JP6004221B2 - Image processing device - Google Patents

Description

  The present invention relates to an image processing apparatus.

  Conventionally, a technique for performing color shading correction of a captured image is known. For example, Patent Document 1 below describes a configuration in which predetermined color shading correction is performed based on subject brightness, color temperature, and the like. Japanese Patent Application Laid-Open No. H10-228561 describes a configuration in which color shading correction is performed using a combination of optical system data and color temperature.

  Patent Document 3 below describes a method of detecting relative luminance in a band near a filter cutoff frequency and selecting a color shading correction coefficient in accordance with the detection result. Patent Document 4 describes a method of correcting a color shading count based on a color temperature estimation result. Patent Document 5 describes a method of determining the type of illumination light based on the color distribution detected by the image sensor and setting a correction amount for color shading correction according to the determination result.

JP 2007-104580 A JP 2008-35282 A JP 2011-091513 A JP 2010-147800 A JP 2009-088800 A

  In the above conventional techniques, color shading is estimated based on subject brightness (luminance), color temperature, and color distribution, but parameters such as subject brightness, color temperature, and color distribution are approximate even when the light source is different. The value may be For this reason, it is difficult to accurately estimate and correct color shading based on subject brightness, color temperature, color distribution, or the like. In particular, in the conventional technique described above, in the case of light sources having different spectrum contents in the vicinity of near-infrared light, shading correction is excessive and insufficient, and it is difficult to correct shading.

  Accordingly, the present invention has been made in view of the above problems, and an object of the present invention is to provide a new and improved image processing apparatus capable of accurately estimating and processing color shading. Is to provide.

  In order to solve the above-described problem, according to an aspect of the present invention, a captured image acquisition unit that acquires a captured image and a result obtained by applying a plurality of different color shading corrections to the captured image are applied to the captured image. There is provided an image processing apparatus comprising: a color shading estimation unit that estimates generated color shading; and a processing unit that performs processing based on the captured image based on the color shading estimated by the color shading estimation unit.

  According to the above configuration, since the color shading generated in the captured image is estimated based on the result of applying a plurality of different color shading corrections to the captured image, the color shading can be estimated with high accuracy.

  In addition, a color shading correction coefficient storage unit that stores a plurality of color shading correction coefficients optimized for each predetermined light source, a block division unit that divides the captured image into a plurality of blocks, and a pixel value statistically for each block The color shading estimation unit multiplies the block statistical value by the color shading correction coefficient, and then outputs an R signal, a G signal, and a B signal for each block. The ratio of the R signal to the signal is calculated, and the color shading correction that minimizes the variation in the ratio of the R signal in the block group that continues from the central portion to the peripheral portion of the image among the plurality of color shading correction coefficients. Select a coefficient. According to this configuration, among the plurality of color shading correction coefficients stored in the color shading correction coefficient storage unit, the variation in the ratio of the R signal in the block group continuous from the central part of the image to the peripheral part is minimized. The color shading correction coefficient is selected. Accordingly, it is possible to compensate for the drop in the R pixel value in the peripheral portion of the image, and it is possible to select an optimal color shading correction coefficient.

  The color shading estimation unit calculates an R signal deviation degree based on a change in the ratio of the R signal in a block group continuous from the central part of the image toward the peripheral part, and all the R signal deviation degrees are calculated. The color shading correction coefficient corresponding to the smallest one is selected. According to this configuration, since the color shading correction coefficient corresponding to the smallest one of all the R signal deviation degrees is selected, when the selected color shading correction coefficient is applied to the divided block, the color shading correction coefficient is selected from the center of the image. The fluctuation of the ratio of the R signal to the peripheral part is the smallest. Accordingly, it is possible to compensate for the drop of the R pixel value in the peripheral portion of the image, and it is possible to select an optimal color shading correction coefficient.

  Further, the R signal deviation degree is calculated using a variance value or a standard deviation of the ratio of the R signals in a block group continuous from the central portion of the image toward the peripheral portion. According to this configuration, the R signal deviation degree is calculated using a variance value or a standard deviation of the ratio of the R signals in the block group. Accordingly, it is possible to select a color shading correction coefficient with the smallest variation in the ratio of the R signal in the block group based on the R signal deviation degree.

  The processing unit corrects the color shading of the captured image based on the color shading correction coefficient selected by the color shading estimation unit. According to this configuration, the color shading of the captured image is corrected based on the color shading correction coefficient selected by the color shading estimation unit. Accordingly, the color shading of the captured image can be corrected with high accuracy.

  In addition, the processing unit determines whether the light source includes a light beam having a wavelength near near infrared light based on the color shading correction coefficient selected by the color shading estimation unit, and estimates the light source. According to this configuration, it is determined based on the color shading correction coefficient selected by the color shading estimation unit whether the light source includes a light beam having a wavelength near near infrared light. Therefore, the type of light source can be determined with high accuracy.

  According to the present invention, it is possible to provide an image processing apparatus capable of accurately estimating and correcting color shading.

It is a figure which shows a mode that the shading generate | occur | produced. It is a block diagram which shows the digital still camera provided with the image processing apparatus which concerns on this embodiment. It is the characteristic figure which compared the difference of the cutoff frequency by the difference in the light reception sensitivity of the R signal of an image sensor, and the incident angle of an infrared cut filter in the center part and peripheral part of the image sensor. It is a schematic diagram which shows the data flow which concerns on this embodiment. It is a flowchart which shows the internal process of a shading estimation part. It is a characteristic view which shows the shading correction coefficient memorize | stored in the shading correction coefficient memory | storage part. It is a figure which shows the result image obtained when shading correction is performed to a captured image. It is a schematic diagram which shows the example of a setting of the block group which calculates R signal fluctuation value. It is a schematic diagram which shows the process in which the optimal shading correction coefficient is calculated.

  Exemplary embodiments of the present invention will be described below in detail with reference to the accompanying drawings. In addition, in this specification and drawing, about the component which has the substantially same function structure, duplication description is abbreviate | omitted by attaching | subjecting the same code | symbol.

  First, shading that occurs in a digital camera or the like will be described. FIG. 1 is a diagram illustrating a state in which shading has occurred. In an electronic imaging device such as a digital camera, an unnatural result (Artifact) called lens shading in which the amount of light at the periphery of an image is attenuated as shown in FIG. 1 due to the influence of an optical system such as a lens. It is known to occur. In particular, color unevenness that occurs due to different shading characteristics for each color of image RGB (red, green, blue) signals obtained from the image sensor is referred to as color shading. For this reason, in the digital camera, processing for correcting shading is performed.

  FIG. 2 is a block diagram illustrating a digital still camera 100 including the image processing apparatus according to the present embodiment. As shown in FIG. 2, the digital camera 100 includes a lens optical system 102, an image sensor 104, an infrared (IR) cut filter 106 mounted on the front surface of the image sensor 104, an AFE circuit (A / D converter) 108, Image signal processing circuit 110, image display unit 112, image recording unit 114, image block statistical circuit 116, driver 118, timing generator (TG) 120, device control unit 122, exposure correction calculation unit 124, white balance (WB) correction coefficient A calculation unit 128, a color shading estimation unit 130, and a shading correction coefficient storage unit 132 are included.

  In the configuration shown in FIG. 2, the lens optical system 102 includes a lens 102 a, a diaphragm 102 b, a shutter 102 c, and the like, and forms a subject image on the imaging surface of the imaging element 104. The lens 102a, the diaphragm 102b, and the shutter 102c of the lens optical system 102 are driven by the device control unit 122 via the driver 118. The exposure correction calculation unit 124 calculates an exposure correction parameter based on a block statistical value described later, and the device control unit 122 drives the aperture 102b and the shutter 102c based on the parameter. The device control unit 122 also controls the timing generator 120 to generate timing for driving the image sensor 104. The white balance correction coefficient calculation unit 128 calculates a white balance correction coefficient based on the block statistical value. The image display unit 112 is configured by a liquid crystal display (LCD) or the like, and displays an image obtained by processing the captured image captured by the image sensor 104 by the image signal processing circuit 110. The image recording unit 114 is a memory that stores an image obtained by processing the captured image captured by the imaging element by the image signal processing circuit 110.

  In an electronic imaging apparatus such as the digital camera 100, an image signal (RGB color signal) acquired from an imaging element 104 that performs photoelectric conversion such as CCD or CMOS is subjected to signal processing by a CDS circuit (not shown). After that, it is A / D converted by the AFE circuit to become a digital signal. The image signal is sent to the image signal processing circuit 110. The image signal processing circuit 110 includes a demosaic processing unit 110a, an edge enhancement unit 110b, a white balance (WB) correction unit 110c, a color shading correction unit 110d, a gamma (Gamma) correction unit 110e, and a light source estimation unit 110f. In each unit, image recording is performed by performing a large number of electronic processing such as demosaicing processing, edge enhancement processing, white balance (WB) correction processing, shading correction processing, and gamma correction processing. As will be described later, the light source estimation unit 110f estimates the light source based on the calculation result of the shading correction coefficient. Note that the color shading correction unit 110d and the light source estimation unit 110f function as processing units that perform processing (shading correction processing, light source estimation processing) based on the captured image based on the color shading estimated by the shading estimation unit 130.

  The color shading correction unit 110d corrects shading that occurs during shooting. In general, the color shading correction unit 110d corrects the state of optical elements such as the zoom and the aperture 102b of the lens 102a of the lens optical system 102, the subject brightness and white balance (WB) correction calculated by the exposure correction calculation unit 124. With reference to light source information such as white balance gain or color temperature calculated by the coefficient calculation unit 128, shading that occurs during shooting is corrected. It should be noted that the color shading correction unit 110d in the drawing can also be configured before the image block (Block) statistical circuit 126.

  On the other hand, in the present embodiment, the color shading correction unit 110d performs the R signal ratio from the center to the periphery of the image from among a plurality of correction tables stored in advance for the block calculated by the image block statistical circuit 116. The shading correction is performed using the correction table selected so that the change in the value becomes the flattest. Therefore, more accurate shading correction can be performed as compared with correction based on information such as subject brightness, white balance gain, and color temperature.

Although it is known that the shading characteristic changes depending on the zoom of the lens optical system 102 and the stop 102b, the characteristic also changes depending on the type of photographing light source. In an electronic imaging device, since it has a light receiving sensitivity in an infrared region having a longer wavelength than visible light, an infrared cut filter (Cut
Infrared light is blocked using a filter. Infrared cut filters are broadly classified into two types, a transmission type and a reflection type (evaporation type). Since the reflective type is formed by vapor deposition on the lens surface, it has the merit that it is easy to reduce the thickness and is cheaper than the transmissive type. On the other hand, the reflection type has a characteristic that the cut-off frequency changes depending on the incident angle of light.

  FIG. 3 shows the difference between the light receiving sensitivity (indicated by the solid line in FIG. 3) of the R (red light) signal of the image sensor and the transmission characteristic (cutoff frequency) due to the difference in the incident angle of the infrared cut filter 106. FIG. 10 is a characteristic diagram (Graph) compared between the central portion and the peripheral portion of the imaging surface (or captured image) of the element 104. In FIG. 3, the horizontal axis indicates the wavelength (nm), and the vertical axis indicates the transmission characteristic (%) of the infrared cut filter 106. In FIG. 3, the light receiving sensitivity of the R signal is indicated by a solid line, the cutoff frequency characteristic at the center of the image sensor 104 is indicated by a broken line, and the cutoff frequency characteristic at the periphery of the image sensor is indicated by a one-dot chain line. Yes.

  The light beam incident on the central portion of the image sensor 104 is incident in a direction that is nearly perpendicular to the imaging surface. On the other hand, a light beam incident on the peripheral portion of the image sensor 104 enters from an oblique direction with respect to the imaging surface. For this reason, as shown in FIG. 3, the cutoff frequency of the infrared cut filter 106 changes between the central portion and the peripheral portion of the image sensor, and the cutoff frequency is higher in the central portion than in the peripheral portion. With the change in the cutoff frequency of the infrared cut filter 106, the light receiving sensitivity characteristic of the R signal in the hatched portion in FIG. 3 shifts to the short wavelength side from the central portion of the image to the peripheral portion. In the example of FIG. 3, when photographing under a light source having a light source wavelength in the vicinity of 660 nm to 700 nm, there is a difference in the area indicated by hatching in the sensitivity characteristic of the R signal between the central portion and the peripheral portion of the image sensor 104. . In this case, light having a wavelength in the vicinity of 660 nm to 700 nm that is imaged without being cut by the infrared cut filter 106 at the central portion is cut by the infrared cut filter 106 at the peripheral portion. Color shading (unevenness) occurs, and the R component is reduced at the periphery rather than at the center. In other words, the difference in R sensitivity does not appear in the captured image under a light source that does not include a wavelength of 660 nm to 700 mn. Through the above process, color shading having different characteristics depending on the light source occurs.

  As a color shading correction technique caused by the infrared cut filter 106, as described in Patent Documents 1 and 2, the subject brightness and the color temperature are calculated, and the possibility of the light source type such as sunlight, a light bulb, and a fluorescent lamp is determined. A method of selecting and applying any one of shading correction tables optimized in advance for a plurality of light sources is known.

  However, even when the color temperature of the light source is approximate, for example, the color characteristics of a light bulb and a light bulb color fluorescent light are different. For this reason, in the conventional method, there is no method other than applying the shading correction table optimized for one of the light sources when the color temperatures are similar as in a light bulb and a light bulb color fluorescent lamp. In the conventional method, when the subject light source is determined to have a low color temperature, the light bulb correction table is applied. Therefore, when the subject light source is actually a light bulb color fluorescent lamp, the shading correction is performed. Is overcorrected, and color unevenness occurs such that the center of the image becomes blue.

  Similarly, in the determination based on the color temperature, since the color temperature approximates that of a fluorescent lamp in a dark outdoor subject such as a sunflower plant, the light source is erroneously determined to be a fluorescent lamp instead of sunlight. In this case, since the fluorescent lamp does not contain much infrared light, shading correction is insufficient, and color shading due to insufficient shading correction occurs such that the center of the image becomes red.

  For this reason, in this embodiment, the color shading resulting from the infrared cut filter 106 is estimated from the captured image, and the color shading that differs for each light source type is effectively removed.

  FIG. 4 is a schematic diagram showing a main part of the configuration related to the shading correction and the data flow (Data flow) according to this embodiment. The image data input from the image sensor 104 is divided into blocks (blocks) composed of n × m blocks (blocks) by the image block statistics circuit 116, and block statistics are calculated for each block. For this reason, the image block statistical circuit 116 includes a block dividing unit 116a that performs block division and a statistical value calculating unit 116b that calculates statistical values. The block statistical value is a pixel integrated value for each RGB or a pixel average value for each RGB in each block.

  The shading correction coefficient storage unit 132 holds a plurality of shading correction coefficients (correction tables) optimized in advance for each light source. Then, the color shading estimation unit 130 applies each of the plurality of correction tables to the block statistical value, reads one shading correction coefficient (color shading correction coefficient) from the shading correction coefficient storage unit 132, and multiplies the block statistical value. The R signal ratio is calculated for each block.

  Next, the color shading estimation unit 130 calculates and temporarily stores the deviation degree of the R signal ratio of the blocks continuous from the central part of the image to the peripheral part using the standard deviation or the variance. Then, a plurality of shading correction coefficients are sequentially read out from the shading correction coefficient storage unit 132, multiplication of block statistics and calculation of R signal ratio deviation are repeated, and R signal ratio deviation corresponding to all shading correction coefficients is repeated. Ask for.

  Then, the color shading estimation unit 130 selects a shading correction coefficient corresponding to the smallest R signal ratio deviation degree among the R signal ratio deviation degrees corresponding to all the shading correction coefficients, and obtains a final shading correction coefficient. Output as.

  The shading correction coefficient output from the color shading estimation unit 130 is sent to the color shading correction unit 110d. The shading correction coefficient is multiplied by the captured image data in the color shading correction unit 110d. Thereby, the color shading of the captured image is corrected.

  FIG. 5 is a flowchart showing internal processing of the color shading estimation unit 130. First, in step S10, a block statistical value is acquired from the image block statistical circuit 116. When the block division is 16 × 16 blocks, block statistics are acquired for each block. In the next step S 12, the shading correction coefficient is read from the shading correction coefficient storage unit 132.

  In the next step S14, the shading correction coefficient read in step S12 is multiplied by each block statistical value to calculate a corrected block statistical value. As described above, since the block statistical value is a pixel integrated value for each RGB or a pixel average value for each RGB in each block, the corrected block statistical value is calculated for each RGB. In the next step S16, the R signal ratio is calculated from the corrected block statistical value obtained in step S14. Specifically, the R signal ratio block statistical value is calculated by calculating R / (R + G + B).

  In the next step S18, the R signal deviation degree is calculated from the R signal ratio block statistical value obtained in step S16. The R signal deviation degree is calculated using the standard deviation or variance for the R signal ratio block statistical value of the blocks continuous from the central part of the image to the peripheral part.

  In the next step S20, it is determined whether or not the calculation of the R signal deviation degree of all the shading correction coefficients stored in the shading correction coefficient storage unit 132 is completed. If not completed, the process returns to step S12. The process after step S12 is performed again. On the other hand, when the calculation of the R signal deviation degree of all the shading correction coefficients is completed, the process proceeds to step S22, and the shading correction coefficient is selected. In step S22, a shading correction coefficient with the smallest R signal deviation is selected.

  After step S22, the captured image (capture image) obtained by imaging with the image sensor 104 is multiplied by the shading correction coefficient selected in step S22.

  FIG. 6 is a characteristic diagram showing the shading correction coefficients stored in the shading correction coefficient storage unit 132. As an example, the shading correction coefficient storage unit 132 holds a shading correction coefficient for each of a plurality of light sources. The shading correction coefficient is held as a shading characteristic or a shading correction table. The shading correction coefficient is set for each RGB and is applied for each RGB.

  In FIG. 6, tungsten (Tangsten), halogen (Halogen), sunlight (Sunlight Sunny), and a fluorescent lamp (Fluorescent) are shown as a plurality of light sources. In FIG. 6, the shading characteristics of each light source are shown on the left side, and the shading correction table (shading correction table) is shown on the right side. In each characteristic and correction table, the center value indicates the value at the center of the captured image, and the left and right peripheral values indicate the values at the peripheral portion of the captured image.

  The shading characteristics shown in FIG. 6 show a state in which the pixel value decreases for each RGB from the center to the periphery of the image. Further, the shading correction table shown in FIG. 6 is set as a characteristic for each RGB in order to compensate for a drop in the pixel value indicated by the shading characteristic. As shown in the shading characteristics of FIG. 6, since the amount of decrease in the pixel value from the central part to the peripheral part differs depending on the light source, the intensity of the correction table for compensating for this also differs depending on the light source. In particular, when the light source is close to near-infrared light, the drop in the pixel value from the central portion to the peripheral portion increases due to the change in the cutoff frequency described in FIG. On the other hand, since the fluorescent lamp shown in FIG. 6 does not contain much infrared light, the drop in the pixel value from the central portion to the peripheral portion is reduced.

  The shading characteristic or the shading correction table may be a coefficient corresponding to the distance from the center of the image, or may be in a two-dimensional bitmap (BitMap) format of an appropriate size. Although the shading characteristics or the shading correction table shows one-dimensional characteristics in FIG. 6, a shading characteristic or a shading correction table suitable for a two-dimensional captured image region is preferable. Or you may interpolate the whole area | region of a captured image by multiplying each one-dimensional characteristic of a vertical direction and a horizontal direction. The shading characteristics or the shading correction table can be created by measuring the shading characteristics from an image obtained by photographing a uniform gray chart (Gray Chart) or the like under a light source registered in advance.

  FIG. 7 is a diagram illustrating a result image obtained when shading correction is performed on a captured image. Here, the upper part of FIG. 7 shows the block statistical value of the captured image, the shading correction coefficient, and the block statistical value of the image after shading correction (result image) as an image. The lower part of FIG. 7 shows the block statistical value of the captured image, the shading correction coefficient, and the block statistical value of the image after shading correction as characteristic values.

  As shown in FIG. 7, each block statistical value is multiplied by a shading correction coefficient to obtain a block statistical value after shading correction. By multiplying the captured image by the shading correction coefficient, the block statistical value of the pixel value of the result image can be flattened from the center to the periphery of the image.

  As shown in FIG. 7, the captured image before being multiplied by the shading correction table has color unevenness due to shading as in FIG. By multiplying the captured image by the shading correction table, it is possible to obtain a shading correction image in which color unevenness due to shading is corrected.

Next, calculation of the R signal ratio (Rr) calculated in step S16 of FIG. 5 will be described. Here, the R signal ratio (Rr) of the corrected block statistical value is calculated by the following formula 1. In Equation 1, R, G, and B indicate block statistical values of the R signal, G signal, and B signal of the corrected block.
Rr = R / (R + G + B) (Formula 1)

  After calculating the R signal ratio of all the corrected block statistics, in step S18 in FIG. 5, the fluctuation of the R signal ratio of the block statistics continuous from the central portion to the peripheral portion of the entire image among the corrected block statistics. A value is calculated, and an average value of a plurality of fluctuation values is temporarily stored as an R signal deviation degree.

  FIG. 8 is a schematic diagram showing a setting example of the block groups UR, UL, V, and H for calculating the R signal deviation degree. Here, a case where the block statistical value is calculated for each block constituted by 16 × 16 blocks by the image block statistical circuit 116 is shown. As in UR, UL, V, and H shown in FIG. 8, the block group is set in a radial shape extending from the central portion to the peripheral portion around the central portion of the image. Then, the fluctuation value of the R signal ratio is obtained for each group of the block groups UR, UL, V, and H including the central portion of the image. The group classification of the block group is not limited to the pattern shown in FIG. 8, and may be a different pattern or the number of groups. The fluctuation value of the R signal ratio for each block group is averaged to obtain the R signal deviation degree.

The variation value ν of the R signal ratio calculated for each group can be calculated using the variance, standard deviation, etc. for each group as shown in the following equation 2. Here, x _i denotes the corrected block statistics of the i-th corrected block. In the example shown in FIG. 8, since each block group UR, UL, V, H is composed of 16 blocks, i = 16.

... (Formula 2)

  The obtained variation value of the R signal ratio for each group is averaged to obtain an R signal deviation degree s after applying the target shading correction. For example, ν1 represents the fluctuation value of the block group UR, and ν2 represents the fluctuation value of the block group UL.

... (Formula 3)

  Then, the plurality of shading correction coefficients stored in the shading correction coefficient storage unit 132 are sequentially read out, and steps S12 to S20 in FIG. 5 are repeated to calculate the R signal deviation degree s for all the shading correction coefficients. In the example of FIG. 6, since four shading correction coefficients are prepared corresponding to each light source, in this case, four R signal deviation degrees s are calculated.

  FIG. 9 is a schematic diagram showing a process in which an optimum shading correction coefficient is calculated from three shading correction coefficients (tables A, B, and C). As shown in FIG. 9, an R signal ratio is calculated for each block statistical value obtained by multiplying a captured image by a shading correction coefficient. The characteristic diagram shown at the bottom of FIG. 9 shows the R signal ratio of each block group UR, UL, V, H. Then, the average value of the fluctuation value of the R signal ratio of each block group UR, UL, V, H is calculated as the R signal deviation degree s, and the shading correction coefficient having the smallest R signal deviation degree is finally subjected to the shading correction. Select as a coefficient.

  In the example shown in FIG. 9, among the three shading correction coefficients, the variation of the R signal ratio calculated by the table B for each block group UR, UL, V, H is the smallest, and the R signal deviation degree of the table B s becomes the smallest. Accordingly, the color shading estimation unit 130 selects the table A as the shading correction coefficient, and the color shading correction unit 110d performs the shading correction by multiplying the captured image by the table A.

  In the process of obtaining the optimum shading correction coefficient, the color shading estimation unit 130 can simplify the process up to obtaining the shading correction coefficient by multiplying the block statistical value by each shading correction coefficient. On the other hand, in the correction of the captured image by the color shading correction unit 110d, the corrected shading correction is performed with higher accuracy by multiplying the R, G, and B signal values of each pixel of the captured image by the shading correction coefficient. Images can be obtained. However, also in the correction of the captured image by the color shading correction unit 110d, the shading correction may be performed by multiplying the block statistical value by each shading correction coefficient.

  The RGB pixel values and block statistical values change according to the subject brightness, but by selecting the table A with the smallest R signal deviation degree s, the pixel values drop from the central part to the peripheral part. It is possible to reliably compensate. Further, by using the R signal ratio, even when the subject luminance changes for each block, it is possible to reliably determine the drop of the R signal and select an optimal table.

  In addition, the light source estimation unit 110f estimates a light source based on the shading correction coefficient. For example, when the shading correction coefficient of tungsten shown in FIG. 6 is selected as the optimum table, the light source estimation unit 110f estimates that the light source of the subject is tungsten. The light source estimation result by the light source estimation unit 110 is applied to white balance correction in the white balance correction unit 110c. As a result, it is possible to significantly increase the estimation accuracy of the light source, and it is possible to reliably suppress the decrease in the white balance correction accuracy caused by the erroneous estimation of the light source.

  In the above-described example, shading is obtained based on the R signal. However, it is also possible to obtain shading based on the G signal and the B signal by the same method.

  As described above, according to the present embodiment, the block statistical value obtained by dividing the captured image into blocks is calculated by multiplying the shading correction coefficient prepared for each light source in advance, and the R signal ratio is used to calculate from the center of the image. The R signal deviation degree of the R signal ratio to the peripheral part is calculated. Then, the shading correction of the captured image is performed using the shading correction coefficient that minimizes the R signal deviation degree. Therefore, since an optimal shading correction coefficient can be applied from among a plurality of shading correction coefficients based on the actual captured image, it is possible to accurately estimate the cause of color shading having different characteristics depending on the light source. Color shading can be corrected with much higher accuracy.

  The preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to such examples. It is obvious that a person having ordinary knowledge in the technical field to which the present invention pertains can come up with various changes or modifications within the scope of the technical idea described in the claims. Of course, it is understood that these also belong to the technical scope of the present invention.

100 Digital camera 110d Color shading correction unit (processing unit)
110f Light source estimation unit 116 Image block statistical circuit 116a Block division unit 116b Statistical value calculation unit 130 Color shading estimation unit 132 Shading correction coefficient storage unit

Claims (5)

A captured image acquisition unit for acquiring a captured image;
A color shading estimation unit that estimates color shading generated in the captured image based on a result of applying a plurality of different color shading corrections to the captured image;
A processing unit that performs processing based on the captured image based on the color shading estimated by the color shading estimation unit;
A color shading correction coefficient storage unit that stores a plurality of color shading correction coefficients optimized for each predetermined light source;
A block division unit for dividing the captured image into a plurality of blocks;
A statistical value calculation unit that calculates a block statistical value by statistically calculating a pixel value for each block, and
The color shading estimation unit
After the block statistic value is multiplied by the color shading correction coefficient, a ratio of the R signal to the R signal, the G signal, and the B signal is calculated for each block. An image processing apparatus, wherein a color shading correction coefficient that minimizes a variation in the ratio of the R signal in a block group continuous toward a peripheral portion is selected .   The color shading estimation unit
  A color corresponding to the smallest one of all the R signal deviation degrees is calculated by calculating the R signal deviation degree based on the fluctuation of the ratio of the R signals in the block group continuous from the central part of the image toward the peripheral part. The image processing apparatus according to claim 1, wherein a shading correction coefficient is selected.   The R signal shift degree is calculated using a variance value or a standard deviation of a ratio of the R signals in a block group continuous from the central part of the image toward the peripheral part. Image processing device.   The image processing apparatus according to claim 1, wherein the processing unit corrects color shading of the captured image based on a color shading correction coefficient selected by the color shading estimation unit. .   The processing unit is configured to estimate a light source by determining whether the light source includes a light beam having a wavelength near near infrared light based on the color shading correction coefficient selected by the color shading estimation unit. The image processing apparatus according to claim 1.