to 13, the distance measuring apparatus or the imaging apparatus according to the third embodiment will be described. This embodiment is a modification of the second embodiment. In the imaging apparatus according to the third embodiment in FIG. 10, the description of the same parts as in the second embodiment is omitted, and only the differences in structure will be mainly described.

  In the third embodiment, in addition to the configuration of the second embodiment, the weight control unit 107 includes a kernel region control unit 116. The effective area setting unit 115 excludes a divided area having a weighting factor W (u, v) of 0 from the DFD kernel. However, the kernel area control unit 116 provides the area for the number of pixels of the divided area excluded from the calculation target of the pixel-by-pixel blur correlation amount around the DFD kernel, To do.

  The effective area designation procedure (subroutine) in steps S21 and S22 of the flowchart of FIG. 11 according to the third embodiment is an improvement of the flowchart of FIG. In the flowchart of FIG. 11, step S35 is added to the flowchart of FIG. The process of step S31-S34 is the same as that of 2nd embodiment.

  In the second embodiment, since the weighting coefficient table is wt1 or wt2, only some pixels of the DFD kernel extracted for ranging accuracy stability are used as effective areas.

  Referring to FIG. 12, in the third embodiment, in step S <b> 35, the kernel area control unit 116 expands the DFD kernel that is the calculation target of the blur correlation amount in units of pixels. Therefore, the kernel area control unit 116 newly sets a pixel in the vicinity (or surrounding) of the DFD kernel as a part of the DFD kernel. Next, the kernel area control unit 116 assigns the same weighting coefficient to the neighboring pixels as the divided area where the distance measurement target pixel exists.

  A specific method for extending the DFD kernel will be described below with reference to the example of FIG.

  The kernel area control unit 116 newly incorporates pixels around the DFD kernel into the DFD kernel according to the arrangement in the DFD kernel of the divided area to which the distance measurement target pixel belongs. The kernel area control unit 116 newly sets surrounding pixels on the side defining the divided area to which the distance measurement target pixel belongs among the sides of the DFD kernel as the DFD kernel area. Specifically, the kernel region control unit 116 pays attention to the label of the divided region to which the distance measurement target pixel belongs, and detects the longest side that has the same label number among the sides that define the DFD kernel. Subsequently, a pixel region in the normal direction of the detected side around the DFD kernel is incorporated into a part of the DFD kernel. This pixel area has the same or equivalent number of pixels as the number of excluded pixels.

  In FIG. 12, pixels having the same label as the distance measurement target area are arranged on the left side of the DFD kernel k1 for the longest time. Therefore, the DFD kernel is expanded by providing pixel areas corresponding to the excluded three pixel areas in the normal direction of the left side. A weighting table wt1 ′ corresponding to the new expanded DFD kernel k1 ′ is defined.

  In the DFD kernel k2, pixels with the same label are arranged on the right side and the lower side of the kernel for the longest time. For this reason, the DFD kernel is expanded by the number of pixels excluded in the normal direction of both sides. A weighting table wt2 ′ corresponding to the new expanded DFD kernel k2 ′ is defined. The weighting coefficients of the weighting tables wt1 ′ and wt2 ′ are all the same value, and if the number of pixels of the DFD kernel is Na, the weighting coefficient is set to 1 / Na.

  Image data is read from the buffer using a new DFD kernel and a corresponding weighting coefficient, and a blur correlation amount calculation and a weighted average calculation are performed for each pixel.

  Next, the operation and effect of the third embodiment will be described.

  When calculating the average blur correlation amount at each pixel position in the DFD kernel, the DFD kernel is divided, and the weight of the divided region to which the distance measurement target pixel does not belong is set to zero. Also, in order to complement the pixels excluded from the DFD kernel with a weight of 0, a pixel area is selected from the neighboring areas around the DFD kernel and used as a new DFD kernel. As a result, the average value of the blur correlation amount is calculated using only the pixels in the divided area where the image of the distance measurement target pixel position and the image with a small subject distance difference are formed and maintaining the number of effective pixels in the DFD kernel. It becomes possible. Therefore, the subject distance of the distance measurement target pixel can be calculated with higher accuracy.

  Furthermore, since a pixel area to be newly used as the DFD kernel is selected according to the arrangement characteristics in the DFD kernel of the divided area to which the distance measurement target pixel belongs, a new DFD kernel can be appropriately configured.

[Fourth embodiment]
With reference to FIGS. 13 and 14, a distance measuring apparatus or an imaging apparatus according to the fourth embodiment will be described.

  The fourth embodiment is a modification of the second embodiment, and the description of the same parts as those of the second embodiment will be omitted, and only the differences in operation will be mainly described.

  The fourth embodiment is different from the second embodiment in the area division method of the DFD kernel. In the fourth embodiment, region division of the DFD kernel is performed using a predetermined feature amount of pixels of the DFD kernel. As the predetermined feature amount, pixel brightness, RGB color information, color density, saturation, brightness, and the like can be used. Note that RGB color information, color density, saturation, and brightness are collectively referred to as color information.

  An example using RGB color information (particularly skin color information) as a predetermined feature amount will be described below.

  The area dividing unit 113 divides an area in the DFD kernel into a specific color area and an unspecified color area, and gives a label (identification symbol) to the pixel. Here, the specific color area indicates an area having a specific color, and the non-specific color area is an area other than the specific color area. The attribute determination unit 114 determines whether the distance measurement target pixel is included in a specific color region or an unspecific color region. The specific color area and the nonspecific color area are collectively referred to as a divided area.

  The effective area setting unit 115 sets a weighting factor W (u, v) to 0 in a divided area that does not include a distance measurement target pixel among the specific color area and the unspecified color area, and includes a distance measurement target pixel. The area weighting factor W (u, v) is set to 1 / N. Here, N is the number of pixels in the area including the distance measurement target pixel in the DFD kernel. The effective area setting unit 115 excludes an area having a weighting factor W (u, v) of 0 from the DFD kernel. Thereby, the pixel correlation amount calculation unit 106 calculates the blur correlation amount G (u, v) for each unit pixel only for the pixels constituting the region including the distance measurement target pixel.

  In addition, after calculating the blur correlation amount of all pixels of the DFD kernel without excluding pixels, the weighted average is obtained by multiplying the weight coefficient W (u, v) by the blur correlation amount G (u, v) in pixel units. A value may be calculated. In this case, since the weight coefficient W (u, v) of the area not including the distance measurement target pixel is 0, the weighted average value is calculated in consideration of only the pixels of the area including the distance measurement target pixel.

  A specific example of specifying an effective area in the DFD kernel will be described with reference to FIGS. 13 (a) and 13 (b). FIG. 13A shows an enlarged view of image data obtained by capturing a person and a background as a subject and a DFD kernel corresponding to a distance measurement target pixel. Examples of distance measurement target pixels are kg1 and kg2, and the corresponding DFD kernels are k1 and k2. Further, skin color information corresponding to the skin color region S corresponding to the face of a person imaged as a subject is preset.

  With reference to FIG. 14, a subroutine showing a procedure for specifying an effective area (S21, S22) of the DFD kernel in the fourth embodiment will be described.

  In step S41, the area dividing unit 113 compares the color information of each pixel of the DFD kernel with the skin color information. If the difference between the color information of each pixel and the skin color information is within a predetermined range, a label indicating a specific color region (skin color region) is assigned to each pixel. If the difference between the color information and the skin color information of each pixel is outside the predetermined range, a label indicating an unspecified color region (non-skin color region) is assigned to each pixel.

  In step S42, the attribute determination unit 114 determines whether the distance measurement target pixel kg belongs to the specific color area or the non-specific color area based on the similarity between the color information of the distance measurement pixel kg and the skin color information. If the difference between the color information of the distance measurement target pixel kg and the skin color information is within a predetermined range, the distance measurement target pixel kg belongs to the specific color region. In the example of FIG. 13A, it is determined that the distance measurement target pixel Kg1 belongs to an unspecified color area and the distance measurement target pixel kg2 belongs to a specific color area.

  In step S43, the effective area setting unit 115 sets the weighting factor of the divided area to which the distance measurement target pixel kg does not belong to zero. On the other hand, when the number of pixels in the divided area to which the distance measurement target pixel kg belongs is N, the effective area setting unit 115 sets the weight coefficient of this divided area to 1 / N. Thereby, the average value of the blur correlation amount in the divided area to which the distance measurement target pixel belongs can be calculated.

  When the number of pixels in the area to which the distance measurement target pixel Kg1 belongs is N1, the weighting coefficient table corresponding to the DFD kernel k1 is wt1 in FIG. 13B. When the number of pixels in the area to which the distance measurement target pixel kg2 belongs is N2, the weighting coefficient table corresponding to the DFD kernel k2 is wt2 in FIG.

  In step S44, pixels having a weight coefficient of 0 are excluded from the DFD kernel, and a pixel unit blur correlation amount is calculated.

  Note that after calculating the blur correlation amount of all pixels of the DFD kernel, the weighted average value may be calculated by multiplying the blur coefficient G (u, v) by the weighting factor W (u, v). .

  In addition, by setting the weighting factor to a value other than 0, it is possible to specify a divided area that does not include a distance measurement target pixel and exclude it from the DFD kernel. This is because both areas can be specified if the weighting coefficients of the area including the distance measurement target pixel and the area not including the distance measurement target pixel are different values. Furthermore, the weighting coefficient in the divided area to which the distance measurement target pixel belongs can be set using the above-described normal distribution function.

  Note that when the pixel luminance is used as the predetermined feature amount, the region dividing unit 113 compares the luminance of each pixel of the DFD kernel with the reference luminance. If the difference between the luminance of the pixel and the reference luminance is within a predetermined threshold value, the pixel belongs to a specific luminance region that shows substantially specific luminance. If the difference between the luminance of the pixel and the reference luminance is greater than a predetermined threshold, it is assumed that the pixel belongs to the unspecified luminance area. In this way, the DFD kernel can also be divided by brightness and darkness, and the effective area setting unit 115 sets the weighting factor of the divided area to which the distance measurement target pixel does not belong to 0.

  Next, the operation and effect of the fourth embodiment will be described.

  When calculating the average blur correlation amount at each pixel position in the DFD kernel, whether to divide the DFD kernel using a predetermined feature amount of pixels and set the weights other than the divided region to which the distance measurement target pixel belongs to 0 Exclude from DFD kernel. In this way, the DFD kernel can be easily divided using only the pixel feature amount. Furthermore, it is possible to calculate the average value of the blur correlation amount using only the pixels in the divided region where the image of the distance measurement target pixel position and the image with a small subject distance difference are formed. Therefore, the subject distance of the distance measurement target pixel can be calculated with high accuracy.

[Other Embodiments]
In the description of each embodiment described above, hardware processing is assumed as processing performed by the distance measuring device, but the present invention is not limited to such a configuration. For example, a configuration in which processing is performed separately by software is also possible. In this case, the distance measuring device includes a main storage device such as a CPU and a RAM, and a computer-readable storage medium storing a program for realizing all or part of the above processing. Here, this program is called a distance measurement program. Then, the CPU reads out the distance measurement program stored in the storage medium and executes the information processing / calculation processing, thereby realizing the above-described processing.

  Here, the computer-readable recording medium refers to a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, a semiconductor memory, and the like. Alternatively, the distance measurement program may be distributed to a computer via a communication line, and the computer that has received the distribution may execute the distance measurement program.

  The present invention is not limited to the above-described embodiment, and it is obvious that various modifications can be made within the scope of the technical idea.

It is a schematic block diagram which shows the imaging device carrying the ranging device which concerns on 1st embodiment. It is a figure which shows the relationship between subject distance U and blur correlation amount (sigma). (A) It is a graph which shows the relationship between the weight coefficient and the distance of the pixel which calculates a blur correlation amount, and a ranging object pixel. (B) It is a graph which shows the relationship between the weight coefficient and the brightness | luminance difference of the pixel which calculates a blur correlation amount, and a ranging object pixel. It is a flowchart which shows the calculation procedure of the object distance which concerns on 1st embodiment. (A) It is a figure which shows the 1st focus lens position L1. (B) It is a figure which shows the 2nd focus lens position L2. It is a schematic block diagram which shows the imaging device carrying the ranging device which concerns on 2nd embodiment. (A) An enlarged view of a subject image and a DFD kernel corresponding to a distance measurement target pixel is shown. (B) It is a figure which shows the weighting coefficient table which concerns on 2nd embodiment. It is a flowchart which shows the calculation procedure of the object distance which concerns on 2nd embodiment. It is a flowchart which shows the procedure of effective area | region designation | designated which concerns on 2nd embodiment. It is a schematic block diagram which shows the imaging device carrying the ranging device which concerns on 3rd embodiment. It is a flowchart which shows the procedure of effective area | region designation | designated which concerns on 3rd embodiment. It is a figure which shows the weighting coefficient table which concerns on 3rd embodiment. (A) An enlarged view of a subject image and a DFD kernel corresponding to a distance measurement target pixel according to the fourth embodiment. (B) It is a figure which shows the weighting coefficient table which concerns on 4th embodiment. It is a flowchart which shows the procedure of effective area | region designation | designated which concerns on 4th embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Optical system 20 Optical system control part 101 Image pick-up element 102 Luminance signal control part 103 Blur correlation amount calculation part 104 Buffer 105 Kernel control part 106 Pixel correlation amount calculation part 107 Weight control part 108 Weighted average calculation part 109 Subject distance determination part 110 LUT Storage unit 111 Memory 301 Lens group 303 Aperture 401 Lens position detection unit 403 Aperture diameter detection unit 405 Focal length detection unit
k1, k2 target area

Claims (12)

A blur correlation amount calculation unit that calculates a correlation amount of blur for each processing target region including a distance measurement target pixel that is a target of distance measurement between a plurality of images with different blurs shot with different shooting parameters;
A subject distance determining unit that calculates a subject distance for each processing target region from a blur correlation amount calculated for each processing target region;
The blur correlation amount calculation unit
A pixel correlation amount calculation unit that calculates a correlation amount of blur in units of pixels for each pixel in the processing target region;
A weighting control unit that sets a weighting coefficient of the blur correlation amount in units of pixels for each pixel in the processing target region;
A weighted average calculation unit that performs a weighted average of the correlation amount of blur in units of pixels based on the weighting factor, and calculates a value of the weighted average as a blur correlation amount of the processing target region, Ranging device.   The weighting control unit sets the weighting coefficient according to an inter-pixel distance between the distance measurement target pixel and a pixel for which a blur correlation amount of the pixel unit is calculated. Ranging device.   The weighting control unit sets the weighting coefficient according to a difference in luminance between the distance measurement target pixel and a pixel for which a blur correlation amount of the pixel unit is calculated. Ranging device. The weight control unit includes:
An area dividing unit for dividing the processing target area into a plurality of divided areas;
An attribute determination unit that determines which of the plurality of divided regions contains the pixel to be measured;
The distance measuring apparatus according to claim 1, further comprising: an effective area setting unit that excludes a divided area to which the distance measuring target pixel does not belong from the processing target area. The weight control unit
A region dividing unit that divides the processing target region into a plurality of divided regions;
An attribute determination unit that determines in which of the plurality of divided regions the pixel to be measured is contained;
The distance measuring apparatus according to claim 1, further comprising: an effective area setting unit that sets the weighting coefficient to 0 for pixels in a divided area to which the distance measurement target pixel does not belong.   The weight control unit includes an area control unit that expands the processing target area so as to incorporate pixels around the processing target area based on the number of pixels in the divided area to which the ranging target pixel does not belong. The distance measuring device according to claim 4 or 5.   The measurement according to claim 6, wherein the area control unit selects a pixel to be incorporated into the processing target area according to an arrangement of the divided area to which the ranging target pixel belongs in the processing target area. Distance device.   The region dividing unit determines position coordinates of a predetermined pixel region in the image, and divides the processing target region into the predetermined pixel region and other regions based on the position coordinates of the predetermined pixel region. 6. The distance measuring device according to claim 4 or 5, wherein:   6. The distance measuring device according to claim 4, wherein the region dividing unit divides the processing target region based on luminance information or color information of pixels in the processing target region. An optical system for imaging a subject;
An image sensor that converts an image formed by the optical system into an electrical signal;
An imaging apparatus comprising: the distance measuring apparatus according to claim 1. A blur correlation amount calculation step for calculating a blur correlation amount for each processing target region including a distance measurement target pixel to be a distance measurement target between a plurality of images having different blurs shot with different shooting parameters;
A subject distance determination step of calculating a subject distance for each processing target region from a blur correlation amount calculated for each processing target region,
The blur correlation amount calculating step includes:
A pixel correlation amount calculating step for calculating a correlation amount of blur in units of pixels for each pixel in the processing target region;
A weight control step for setting a weighting coefficient of the correlation amount of blur of the pixel unit for each pixel in the processing target region;
A weighted average calculation step of performing a weighted average of the correlation amount of blur of the pixel unit based on the weighting factor, and calculating a value of the weighted average as a blur correlation amount of the processing target region, Ranging method to do. A blur correlation amount calculation procedure for calculating a correlation amount of blur for each processing target region including a distance measurement target pixel to be a distance measurement target between a plurality of images with different blurs shot with different shooting parameters;
A subject distance determination procedure for calculating a subject distance for each processing target region from a blur correlation amount calculated for each processing target region,
The blur correlation amount calculation procedure is as follows:
For each pixel in the processing target area, a pixel correlation amount calculation procedure for calculating a correlation amount of blur in pixel units,
A weighting control procedure for setting a weighting coefficient of a blur correlation amount in units of pixels for each pixel in the processing target region;
A weighted average calculation step of performing a weighted average of the blur correlation amount of the pixel unit based on the weighting factor and calculating a value of the weighted average as a blur correlation amount of the processing target region. Ranging program to do.