JP4947624B2 - White balance adjustment method - Google Patents

Description

  The present invention relates to a white balance adjustment method, and more particularly to an auto white balance adjustment function mounted on a digital still camera, a digital video camera, or the like.

  Conventionally, white balance adjustment using a gray world algorithm is known (for example, Patent Document 1 below). The gray world algorithm is an algorithm that presupposes the statistical fact that if all colors in the screen are averaged, it becomes close to an achromatic color, and R (red), G (green), and B (blue) in the screen are assumed. The white balance gain is set so that the average signal level of each color becomes equal.

JP 2002-185977 A

  However, in general, according to the white balance adjustment method using the gray world algorithm, chromatic subjects occupy a wide area in the shooting screen, such as landscape shooting including blue sky, macro shooting with close-up of fruits and flowers, etc. When shooting is performed, the white balance adjustment may not be performed correctly due to the strong influence of the color of the chromatic subject.

  The present invention has been made in view of such circumstances, and even when a chromatic single-color region (hereinafter referred to as a “chromatic single-color region”) is included in a wide range in a shooting screen, the chromatic single color is included. It is an object of the present invention to obtain a white balance adjustment method capable of excluding the influence of the region and performing a good white balance adjustment.

The white balance adjustment method according to the first invention includes (A) a step of defining a plurality of evaluation areas in the image, and (B) for each of the plurality of evaluation areas, the evaluation area is a chromatic monochrome region. and whether the determining whether, (C) If a part of the evaluation area of the plurality of evaluation area is determined to be a chromatic monochromatic area, among the plurality of evaluation areas, Arisa Adjusting the white balance of the image using a gray world algorithm in which only the evaluation area determined to be a single color area is excluded from the calculation target.

  The white balance adjustment method according to the second invention is the white balance adjustment method according to the first invention, particularly when (D) it is determined in advance that all of the plurality of evaluation areas are chromatic monochromatic regions. The method further includes the step of adjusting the white balance of the image using a set default value.

  The white balance adjustment method according to a third aspect of the invention is the white balance adjustment method according to the first or second aspect of the invention. In particular, the step (B) includes (B-1) dividing the evaluation area into a plurality of small areas. (B-2) obtaining a color vector for each of the plurality of small areas; and (B-3) calculating an average value and uniformity of the plurality of color vectors obtained for the plurality of small areas. And determining whether or not the evaluation area is a chromatic single-color region.

  The white balance adjustment method according to a fourth aspect of the present invention is the white balance adjustment method according to the first or second aspect of the invention. In particular, the step (B) includes (B-1) dividing the evaluation area into a plurality of small areas. (B-2) obtaining a color vector and a luminance value for each of the plurality of small areas, and (B-3) an average value of the plurality of color vectors obtained for the plurality of small areas. And determining whether or not the evaluation area is a chromatic monochrome region based on the uniformity and the uniformity of a plurality of luminance values.

  The white balance adjustment method according to a fifth aspect of the present invention is the white balance adjustment method according to the third or fourth aspect of the invention. In particular, the step (B-3) relates to (B-3-1) the plurality of color vectors. A step of obtaining an average color vector; (B-3-2) a step of obtaining a difference value between the average color vector and each of the plurality of color vectors; and (B-3-3) the step (B-3- A step of obtaining a mean square value of the plurality of difference values obtained in 2); and (B-3-4) a vector length of the mean color vector is greater than a predetermined first threshold value, and Determining that the evaluation area is a chromatic monochrome region when the mean square value is equal to or less than a predetermined second threshold value.

  The white balance adjustment method according to the sixth aspect of the invention is the white balance adjustment method according to the fifth aspect of the invention. In particular, in the step (B-3-3), all the values obtained in the step (B-3-2) are obtained. Among the difference values, the difference value of the upper predetermined ratio having a large value is excluded from the calculation target, and the root mean square value is obtained.

  The white balance adjustment method according to a seventh aspect of the invention is the white balance adjustment method according to any one of the first to sixth aspects of the invention, particularly in the step (C), the evaluation area itself and the vicinity of the evaluation area. The region is excluded from the calculation target of the gray world algorithm.

  The white balance adjustment method according to an eighth invention is the white balance adjustment method according to any one of the first to seventh inventions, and in particular, the step (A) includes (A-1) the horizontal direction of the image and Defining a plurality of evaluation areas arranged in the vertical direction, and (A-2) defining an evaluation area located in a central area of the image and an evaluation area located in a peripheral area of the image. The steps (B) and (C) are executed in order corresponding to the steps (A-1) and (A-2), respectively.

  A white balance adjustment method according to a ninth aspect of the invention is the white balance adjustment method according to any one of the first to seventh aspects of the invention, in particular, in the step (A), an evaluation area located in a central area of the image, A plurality of evaluation areas arranged in a horizontal direction and a vertical direction of the image are defined in a peripheral region of the image.

  The white balance adjustment method according to the tenth invention is the white balance adjustment method according to any one of the first to tenth inventions, particularly using a gray world algorithm in which the evaluation area of the chromatic monochrome region is excluded from the calculation target. White balance adjustment and white balance adjustment using the gray world algorithm that does not exclude the evaluation area of the chromatic single-color area from the calculation target are temporally continuous in any order, from the adjustment amount in the previous white balance adjustment The transition to the adjustment amount in the subsequent white balance adjustment is performed in stages.

  According to the white balance adjustment method according to the first aspect of the invention, the white balance adjustment of the image is performed using the gray world algorithm in which the evaluation area of the chromatic monochrome region is excluded from the calculation target. Therefore, even when the chromatic single-color region is included in a wide range in the shooting screen, it is possible to eliminate the influence of the chromatic single-color region and perform good white balance adjustment.

  According to the white balance adjustment method according to the second aspect of the invention, when all the evaluation areas are chromatic monochromatic areas, the preset default value is used instead of the white balance adjustment using the gray world algorithm. The white balance adjustment was made. Therefore, by setting an appropriate default value according to the type of light source such as sunlight or fluorescent light, even when all the evaluation areas are chromatic monochromatic areas, good white balance adjustment is possible. It can be carried out.

  According to the white balance adjustment method according to the third aspect of the present invention, whether or not the evaluation area is a chromatic monochrome region is determined easily and with high accuracy by a determination method based on the average value and uniformity of a plurality of color vectors. can do.

  According to the white balance adjustment method of the fourth aspect of the invention, whether or not the evaluation area is a chromatic monochromatic region by a determination method based on the average value and uniformity of a plurality of color vectors and the uniformity of a plurality of luminance values. Can be determined easily and with high accuracy.

  According to the white balance adjustment method according to the fifth aspect of the present invention, it is possible to determine whether or not the evaluation area is a chromatic monochrome area by a simple calculation of obtaining a difference value or a root mean square value.

  According to the white balance adjustment method of the sixth aspect of the invention, it is possible to eliminate the influence of noise and the influence caused by a rare color vector corresponding to the noise, so that good white balance adjustment can be performed.

  According to the white balance adjustment method according to the seventh aspect of the invention, not only the evaluation area itself of the chromatic monochrome area but also the neighboring area of the evaluation area is excluded from the calculation object of the gray world algorithm, thereby protruding from the evaluation area. It is possible to eliminate the influence of the chromatic monochromatic region in the portion where the image is present.

  According to the white balance adjustment method of the eighth invention, by specifying the evaluation area with a plurality of patterns, the evaluation area of the chromatic monochrome area is specified regardless of the position of the chromatic monochrome object in the shooting screen. can do.

  According to the white balance adjustment method according to the ninth aspect of the present invention, by defining the evaluation area in the screen center area that is likely to include the main subject, and by defining a plurality of evaluation areas in the screen peripheral area, The evaluation area of the chromatic / monochromatic area can be specified regardless of the position of the chromatic / monochromatic subject in the shooting screen.

  According to the white balance adjustment method according to the tenth aspect of the present invention, it is possible to prevent screen flickering caused by frequent changes in the white balance adjustment amount.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In addition, the element which attached | subjected the same code | symbol in different drawing shall show the same or corresponding element.

  FIG. 1 is a block diagram schematically showing the overall configuration of a digital camera according to an embodiment of the present invention. An analog signal processing circuit 2 is connected to the subsequent stage of the imaging unit 1. An SPU (Sensor Processing Unit) 3 a is connected to the subsequent stage of the analog signal processing circuit 2. An RPU (Real-time Processing Unit) 4 is connected to the subsequent stage of the SPU 3a. The SPU 3 a and RPU 4 are connected to the main bus 5. In addition to the SPU 3 a and RPU 4, a DMA controller 6 and a memory interface 7 are connected to the main bus 5. A memory 8 such as an SDRAM is connected to the memory interface 7.

  The imaging unit 1 has an imaging element (image sensor) such as a CCD. In addition, the imaging unit 1 includes three color filters such as R (red), G (green), and B (blue), and images obtained using an optical lens are captured. The signal S1 is output. Note that another image pickup device such as a CMOS image sensor may be provided instead of the CCD. Instead of the three-color color filters, four-color color filters such as Y (yellow), M (magenta), C (cyan), and W (white) may be provided.

  The analog signal processing circuit 2 performs signal processing such as noise removal, signal amplification, and A / D conversion on the analog image signal S1 input from the imaging unit 1, and outputs a digital image signal D1.

  The SPU 3a performs various types of signal processing related to the problems caused by the characteristics of the image sensor on the image signal D1 input from the analog signal processing circuit 2, and outputs the image signal D2. In particular, the SPU 3a includes an auto white balance processing unit (hereinafter referred to as “AWB processing unit”) 9 that performs white balance adjustment on the image signal D1.

  The RPU 4 performs various types of signal processing such as pixel interpolation, color space conversion, edge enhancement, and false color suppression on the image signal D2 input from the SPU 3a, and outputs an image signal D3. The image signal D3 output from the RPU 4 is stored in the memory 8 via the main bus 5 and the memory interface 7 under the control of the DMA controller 6.

  Between the SPU 3 a and the memory 8, the image signal D 1 can be mutually transmitted via the main bus 5 and the memory interface 7 under the control of the DMA controller 6.

  FIG. 2 is a flowchart showing a flow of white balance adjustment processing by the AWB processing unit 9 shown in FIG. In this example, the image signal D1 before white balance adjustment is temporarily stored in the memory 8, and the AWB processing unit 9 reads the image signal D1 from the memory 8 for each frame and performs white balance adjustment processing.

  First, in step SP1, the AWB processing unit 9 defines a plurality of evaluation areas in the image represented by the image signal D1 read from the memory 8. FIG. 3 is a diagram illustrating an example in which four evaluation areas A1 to A4 are defined in the image 10. The evaluation areas A1, A2, A3, and A4 are respectively defined in an upper left area, an upper right area, a lower left area, and a lower right area obtained by dividing the image 10 into four equal parts. Evaluation areas A1 and A2 are arranged in the horizontal direction of the image 10, and similarly, evaluation areas A3 and A4 are arranged in the horizontal direction of the image 10. The evaluation areas A1 and A3 are arranged in the vertical direction of the image 10, and similarly, the evaluation areas A2 and A4 are arranged in the vertical direction of the image 10.

  Referring to FIG. 2, next, in step SP2, the AWB processing unit 9 determines whether or not each of the evaluation areas A1 to A4 is a chromatic monochrome region. That is, it is determined whether or not each of the evaluation areas A1 to A4 is an evaluation area composed of a chromatic and monochromatic image area (or an image area close thereto). A specific method for such determination will be described later. FIG. 4 is a diagram illustrating a first example of the image 10. The blue sky 41 is reflected in the background of the mountain 40. In the case of the first example, since only the blue color of the blue sky 41 is included in the evaluation areas A1 and A2, it is determined that the evaluation areas A1 and A2 are chromatic monochrome regions.

  Referring to FIG. 2, when a chromatic single color area exists in evaluation areas A1 to A4 (that is, when the determination result in step SP2 is “YES”), in step SP3, AWB processing unit 9 It is determined whether or not all of the evaluation areas A1 to A4 are chromatic monochromatic areas.

  If a part of the evaluation areas A1 to A4 is a chromatic monochrome region (that is, if the determination result in step SP3 is “NO”), then in step SP4, the AWB processing unit 9 evaluates the chromatic monochrome region. The value of the white balance gain is determined by the gray world algorithm excluding the area from the calculation target, and the white balance of the image 10 is adjusted.

  FIG. 5 is a diagram showing image areas to be excluded from the calculation target of the gray world algorithm, corresponding to the first example shown in FIG. An image area B1 (shaded hatched portion) including the evaluation area A1 itself and the vicinity area of the evaluation area A1 is excluded from the calculation target. Similarly, the image area B1 (scattered hatched portion) including the evaluation area A2 itself and the vicinity area of the evaluation area A2 is excluded from the calculation target. Therefore, the AWB processing unit 9 determines the value of the white balance gain based on the image signal corresponding to the lower half area of the image 10. In this way, by excluding the area wider than the evaluation areas A1 and A2 from the calculation target, the influence of the chromatic monochrome area (the blue sky 41 in the first example) of the portion that protrudes from the evaluation areas A1 and A2 can be obtained. Can be eliminated.

  Referring to FIG. 2, when all of evaluation areas A1 to A4 are chromatic monochromatic regions in step SP3 (that is, when the determination result in step SP3 is “YES”), in step SP7, an AWB process is performed. The unit 9 performs white balance adjustment of the image 10 using a predetermined default value. As the default value, a white balance gain value corresponding to the type of light source such as sunlight or fluorescent lamp is set in advance. The AWB processing unit 9 calculates an average luminance value of the image 10, and when the average luminance value is equal to or greater than a predetermined threshold, the white balance is set using the white balance gain value set for sunlight. Make adjustments. On the other hand, when the average luminance value of the image 10 is less than the predetermined threshold value, white balance adjustment is performed using the value of the white balance gain set for the fluorescent lamp. By appropriately setting each white balance gain value, even if all evaluation areas are chromatic single-color areas, it is possible to perform good white balance adjustments regardless of the gray world algorithm. It becomes.

  In step SP2, if there is no chromatic single color region in the evaluation areas A1 to A4 (that is, if the determination result in step SP2 is “NO”), then in step SP5, the AWB processing unit 9 performs the processing shown in FIG. An evaluation area of a prescribed pattern different from the prescribed patterns of the evaluation areas A1 to A4 shown is defined in the image 10. FIG. 6 is a diagram showing an example in which two evaluation areas A5 and A6 are defined in the image 10. As shown in FIG. The evaluation area A5 is defined as a central area including the center of the image 10. The evaluation area A6 is defined in the peripheral area of the image 10 so as to surround the periphery of the evaluation area A5.

  Referring to FIG. 2, next, in step SP6, AWB processing unit 9 determines whether or not each of evaluation areas A5 and A6 is a chromatic monochrome region. FIG. 7 is a diagram illustrating a second example of the image 10. An apple 50 is placed on a desk 51 placed in contact with the wall 52. In the case of the second example, since only the red color of the apple 50 is included in the evaluation area A5, it is determined that the evaluation area A5 is a chromatic monochrome region. FIG. 8 is a diagram illustrating a third example of the image 10. The blue sky 61 is reflected in the background of the hot air balloon 60. In the case of the third example, since only the blue color of the blue sky 61 is included in the evaluation area A6, it is determined that the evaluation area A6 is a chromatic monochrome region. In this way, by defining the evaluation area with a plurality of prescribed patterns, it is possible to specify the evaluation area of the chromatic monochrome region regardless of the position of the chromatic monochrome subject in the shooting screen.

  Referring to FIG. 2, if a chromatic single color area exists in evaluation areas A5 and A6 in step SP6 (that is, if the determination result in step SP6 is “YES”), then in step SP3, an AWB process is performed. The unit 9 determines whether or not all of the evaluation areas A5 and A6 are chromatic monochromatic areas.

  When all of the evaluation areas A5 and A6 are chromatic monochromatic areas (that is, when the determination result in step SP3 is “YES”), in the next step SP7, the AWB processing unit 9 determines in advance as described above. The white balance adjustment of the image 10 is performed using the default value. On the other hand, when a part of the evaluation areas A5 and A6 is a chromatic monochrome region (that is, when the determination result in step SP3 is “NO”), in the next step SP4, the AWB processing unit 9 The white balance gain value is determined by the gray world algorithm excluding the evaluation area of the chromatic monochrome region from the calculation target, and the white balance of the image 10 is adjusted.

  FIG. 9 is a diagram showing image regions to be excluded from the calculation target of the gray world algorithm, corresponding to the second example shown in FIG. The image area B5 (shaded hatched portion) including the evaluation area A5 itself and the vicinity area of the evaluation area A5 is excluded from the calculation target. Accordingly, the AWB processing unit 9 determines the value of the white balance gain based on the image signal corresponding to the remaining image area obtained by removing the image area B5 from the image 10. FIG. 10 is a diagram showing image areas to be excluded from the calculation target of the gray world algorithm, corresponding to the third example shown in FIG. An image area B6 (shaded hatched portion) including the evaluation area A6 itself and the vicinity area of the evaluation area A6 is excluded from the calculation target. Accordingly, the AWB processing unit 9 determines the value of the white balance gain based on the image signal corresponding to the remaining image area obtained by removing the image area B6 from the image 10.

  Referring to FIG. 2, if there is no chromatic single color region in evaluation areas A5 and A6 in step SP6 (that is, if the determination result in step SP6 is “NO”), then in step SP8, an AWB process is performed. The unit 9 performs white balance adjustment by a normal gray world algorithm. In other words, the white balance gain value is determined by calculation for the entire area of the image 10 and white balance adjustment is performed.

  In the above description, an example in which there are two types of defined patterns in the evaluation area has been described, but there may be three or more defined patterns. In addition, an example has been described in which re-determination is performed using the second defined pattern (FIG. 6) when a chromatic monochrome region is not found in the determination using the first defined pattern (FIG. 3). Even if a chromatic monochromatic region is found by the determination using the first specified pattern, the determination using the second specified pattern may be performed.

  On the contrary, it is also possible to perform only one determination using one type of specified pattern instead of performing the determination twice using two types of specified patterns. FIG. 11 is a diagram showing an example in which five evaluation areas A7 to A11 are defined in the image 10. As shown in FIG. The evaluation area A7 is defined as a central region including the center of the image 10. The evaluation areas A8 to A11 are defined in the peripheral area of the image 10 so as to surround the periphery of the evaluation area A7. Evaluation areas A8, A9, A10, and A11 are respectively defined in an upper left area, an upper right area, a lower left area, and a lower right area obtained by dividing the image 10 into four equal parts. The evaluation areas A8 and A8 are arranged in the horizontal direction of the image 10, and similarly, the evaluation areas A10 and A11 are arranged in the horizontal direction of the image 10. The evaluation areas A8 and A10 are arranged in the vertical direction of the image 10, and similarly, the evaluation areas A9 and A11 are arranged in the vertical direction of the image 10. As described above, by defining the evaluation area A7 in the center area of the screen where the main subject is likely to be included, and by defining the plurality of evaluation areas A8 to A11 in the peripheral area of the screen, the chromaticity in the shooting screen is displayed. Regardless of the position of the monochrome subject, the evaluation area of the chromatic monochrome region can be specified by one determination.

  Now, a process for determining whether or not each of the evaluation areas A1 to A11 (hereinafter collectively referred to as “evaluation area A”) is a chromatic monochrome region will be described in detail.

  FIG. 12 is a flowchart showing a flow of determination processing for determining whether or not each evaluation area A is a chromatic monochrome region by the AWB processing unit 9 shown in FIG. First, in step SP20, the AWB processing unit 9 divides each evaluation area A into a plurality of small areas 11 as shown in FIGS. 3, 6, and 11 illustrate an example in which not only the evaluation area A but the entire area of the image 10 is divided into small areas 11, image areas other than the evaluation area A are not necessarily included in the small area 11. There is no need to split.

  FIG. 13 is an enlarged view showing one small area 11. In the example shown in FIG. 13, a total of 64 pixels of 8 rows × 8 columns are included in one small area 11. However, the number of pixels included in one small area 11 is not limited to this.

Referring to FIG. 12, next, in step SP21, AWB processing unit 9 obtains the color vectors of a plurality of small areas 11 included in each evaluation area A, respectively. Specifically, referring to FIG. 13, the average pixel values R _AVE , G _AVE , B _AVE are obtained by averaging a plurality of pixel values of R, G, B colors included in the small area 11 for each color. Is calculated. Next, the Y value, U value, and V value are calculated by converting the average pixel values R _AVE , G _AVE , and B _AVE into the YUV color space. In the coordinate system with the U value as the vertical axis and the V value as the horizontal axis, the color vector of the small area 11 is a vector connecting the origin of the coordinates and the point where the U value and V value calculated above are plotted. (In other words, a color difference signal) is obtained. In addition to the YUV color space, any color space may be used as long as it is a color space that can express a color vector in a two-dimensional or higher coordinate system, such as an R / G, B / G gain space. .

  Referring to FIG. 12, in step SP22, the AWB processing unit 9 averages the plurality of color vectors obtained in step SP21 for each evaluation area A, thereby obtaining an average color vector for each evaluation area A. calculate.

In step SP23, AWB processing unit 9, the vector length of the mean color vector (saturation in other words) is, whether a predetermined threshold value TH ₁ below a predetermined, each evaluation area A Judge every time. The threshold value TH ₁ can be determined by an experiment or the like.

A short average color vector length (that is, low saturation) means that the average color vector is close to gray. Accordingly, if such an evaluation area A is excluded from the calculation target of the gray world algorithm, accurate white balance adjustment cannot be expected. Therefore, when the average color vector is equal to or less than the threshold value TH ₁ (that is, when the determination result in step SP23 is “YES”), the AWB processing unit 9 determines that the evaluation area A is not a chromatic monochrome region. To do. Thereby, the situation where the evaluation area A is excluded from the calculation target of the gray world algorithm is avoided.

When the vector length of the average color vector is larger than the threshold value TH ₁ (that is, when the determination result in step SP23 is “NO”), in step SP24, the AWB processing unit 9 performs the following for each evaluation area A. A difference value between the average color vector and each of the plurality of color vectors related to the plurality of small areas 11 is calculated.

  Next, in step SP25, the AWB processing unit 9 specifies, for each evaluation area A, a difference value of a higher predetermined ratio (for example, higher 5%) having a larger value among the plurality of difference values obtained in step SP24. . The identified difference value is excluded from the calculation target in the subsequent step SP26. As a result, the influence of noise and the influence caused by a rare color vector corresponding to the noise can be eliminated, so that good white balance adjustment can be performed. However, instead of excluding the upper predetermined ratio difference value, a difference value larger than a predetermined threshold value may be excluded.

  Next, in step SP26, for each evaluation area A, the AWB processing unit 9 squares the remaining difference values obtained by excluding the upper predetermined ratio difference values obtained in step SP25 from the plurality of difference values obtained in step SP24. An average value (Root Means Square: RMS) is calculated. Here, the standard deviation may be obtained instead of the mean square value, but since it has already been normalized in the form of the difference value from the average color vector, it is not necessary to perform complicated standard deviation calculations. It is sufficient to calculate a mean square value.

In step SP27, AWB processing unit 9, mean square values calculated in step SP26 is, whether a predetermined threshold value TH ₂ below a predetermined determines for each evaluation area A. The threshold value TH ₂ can be determined by an experiment or the like.

A large mean square value means that the uniformity of a plurality of color vectors related to the evaluation area A is low, that is, a plurality of colors exist in the evaluation area A. Therefore, when the mean square value is larger than the threshold value TH ₂ (that is, when the determination result in step SP27 is “NO”), the AWB processing unit 9 determines that the evaluation area A is not a chromatic monochrome region. .

On the other hand, when the mean square value is equal to or less than the threshold value TH ₂ (that is, when the determination result in step SP27 is “YES”), the AWB processing unit 9 determines that the evaluation area A is a chromatic monochrome region. To do.

In the above description, whether or not each evaluation area A is a chromatic monochrome region is determined based on the average value and the uniformity of the plurality of color vectors obtained for the plurality of small areas 11. In addition to the U value and the V value shown in FIG. 6, a Y value (luminance value) can also be used for the determination. That is, in addition to the determination based on the color vector described above, the luminance value for each small area 11 is calculated as an average value of the luminance values of a plurality of pixels in step SP21, and the luminance values of the plurality of small areas 11 are calculated in step SP22. An average luminance value for each evaluation area A is calculated as an average value, a difference value between the average luminance value and the luminance value of each small area 11 is calculated in step SP24, and the upper predetermined ratio difference value is excluded in step SP25. The mean square value is calculated in step SP26. In step SP27, when the mean square value of the color vector is equal to or less than the threshold value TH ₂ and the mean square value of the luminance is equal to or less than a predetermined threshold value TH ₃ , the AWB processing unit 9 Determines that the evaluation area A is a chromatic monochrome region. The threshold value TH ₃ can be determined by an experiment or the like. As described above, whether or not each evaluation area A is a chromatic monochrome region is determined easily and with high accuracy by a determination method based on the average value and uniformity of a plurality of color vectors and the uniformity of a plurality of luminance values. can do.

In the above description, it is assumed that the threshold value TH ₁ and the threshold value TH ₃ are fixed values, but the values of the threshold value TH ₁ and the threshold value TH ₃ are changed according to the situation. This is also effective for performing more accurate white balance. For example, when the average color vector has a large vector length, it is predicted that the variation in the color vector also increases. Therefore, the probability that a chromatic single color can be detected increases when the threshold value TH ₃ is set large. May change the value of the stepwise threshold TH ₃ in accordance with the vector length of the mean color vector, it may change the value of the continuous threshold TH ₃ in accordance with the vector length of the average color vector . In addition, when the vector length of the average color vector is very large, it is also effective to set the threshold value TH ₃ to zero and detect a chromatic single color using only the average color vector as a judgment material.

  FIG. 14 is a diagram illustrating a situation in which the white balance gain is set for a plurality of consecutive frames. In the white balance adjustment method according to the present embodiment, as shown in FIG. 2, AWB adjustment (step SP4) by the gray world algorithm in which the evaluation area of the chromatic monochrome region is excluded from the calculation target, and white using the default value Three types of AWB adjustment methods are prepared: balance adjustment (step SP7) and AWB adjustment (step SP8) using a normal gray world algorithm, and an appropriate adjustment method is selected according to the image.

  In FIG. 14, the white balance gain G1 is set by the AWB adjustment in step SP4 before the n-1th frame, and the white balance gain G4 should be set by the AWB adjustment in step SP8 after the nth frame. . However, the type and order of the AWB adjustment methods are not limited to this example, and are arbitrary.

  The presence / absence, size, position, and the like of a chromatic single-colored object in the shooting screen are changed by the framing operation by the photographer. Accordingly, it is assumed that the AWB adjustment method is switched every time the framing is changed, and the value of the white balance gain is changed accordingly. In the example shown in FIG. 14A, a transition from a small white balance gain G1 to a large white balance gain G4 is performed between adjacent frames. Therefore, if the AWB adjustment method is frequently switched, the screen may flicker due to a change in white balance gain.

  In contrast, in the example shown in FIG. 14B, the transition from the white balance gain G1 to the white balance gain G4 is performed by multiplying a plurality of frames (three frames in this example). That is, the AWB processing unit 9 shown in FIG. 1 performs the AWB adjustment (step SP4 in FIG. 2) by the gray world algorithm in which the evaluation area of the chromatic monochrome region is excluded from the calculation target, and other AWB adjustments (in FIG. 2). In a situation where steps SP7 and SP8) are temporally continuous in an arbitrary order, the transition from the gain G1 in the previous AWB adjustment to the gain G4 in the subsequent AWB adjustment is performed in stages over a plurality of frames. Accordingly, it is possible to prevent or suppress flickering of the screen due to frequent changes in the white balance adjustment amount.

  Thus, according to the white balance adjustment method according to the present embodiment, as shown in FIG. 2, some of the evaluation areas A among the plurality of evaluation areas A defined in the image 10 are chromatic monochromatic regions. Is determined, the white balance adjustment of the image 10 is performed using a gray world algorithm in which the evaluation area of the chromatic monochrome region is excluded from the calculation target. Therefore, even when the chromatic single-color region is included in a wide range in the shooting screen, it is possible to eliminate the influence of the chromatic single-color region and perform good white balance adjustment.

FIG. 15 is a block diagram showing a modification of the digital camera according to the present embodiment, corresponding to FIG. In FIG. 1, the image signal D1 before white balance adjustment is temporarily stored in the memory 8, and the AWB processing unit 9 reads the image signal D1 from the memory 8 for each frame and performs white balance adjustment. On the other hand, in FIG. 15, an SPU 3b having an AWB processing unit 9 and an average value calculation circuit 10 is provided instead of the SPU 3a shown in FIG. The average value calculation circuit 10 divides the image 10 represented by the image signal D1 input from the analog signal processing circuit 2 into a plurality of small areas 11, and, as in FIG. Average pixel values R _AVE , G _AVE , and B _AVE are calculated by averaging a plurality of pixel values for each color of G and B for each color. Then, data D4 related to the average pixel values R _AVE , G _AVE , and B _AVE for each small area 11 is stored in the memory 8 via the main bus 5 and the memory interface 7. The AWB processing unit 9 reads the data D4 from the memory 8 for each frame, and thereafter performs white balance adjustment by the same method as described above. Thereby, the amount of data stored in the memory 8 can be reduced. Note that the data D4 may be stored not in the memory 8 but in a register or the like connected to the SPU 3b.

1 is a block diagram schematically showing an overall configuration of a digital camera according to an embodiment of the present invention. 3 is a flowchart showing a flow of white balance adjustment processing by an AWB processing unit shown in FIG. 1. It is a figure which shows the example in which four evaluation areas were prescribed | regulated in the image. It is a figure which shows the 1st example of an image. FIG. 5 is a diagram showing image regions to be excluded from the calculation target of the gray world algorithm corresponding to the first example shown in FIG. 4. It is a figure which shows the example in which two evaluation areas were prescribed | regulated in the image. It is a figure which shows the 2nd example of an image. It is a figure which shows the 3rd example of an image. FIG. 8 is a diagram illustrating image regions to be excluded from the calculation target of the gray world algorithm, corresponding to the second example illustrated in FIG. 7. FIG. 9 is a diagram illustrating image regions to be excluded from the calculation target of the gray world algorithm corresponding to the third example illustrated in FIG. 8. It is a figure which shows the example in which five evaluation areas were prescribed | regulated in the image. It is a flowchart which shows the flow of the determination process whether each evaluation area is a chromatic monochrome area | region. It is a figure which expands and shows one small area. It is a figure which shows the condition which sets a white balance gain regarding several continuous frames. FIG. 6 is a block diagram showing a modification of the digital camera according to the embodiment of the present invention corresponding to FIG. 1.

Explanation of symbols

3a, 3b SPU
9 AWB processing unit 10 Image 11 Small area A1 to A11 Evaluation area B1, B2, B5, B6 Image area

Claims (10)

(A) defining a plurality of evaluation areas in the image;
(B) for each of the plurality of evaluation areas, determining whether the evaluation area is a chromatic monochrome region;
(C) When it is determined that some of the plurality of evaluation areas are chromatic monochromatic areas, the evaluation area determined to be a chromatic monochromatic area among the plurality of evaluation areas. Adjusting the white balance of the image using a gray world algorithm in which only the image is excluded from the calculation target. (D) The method further includes the step of adjusting the white balance of the image using a preset default value when it is determined that all of the plurality of evaluation areas are chromatic single-color regions. The white balance adjustment method described in 1. The step (B)
(B-1) dividing the evaluation area into a plurality of small areas;
(B-2) obtaining a color vector for each of the plurality of small areas;
(B-3) determining whether or not the evaluation area is a chromatic monochrome region based on an average value and uniformity of a plurality of color vectors obtained with respect to the plurality of small areas. The white balance adjustment method according to 1 or 2. The step (B)
(B-1) dividing the evaluation area into a plurality of small areas;
(B-2) obtaining a color vector and a luminance value for each of the plurality of small areas;
(B-3) Whether or not the evaluation area is a chromatic monochromatic area based on the average value and uniformity of a plurality of color vectors and the uniformity of a plurality of luminance values obtained for the plurality of small areas. The white balance adjustment method according to claim 1, further comprising a step of determining. The step (B-3)
(B-3-1) obtaining an average color vector for the plurality of color vectors;
(B-3-2) obtaining a difference value between the average color vector and each of the plurality of color vectors;
(B-3-3) obtaining a mean square value of a plurality of difference values obtained in the step (B-3-2);
(B-3-4) The vector length of the average color vector is larger than a predetermined first threshold value, and the mean square value is equal to or smaller than a predetermined second threshold value. The white balance adjustment method according to claim 3, further comprising: determining that the evaluation area is a chromatic monochrome region.   In the step (B-3-3), out of all the difference values obtained in the step (B-3-2), the difference value of the upper predetermined ratio having a large value is excluded from the calculation target, The white balance adjustment method according to claim 5, wherein a mean square value is obtained.   The white balance adjustment method according to any one of claims 1 to 6, wherein in the step (C), the evaluation area itself and a neighboring region of the evaluation area are excluded from a calculation target of a gray world algorithm. . The step (A)
(A-1) defining a plurality of evaluation areas arranged in the horizontal and vertical directions of the image;
(A-2) defining an evaluation area located in a central area of the image and an evaluation area located in a peripheral area of the image;
The white balance according to any one of claims 1 to 7, wherein the steps (B) and (C) are sequentially executed corresponding to the steps (A-1) and (A-2). Adjustment method.   In the step (A), an evaluation area located in a central region of the image and a plurality of evaluation areas arranged in a horizontal direction and a vertical direction of the image in a peripheral region of the image are defined. 8. The white balance adjustment method according to any one of 7 above. The white balance adjustment using the gray world algorithm that excludes the evaluation area of the chromatic monochrome area from the calculation target and the white balance adjustment using the gray world algorithm that does not exclude the evaluation area of the chromatic monochrome area from the calculation target are optional. The transition from the adjustment amount in the previous white balance adjustment to the adjustment amount in the subsequent white balance adjustment is performed in a stepwise manner in the case of continuous in time in order. White balance adjustment method.

Images