JP2003223636A - Image processor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processor capable of improving the tone of an image photographed under restricted exposure conditions according to utilizations and features of an image input device. <P>SOLUTION: A Y/C conversion part 18 converts R, G and B data of the inputted image into a brightness signal Y and chroma signals Cr and Cb. A Y histogram calculation part 20 produces a histogram of the brightness signal Y. A LUT (look-up table) calculation part 22 calculates a correction coefficient for flattening the histogram and calculates the LUT for correcting the brightness signal Y. The correction coefficient is adjusted by an intensity coefficient inputted from an intensity coefficient input part. The intensity coefficient is a coefficient for applying an inclination to the correction of the brightness signal Y and preset by the type, characteristics, an installation place, and the utilizations of the image input device 12. The intensity coefficient is set according to the system to be used so as to correct the tone according to the system. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, and more particularly to an image processing apparatus for correcting gradation of an image taken by a surveillance camera or the like.

[0002]

2. Description of the Related Art Conventionally, there has been proposed a technique for improving the gradation by converting the contrast of an image signal of a television camera, a monitor or the like.

For example, in Japanese Patent Laid-Open No. 7-73308, a histogram is created from input data, and the output of an interpolation circuit for converting the contrast of the input data using this histogram and the input data are added and averaged at a predetermined ratio. By doing so, a technique capable of performing optimum contrast conversion even in an image portion having a low histogram frequency is disclosed.

By the way, there is a system for displaying an image captured by an image input device such as a surveillance camera installed in a bank or a supermarket or an intercom camera installed at the entrance of a house on an image display device such as a CRT or LCD. is there.

The image input device as described above has a fixed installation position and cannot use auxiliary light such as a strobe or a light due to the nature of its function. Therefore, the aperture, exposure time, sensitivity, etc. are exposed. The condition is restricted, and the brightness of the subject varies greatly depending on the installation position, shooting time zone, and the like. Therefore, it may not be possible to take an image with appropriate brightness.

When an image photographed by such an image input device is displayed on the above-mentioned image display device, there are problems that the gradation is poor and the image is difficult to recognize. Also,
An image captured by such an image input device is disclosed in Japanese Patent Laid-Open No. 7-
If contrast conversion is performed by the technique described in Japanese Patent No. 3308, it is possible to maintain good gradation even in an image portion with a low histogram frequency. It is not possible to optimally improve the overall gradation.

The present invention has been made in consideration of the above facts, and improves the gradation of an image taken under the condition that the exposure condition is restricted, depending on the use and the nature of the image input device. An object of the present invention is to provide an image processing device capable of performing the above.

[0008]

In order to achieve the above-mentioned object, the invention according to claim 1 is to create a luminance data of each pixel of the photographed image from image data of the photographed image, and the luminance data. Histogram calculating means for calculating the histogram, correction coefficient calculating means for calculating the correction coefficient for correcting each of the luminance data so that the histogram is flattened based on the histogram, Input means for inputting an intensity coefficient for giving a tendency to correction, and conversion means for converting the luminance data based on the correction coefficient to which the tendency is given by the intensity coefficient inputted by the input means. It is characterized by having.

According to the present invention, the creating means creates the brightness data of each pixel of the taken image from the image data of the taken image, and the histogram calculating means calculates the histogram of the brightness data. That is, the appearance frequency (cumulative number of pixels) of the brightness data is calculated.

Then, the correction coefficient calculating means calculates, based on the histogram calculated by the histogram calculating means, a correction coefficient for correcting each of the luminance data so that the histogram is flattened. That is, when an extreme peak exists in the specific area of the histogram, the correction coefficient is calculated so that the brightness data near this peak is dispersed around. This makes it possible to correct the brightness data so that the bright part becomes dark and the dark part becomes bright.

The conversion means adjusts the correction coefficient with an intensity coefficient for giving a tendency to the correction of the brightness data input by the input means, and converts the brightness data based on the adjusted correction coefficient. This intensity coefficient may be predetermined and input according to the type and characteristics of the image input device that captured the captured image, the installation location, the application of the captured image, or may be stored in advance in the memory. Good.

As described above, since it is possible to give a tendency to the correction of the brightness data, that is, the gradation correction by the input intensity coefficient, by setting the intensity coefficient according to the system, the optimum intensity according to the system can be obtained. Gradation correction can be performed.

As described in claim 2, the intensity coefficient is low and the first intensity coefficient that gives a tendency to a correction coefficient for correcting the brightness data to be high and the brightness data are low. The correction coefficient for the correction may be composed of a second intensity coefficient that gives a tendency. As a result, correction to increase the brightness data,
That is, the correction for making the image brighter and the correction for making the luminance data lower, that is, the correction for making the image darker can be independently imparted, so that the system can be optimized more optimally. Gradation correction can be performed.

By the way, when there is a solid background (for example, blue sky) having a predetermined area or more in the photographed image, the luminance data becomes more than a predetermined frequency and noise may appear in the corrected image. is there. Therefore, as described in claim 3, the histogram further includes a reduction unit that reduces the frequency of the luminance data of a predetermined frequency or more, and the correction coefficient calculation unit includes the histogram after the reduction by the reduction unit. The correction coefficients may be calculated based on the above.

This makes it possible to prevent noise from appearing in the corrected solid background even when there is a solid background having a predetermined area or more in the captured image.

The histogram calculating means divides the photographed image into a plurality of areas, assigns a predetermined weight to each of the divided areas to the corresponding brightness data, and creates a histogram based on the weighted brightness data. It may be calculated. For example, when the probability that the subject is at the center of the captured image is high, the weight of the central portion of the captured image is increased and the weight of the peripheral portion is reduced. Thereby, the gradation correction can be performed more appropriately.

Further, a characteristic amount calculating means for calculating a characteristic amount indicating a degree of correction based on the correction coefficient to which the tendency is given, and noise for performing noise removal processing on the luminance data converted according to the characteristic amount. The removing means may be further provided.

For example, if the correction tendency is strong in the direction of making bright, noise may occur in the corrected image. Therefore, the feature amount is defined as a physical amount that represents the relative relationship between the degree of correction that brightens the image and the noise that occurs in the corrected image. Then, for example, the difference between the degree of correction for making the image brighter and the degree of correction for making the image darker is calculated as a feature amount, and as the feature amount becomes larger,
Increase the strength of noise removal. As a result, it is possible to prevent the deterioration of the image due to the correction.

Further, a chroma signal generating means for generating a chroma signal relating to the saturation of the photographed image from the image data, a cumulative histogram calculating means for calculating a chroma cumulative histogram from the generated chroma signal, and a cumulative histogram calculating means for calculating the chroma cumulative histogram. The configuration may further include a correction value calculation unit that calculates a correction value for correcting the chroma signal, and a correction unit that corrects the chroma signal based on the calculated correction value.

For example, the correction value calculating means judges from the calculated cumulative histogram whether or not the saturation is equal to or higher than a predetermined saturation, and when the saturation is equal to or higher than the predetermined saturation, the correction is performed so that the chroma signal becomes small. Calculate the value. Then, the chroma signal is corrected by the correction means based on the calculated correction value. As a result, the saturation of the captured image can be corrected.

[0021]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, an example of an embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 shows a schematic block diagram of the image processing system 10. The image processing system 10 includes an image input device 12, an image processing device 14, and an image output device 16.

The image input device 12 is, for example, a device such as a surveillance camera or a doorphone camera that cannot use auxiliary light such as a strobe or a light due to the nature of its function and application.

Although illustration of the image input device 12 is omitted,
For example, it is configured to include a lens, a CCD as an image pickup element, a drive circuit for driving the CCD, and the like. The subject image formed on the light receiving surface of the CCD through the lens is converted by each sensor into a signal charge of an amount corresponding to the incident light amount of light,
The R, G, and B data corresponding to the signal charges are output to the image processing device 14 as image data.

In the image processing device 14, gradation correction and noise removal based on the input R, G and B data of each pixel,
Perform processing such as color difference gain correction, for example, CRT or L
The image is output to the image output device 16 such as a CD.

The image processing device 14 includes a Y / C converter 18,
Y histogram calculation unit 20, LUT calculation unit 22, intensity coefficient input unit 23, LUT conversion unit 24, feature amount calculation unit 26,
Noise removal filter 28, chroma conversion unit 30, chroma cumulative histogram calculation unit 32, chroma correction value calculation unit 34,
The color difference gain processing unit 36 and the RGB conversion unit 38 are included.

The Y / C converter 18 converts the R, G, B data of each pixel output from the image input device 12 into a luminance signal Y of each pixel and chroma signals Cr, Cb. The converted luminance signal Y is stored in the Y histogram calculation unit 20 and the LU.
The chroma signals Cr and Cb output to the T conversion unit 24 are output to the chroma conversion unit 30 and the color difference gain processing unit 36.

First, the processing based on the luminance signal Y will be described.

The Y histogram calculation unit 20 creates a histogram of the luminance signal Y of each pixel output from the Y / C conversion unit 18. That is, the Y histogram calculation unit 20
Is the gradation value i from the luminance signal (gradation value) Y of each pixel.
The frequency (cumulative number of pixels) of (= 0, 1, 2, ..., N-1: N is the number of gradations) is calculated. In this embodiment, the number of gradations N is 256, and the frequency is calculated for each of the gradation values i of 0 to 255.

The LUT calculation unit 22 converts the input gradation value Y into the output gradation value T (Y) based on the frequency of each gradation value i calculated by the Y histogram calculation unit 20. Up table). Specifically, first, a correction coefficient BAI (i) for correcting the gradation value i is calculated by the following equation for each of i = 0 to 255.

[0031]

[Equation 1]

However, D (i) is the input gradation value Y of each pixel
Is the frequency of the gradation value i calculated from Upper is a cumulative value of each frequency from the gradation value i to 0 and a predetermined threshold value Upp, and Lower is a cumulative value of each frequency from the gradation value i to 0 and a predetermined threshold value Low. is there. The predetermined threshold value Upp is basically N-1 (in this case, 2).
55), the predetermined threshold Low is 0. Also, BAI
When (i) <0, BAI (i) = 0 is forcibly set.

Correction coefficient B calculated by the above equation (1)
AI (i) has a function of correcting the input gradation value Y so that the histogram of the input gradation value Y is flattened when the input gradation value Y is corrected by the correction coefficient BAI (i). For example, in FIG.
When a histogram in which the frequency of the dark portion and the frequency of the bright portion are extremely different, such as the histogram before correction shown by the solid line of, is corrected by the correction coefficient BAI (i),
Like the corrected histogram shown by the dotted line in FIG. 2, the dark portion and the bright portion are flattened histograms.

Next, the correction coefficient BAI (i) calculated by the above equation (1) is adjusted by the following equation.

BAI ′ (i) = BAI (i) ^KO (BAI (i) ≧ 1.0) (2) BAI ′ (i) = BAI (i) ^KU (BAI (i) <1.0) (3) Here, KO is for giving a tendency to the correction when BAI (i) is 1 or more, that is, in the case of a correction coefficient for correcting the input gradation value Y to be high. Intensity coefficient, that is, an intensity coefficient for adjusting the correction intensity, and KU is when BAI (i) is less than 1, that is,
In the case of a correction coefficient for correcting the input gradation value Y to be low, it is an intensity coefficient for adjusting the correction strength.

Therefore, as the value of KO increases,
As the input gradation value Y is corrected to a higher degree and the value of KU increases, the input gradation value Y increases.
The degree to which the correction is made to be low becomes large. That is, by setting so that KO> KU, the dark image is corrected to become a bright image, and KO <
By setting KU, a bright image is corrected to a dark image. In addition, KO = KU
In the case of = 0, no correction is made, and in the case of KO = KU = 1, the correction intensity is not adjusted and the correction is made by BAI (i).

The intensity coefficients KO and KU can be input from the intensity coefficient input section 23, and the image processing system 1
0 is set according to the application of the system in which it is used.
That is, when the emphasis is placed on the correction so that the dark image becomes a bright image, the setting is made such that KO> KU, and the correction is made so that the bright image becomes a dark image. If it is set, KO <KU is set. The strength coefficients KO and KU may be manually input, or the strength coefficients KO and KU may be stored in a memory in advance and read. In addition, strength coefficient KO, K
U may be stored in a rewritable memory so that it can be rewritten according to the applied system.

As described above, the degree to which the input gradation value Y is corrected to be higher (brighter) and the degree to which the input gradation value Y is corrected to be lower (darker) are set. Since both can be adjusted independently, it is possible to perform optimum gradation correction according to the system used.

Next, using the corrected correction coefficient BAI 'calculated by the above equation (2) or equation (3), the gradation value i
The gradation value T (i) corresponding to is calculated.

T (i) = i × BAI ′ (i) (4) The gradation value i and the gradation value T (i) are calculated by performing the calculation of the equation (4) for all gradation values i. It is possible to obtain an LUT that represents the correspondence relationship with. Note that T (i)> N-1
In the case of, T (i) = N-1 is forcibly set.

In the LUT conversion unit 24, the Y / C conversion unit 18
The input grayscale value Y of each pixel output from is converted into the output grayscale value T (Y) by the LUT calculated by the LUT calculation unit 22. Output gradation value T converted by the LUT conversion unit 24
(Y) is the feature amount calculation unit 26 and the noise removal filter 2
8 is output.

An example of such an LUT is shown in FIG. The LUT shown in FIG. 3 is an LUT calculated for a histogram in which the frequency of dark portions and the frequency of bright portions are extremely different as shown by the solid line in FIG. LU like this
By correcting the input gradation value Y with T, a dark part and a bright part are each flattened like the corrected histogram shown by the dotted line in FIG.

When KO = KU = 0, BAI '(i) = 1 from the above equations (2) and (3), so there is no correction, and as the value of KO or KU increases, It can be seen that the degree of correction is large. Therefore, if it is important to correct the dark image to be a bright image, it is necessary to set KO> KU so that the bright image is a dark image. When giving priority, by setting so that KO <KU, optimum gradation correction can be performed according to the system.

Further, as described above, the predetermined threshold value Upp is basically N-1 (255 in this case), and the predetermined threshold value Lo.
w is 0. In this case, as shown by the chain double-dashed line in FIG. 4, all input gradation values Y of 0 to 255 are corrected,
Depending on the system, there are cases where it is desired to convert gradation values above a predetermined threshold value or gradation values below a predetermined threshold value into substantially the same value. In this case, the predetermined threshold value Upp may be set to a value of (N-1) or less and the predetermined threshold value Low may be set to a value larger than 0. As a result, as shown by the solid line in FIG. 4, the input gradation value equal to or greater than the predetermined threshold Upp or the predetermined threshold L
It is possible to correct input grayscale values equal to or less than ow to substantially the same value.

When the Y histogram calculation unit 20 calculates a histogram, the photographed image is divided into a plurality of areas, and a weight predetermined according to the system used for each divided area is input. The gradation value Y may be multiplied and a histogram may be calculated for the input gradation value Y multiplied by the weight.

For example, when the image input device 12 is a doorphone camera or a digital camera built in a mobile phone,
There is a high probability that the main subject is in the center of the screen. In such a case, the weight of the central portion of the image is increased and the weight of the peripheral portion of the screen is reduced. As a result, gradation correction can be properly performed according to the system.

Further, when the photographed image has a solid background (for example, blue sky) having an area equal to or larger than a predetermined area, the frequency of gradation values near the solid background becomes extremely high as shown in FIG. . When the LUT is calculated for an image having such characteristics, the inclination near the solid background becomes steep as shown in FIG. 6, and the output gradation value T with respect to the gradation value i near the solid background.
The variation in (i) may be large, and noise may appear near the solid background of the corrected image.

Therefore, when calculating the histogram in the Y-histogram calculating section 20, the frequency of the gradation value i having a frequency equal to or higher than a predetermined threshold value may be reduced. For example, a gradation value i having a frequency equal to or higher than a predetermined threshold value
A predetermined value is subtracted from the frequency of, or set to 0. As a result, it is possible to prevent noise from appearing near the solid background of the corrected image.

In the feature quantity calculation unit 26, the LUT calculation unit 22
On the basis of the LUT calculated in step 1, the characteristic amount, which is a physical amount indicating the relative relationship between the degree of correction for brightening the input image and the noise generated in the corrected image, is calculated. In particular,
For example, as indicated by the dotted line in FIG. 7, the gradation value i and the gradation value T
An area closed by a line having the same value as (i) and a line showing the relationship between the gradation value i and the gradation value T (i) shown by a solid line (curve), and more than the dotted line. The area of the region A formed on the upper side, that is, the gradation value i is increased (brightened)
The difference (AB) between the degree of correction and the area of the region B formed below the dotted line, that is, the degree of correction that lowers (darkens) the gradation value i, is calculated, and this is used as the feature amount. F
And

This feature amount F is the noise removal filter 28.
Is output to. The noise removal filter 28 is configured to include, for example, a low-pass filter and the like, and noise is generated with respect to the gradation value T (Y) output from the LUT conversion unit 24 according to the feature amount F output from the feature amount calculation unit 26. The removal processing is performed and the gradation value T ′ (Y) is output to the RGB conversion unit 38. Specifically, for example, when the feature amount F is large, that is, when the area of the region A is larger than the area of the region B, the degree of bright correction is increased and the degree of noise is increased. , Increase the intensity of noise removal processing. On the other hand, when the feature amount F is small, that is,
When the area of the area A is the same as the area of the area B, the intensity of the noise removal processing is reduced, or the noise removal processing is not performed and output to the RGB conversion unit 38. In this way, since the intensity of noise removal can be changed according to the degree (trend) of correction, it is possible to optimally remove noise according to the degree of correction.

Next, the processing based on the chroma signals Cr and Cb will be described.

The chroma conversion section 30 calculates the chroma value C relating to the saturation of each pixel based on the chroma signals Cr and Cb of each pixel and outputs it to the chroma cumulative histogram calculation section 32. The chroma value C can be obtained by the following equation.

[0053]

[Equation 2]

In the chroma cumulative histogram calculation section 32,
A cumulative histogram of the chroma values is calculated from the chroma value C of each pixel and output to the chroma correction value calculation unit 34. FIG. 8 shows an example of the cumulative histogram. The cumulative histogram shown by the solid line in FIG. 8 is a cumulative histogram of an image with high saturation as shown by the solid line in FIG. 9, and the histogram shown by the dotted line in FIG. 8 is that of the image with low saturation as shown by the dotted line in FIG. It is a cumulative histogram.

The chroma correction value calculation unit 34 calculates a correction value G for correcting the chroma signals Cr and Cb based on the chroma cumulative histogram calculated by the chroma cumulative histogram calculation unit 32, and the color difference gain processing unit 36. Output to. Specifically, for example, the chroma value Ct at which the cumulative frequency of the chroma value C becomes equal to or higher than a predetermined threshold THr (95%, for example) is determined. In the case of the cumulative histogram shown by the dotted line in FIG. 8, Ct = C1, and in the case of the cumulative histogram shown by the solid line in FIG. 8, Ct = C2. When the chroma value Ct is equal to or higher than a predetermined threshold THc for determining whether or not the image has high saturation of a predetermined value or more, it is determined that the image has a strong tendency of high saturation, and the color difference gain is set. A correction value G for suppressing is calculated, and the color difference gain processing unit 3
Output to 6.

A predetermined threshold value Ct ₂ for determining whether or not the chroma value Ct has a low saturation of a predetermined value or less is set, and it is determined whether or not Ct is Ct ₂ or less, and the predetermined value is determined. When it is equal to or less than the threshold value, it may be determined that the image has a strong tendency of low saturation, and a correction value G for increasing the color difference gain may be calculated and output to the color difference gain processing unit 36. In addition, the chroma value Ct is calculated according to a predetermined formula for adjusting the saturation.
The correction value G according to the above may be calculated.

The color difference gain processing unit 36 corrects the chroma signals Cr and Cb according to the following equation, and the corrected chroma signal C
The r ′ and Cb ′ are output to the RGB conversion unit 38.

Cr ′ = Cr × G (6) Cb ′ = Cb × G (7) That is, in the chroma correction value calculation unit 34, when it is desired to correct an image having low saturation to the high saturation side, that is, the color difference gain is set. When it is desired to increase the value, a correction value G is calculated so that G> 1, and when it is desired to correct a high saturation image to the low saturation side, that is, to suppress the color difference gain, G <1 is set. The correction value G is calculated. Further, when the saturation correction is not necessary, G = 1.

In the RGB conversion section 38, the output gradation value T ′ (Y) output from the noise removal filter 28 and the chroma signals Cr ′ and C output from the color difference gain processing section 36.
b, and R, G, and B signals are created and output to the image output device 16.

The image output device 16 outputs an image based on the R, G and B signals output from the image processing device 14.

As described above, in the present embodiment, the luminance signal is corrected so that the histogram is flattened based on the histogram of the luminance signal of the image input by the image input device 12. It is possible to correct an image into a dark image and a dark image into a bright image. In addition, since the correction intensity for making the image brighter and the correction intensity for making the image darker can be adjusted independently, they can be adjusted according to the type and characteristics of the image input device 12, the setting place, the use, etc. By adjusting the intensity of, it is possible to perform optimum gradation correction according to each system.

In the present embodiment, the case where the luminance signal Y is created from the RGB signals and the histogram is calculated and the LUT is calculated based on the calculated histogram has been described, but the present invention is not limited to this. Is calculated, and the LUT is calculated based on the calculated histogram.
Alternatively, the RGB signal may be converted into an L ^* a ^* b ^* signal, a histogram of the L ^* signal representing the brightness may be calculated, and an LUT may be created based on the calculated histogram.

[0063]

As described above, according to the present invention, the correction coefficient for correcting each of the brightness data so that the histogram of the brightness data of each pixel calculated from the image data of the captured image is flattened. Can be added to each correction coefficient, and the correction coefficient calculated from the intensity coefficient can be given a tendency. Accordingly, it has an effect that it can be optimally corrected.

[Brief description of drawings]

FIG. 1 is a schematic block diagram of an image processing system.

FIG. 2 is a diagram showing a relationship between gradation value and frequency.

FIG. 3 is a diagram showing a relationship between an input gradation value and an output gradation value.

FIG. 4 is a diagram showing a relationship between a gradation value and a frequency in an image with a solid background.

FIG. 5 is a diagram showing a relationship between an input gradation value and an output gradation value in an image having a solid background.

FIG. 6 is a diagram showing a relationship between an input gradation value and an output gradation value.

FIG. 7 is a diagram for explaining a feature amount.

FIG. 8 is a diagram showing a relationship between chroma value and cumulative frequency.

FIG. 9 is a diagram showing a relationship between chroma value and frequency.

[Explanation of symbols]

Reference Signs List 10 image processing system 12 image input device 14 image processing device 16 image output device 18 Y / C conversion unit (creating means) 20 Y histogram calculation unit (histogram calculation means,
Reduction unit 22 LUT calculation unit (correction coefficient calculation unit) 23 Strength coefficient input unit 24 LUT conversion unit (conversion unit) 26 Feature amount calculation unit 28 Noise removal filter 30 Chroma conversion unit 32 Chroma cumulative histogram calculation unit 34 Chroma correction value calculation Part 36 Color difference gain processing unit

   ─────────────────────────────────────────────────── ─── Continued front page    F term (reference) 5B057 AA19 BA02 CA01 CA08 CA12                       CA16 CB01 CB08 CB12 CB16                       CE11 DA17 DB02 DB06 DB09                       DC23 DC36                 5C021 PA17 PA72 PA77 PA80 XA35                 5C054 AA01 CC02 CH01 DA08 EA05                       ED11 EJ01 EJ02 FB03 FC03                       HA18                 5C077 LL04 LL19 MP01 NN02 PP02                       PP15 PP44 PQ12 PQ19 PQ23                       TT09

Claims (3)

[Claims] 1. A creating means for creating brightness data of each pixel of the taken image from image data of the taken image; a histogram calculating means for calculating a histogram of the brightness data; and a flattening of the histogram based on the histogram. Correction coefficient calculation means for calculating correction coefficients for correcting each of the brightness data, input means for inputting an intensity coefficient for giving a tendency to correction of the brightness data, and the input means An image processing apparatus, comprising: a conversion unit that converts the luminance data based on a correction coefficient to which the tendency is given by the intensity coefficient input by the above. 2. The intensity coefficient is a first intensity coefficient that gives a tendency to a correction coefficient for correcting the luminance data to be high and a correction coefficient for correcting the luminance data to be low. The image processing apparatus according to claim 1, wherein the image processing apparatus comprises a second intensity coefficient for imparting a tendency. 3. The histogram further includes a reduction unit that reduces the frequency of luminance data that is equal to or greater than a predetermined frequency, and the correction coefficient calculation unit includes the correction coefficient based on the histogram reduced by the reduction unit. The image processing apparatus according to claim 1 or 2, wherein the image processing apparatus calculates each of the following.