JP2006128986A - Image processing device - Google Patents

Abstract

Even when the gradation distribution of an input image is concentrated in a part or the center of the distribution is in a high gradation part or a low gradation part, the shape of the gradation distribution is not destroyed and the gradation distribution Provided is an image processing apparatus capable of performing gradation correction in which the center does not change greatly.
A gradation histogram for each color component of an input image signal is calculated, a maximum value and a minimum value of the gradation histogram are calculated, and a gradation correction maximum value for each color component is calculated from the calculated maximum value. The tone correction minimum value for each color component is calculated from the minimum value calculated from the minimum value, and the tone for each color component is set so that the maximum value becomes the tone correction maximum value and the minimum value becomes the tone correction minimum value. Tone correction is performed by widening the distribution.
[Selection] Figure 1

Description

  The present invention relates to a technique applicable to image equipment such as a video camera, an electronic still camera, a mobile phone with a camera, and a printer, and relates to control of gradation of an image signal.

  Conventionally, as a method of correcting the gradation of the input image signal, the maximum value and the minimum value of the gradation of the input image signal are detected, and the detected maximum value and minimum value of the detected gradation are the dynamic range of the image signal system. There is a method of performing gradation correction by widening the gradation distribution so as to be the maximum value and the minimum value. In this method, when the difference between the maximum value and the minimum value of the gradation is small, that is, when the distribution width is small, the gradation correction amount becomes too large, and the image quality may be deteriorated such as gradation skip or noise increase. is there. Therefore, there is a method of performing gradation correction by limiting the amount of gradation correction by correcting the maximum value and the minimum value according to the distribution width (see, for example, Patent Document 1).

This conventional example is shown in FIG. In the conventional example, a luminance histogram of a multi-valued image as an input image is created (S1302), and then the upper limit value (Hi point), lower limit value (Lo point) of the luminance histogram distribution, and the difference between the upper limit value and the lower limit value. A certain distribution width is calculated (S1303), and if the distribution width falls below a predetermined distribution width lower limit value in the contrast excessive measure (S1304), the upper limit value and the lower limit value are set again according to the distribution width lower limit value. Then, a correction table in which the set upper limit value and lower limit value are the maximum distribution width of the multi-valued image is generated (S1305), and image correction processing is performed based on the correction table (S1306).
JP 2001-148785 A

  However, when gradation correction is performed as in the conventional example, the upper limit value and the lower limit value are reset according to the distribution width. Therefore, when the distribution width is small and the center of the distribution is in the high gradation portion or the low gradation portion, In some cases, the center of the gradation changes greatly, or the shape of the gradation distribution changes greatly.

For example, when the center of gradation distribution is in the high gradation part as shown in FIG. 14A and when the center of gradation distribution is in the low gradation part as shown in FIG. ) And lower limit (Lo point) are corrected to the new Hi point and the new Lo point, respectively, and the gradation distribution is adjusted so that the new Hi point becomes the maximum value of the gradation and the new Lo point becomes the minimum value of the gradation. A widening process is performed, and gradation correction is performed as shown in FIG. 14B in FIG. 14A and in FIG. 15B in FIG. 15A. 14 and 15, the dotted line is a line representing the center of gradation, and the difference between the center before correction and after correction is ΔT. In the conventional example, since the gradation correction amount is determined by the distribution width, when the center of the gradation is in the high gradation part as shown in FIG. 14, the gradation spreads greatly on the low gradation side and the value of ΔT is large. From this, it can be seen that the center of the gradation moves greatly. Also, the distribution shape changes greatly.

  Further, when the center of the gradation is in the low gradation part as shown in FIG. 15, the gradation spreads greatly toward the high gradation side and the value of ΔT increases, so that the distribution center moves greatly to the high gradation side. You can see that Also, the distribution shape changes greatly.

  When the gradation center is moved greatly by gradation correction, the phenomenon that the brightness of the image changes greatly occurs, and when the gradation distribution shape changes greatly, problems such as gradation skip and noise increase occur. As a result, the image quality has deteriorated.

  For example, there is an underwater image as a representative image having a small gradation distribution width. In the case of an underwater image, the gradation distribution differs greatly for each color component, and the characteristics shown in FIGS. 16 to 19 are obtained. 16 shows the red component gradation distribution of the underwater image, FIG. 17 shows the green component gradation distribution of the underwater image, FIG. 18 shows the blue component gradation distribution of the underwater image, and FIG. 19 shows the gradation distribution of the luminance component of the underwater image. In the underwater image, since the red component is absorbed by water, it can be seen that the red component is extremely concentrated and distributed in the low gradation portion. For this reason, even if gradation correction is performed from the luminance distribution as in the conventional example, there is a case where sufficient gradation correction cannot be performed for the red component.

  The present invention is for solving the above-described problems, and calculates the maximum value and the minimum value of the gradation histogram of the input image for each color component, and calculates the maximum gradation correction value from the maximum value of the gradation histogram. The tone distribution is calculated so that the tone correction minimum value is calculated from the tone histogram minimum value, the tone histogram maximum value becomes the tone correction maximum value, and the tone histogram minimum value becomes the tone correction minimum value. An object of the present invention is to provide a gradation correction device that can perform optimum correction even when the gradation distribution is concentrated in a part by widening the image, and can perform optimum correction even when the input image is an underwater image. And

  In order to solve the conventional problem, the image processing apparatus of the present invention is calculated by a gradation histogram calculation unit that calculates gradation distribution information for each color component of an input image signal and the gradation histogram calculation unit. Maximum value calculating means for calculating the maximum value of the gradation histogram for each color component, minimum value calculating means for calculating the minimum value of the gradation histogram for each color component calculated by the gradation histogram calculating means, and the maximum Gradation correction maximum value calculating means for calculating a gradation correction maximum value for each color component from the maximum value calculated from the value calculating means, and gradation correction minimum value for each color component from the minimum value calculated from the minimum value calculating means Gradation correction minimum value calculation means for calculating the maximum correction value, the maximum value calculated from the maximum value calculation means becomes the tone correction maximum value, and the minimum value calculated from the minimum value calculation means becomes the gradation correction minimum value. color And having a gradation correction means to broaden the gradation distribution, the per minute.

Further, the maximum value calculating means of the image processing apparatus of the present invention sequentially decreases the gradation from the maximum value of the dynamic range that can be taken by the input image signal in the gradation histogram, and cumulatively adds the number of pixels for each gradation. Is calculated as the maximum value when the number of pixels exceeds the preset number of pixels,
The minimum value calculation means sequentially increases the gradation from the minimum value of the dynamic range that can be taken by the input image signal in the gradation histogram, cumulatively adds the number of pixels for each gradation, and the cumulative value is a pixel in which the cumulative value is set in advance Calculate the gradation value when the number is exceeded as the minimum value,
It is characterized by that.

  The maximum value calculation means of the image processing apparatus of the present invention calculates the maximum value of the gradation histogram for each color component of the input image signal, and maximizes the largest value of the calculated maximum values of the gradation histogram for each color component. The minimum value calculating means calculates a minimum value of the gradation histogram for each color component of the input image signal, and sets the smallest value of the calculated minimum values of the gradation histogram for each color component as the minimum value. It is characterized by.

  Further, the gradation correction maximum value calculating means of the image processing apparatus of the present invention is configured to set the maximum value to the dynamic value when the maximum value of the gradation histogram calculated from the maximum value calculating means is equal to or greater than a predetermined first threshold value. The maximum value of the range is used. In other cases, the gradation correction maximum value is calculated from the maximum value using a predetermined conversion formula, and the gradation correction minimum value calculation means is calculated from the minimum value calculation means. When the minimum value of the gradation histogram is less than or equal to a predetermined second threshold value, the minimum value is set as the minimum value of the dynamic range. In other cases, a conversion equation determined from the minimum value is used. The minimum gradation correction value is calculated.

  In addition, the image processing apparatus of the present invention includes an average luminance value calculation unit before calculating gradation that calculates an average luminance value before gradation correction of an input image signal, and after gradation correction that calculates an average luminance value after gradation correction. Average luminance calculation means, luminance correction amount calculation means for calculating a luminance correction amount from the average luminance value before gradation correction and the average luminance value after gradation correction, and after gradation correction by the luminance correction amount calculation means Brightness correction means for correcting the brightness of the image signal, and the average brightness value after gradation correction is made the same as the average brightness value before gradation correction.

  The image processing apparatus of the present invention further includes an underwater image determination unit that determines whether or not the input image is an underwater image. Only when the input image is determined to be an underwater image by the underwater image determination unit, the red component level is determined. This is characterized in that the tone correction maximum value is brought close to the maximum value of the dynamic range.

  The image processing apparatus of the present invention further includes a color difference signal calculating unit that calculates a color difference signal from an input image, and a color difference average value calculating unit that calculates an average value of the color difference signals, and the underwater image is calculated based on the average value of the color difference signals. It is characterized in that an underwater image is judged by the judging means.

  According to the image processing apparatus of the present invention, even when the gradation distribution is concentrated in a part, or even when the center of the distribution is in the high gradation part or the low gradation part, the shape of the gradation distribution is not destroyed. Tone correction can be performed without significant change in the center of the tone distribution.

  According to the image processing apparatus of the present invention, the maximum value and the minimum value of the gradation histogram can be accurately calculated even when there is noise in the input image, so that accurate gradation correction can be performed.

  Further, according to the image processing apparatus of the present invention, the maximum value and the minimum value of the gradation are calculated for each color signal of the input image signal, and the largest value among the calculated maximum values for each color signal is set as the maximum value. By performing correction with the smallest value for each minimum value as the minimum value, the gradation correction amount for each color component becomes the same, so that gradation correction can be performed without losing white balance.

  Further, according to the image processing apparatus of the present invention, the gradation correction maximum value calculating means calculates the gradation correction maximum value in proportion to the maximum gradation value, and the gradation correction minimum value calculating means is the minimum gradation correction value. Since the gradation correction minimum value is calculated in proportion to the value, gradation correction can be performed according to the maximum and minimum values of the input image. Even when the center of the gray level is in the high gray level portion or the low gray level portion, the gray level correction can be performed without greatly changing the center of the gray level distribution and without destroying the shape of the gray level distribution.

  Further, according to the image processing apparatus of the present invention, it is possible to calculate the average luminance value before and after gradation correction, and to make the luminance value after gradation correction the same as the luminance value before gradation correction. Therefore, it is possible to prevent the luminance from changing due to the gradation correction.

  Further, according to the image processing apparatus of the present invention, only when the input image is determined to be an underwater image, the gradation correction amount of the red component is adjusted by bringing the gradation correction maximum value of the red component close to the maximum value of the dynamic range. Therefore, it is possible to perform gradation correction suitable for an underwater image with few red components.

  In addition, according to the image processing apparatus of the present invention, the gradation correction of the underwater image is performed by a simple method in order to calculate whether the image is an underwater image based on the average value of the color difference signals.

  Embodiments of an image processing apparatus according to the present invention will be described below in detail with reference to the drawings.

  FIG. 1 is a block diagram showing the configuration of an image processing apparatus according to the first embodiment of the present invention.

  In FIG. 1, 101 is a gradation histogram calculating means for calculating a gradation histogram of an input image signal, 102 is a maximum value calculating means for calculating the maximum value of the gradation histogram, and 103 is a minimum for calculating the minimum value of the gradation histogram. A value calculation unit 104 is a gradation correction maximum value calculation unit 104 that calculates a gradation correction maximum value from the maximum value, 105 is a gradation correction minimum value calculation unit that calculates a gradation correction minimum value from the minimum value, and 106 is a gradation It is a correction means. Hereinafter, the operation in the case where the input image is 8 bits will be described.

  First, a gradation histogram is calculated for each color component by the gradation histogram calculation means 101 from the input image signal. The gradation histogram represents the gradation distribution of the input image, for example, as shown in FIG. In FIG. 2, the horizontal axis represents the gradation level, and the vertical axis represents the number of pixel distributions for each gradation. P1 represents the minimum value of the gradation distribution, and P2 represents the maximum value of the gradation distribution.

  Next, the maximum value calculation means 102 and the minimum value calculation means 103 calculate the maximum value and the minimum value of the gradation distribution. The calculation method will be described below.

  The maximum value calculation means 102 calculates the maximum value of the gradation distribution. For the maximum value of the gradation distribution, the number of gradation distributions is examined from 255, which is the maximum value of the gradation level, to 0, which is the minimum value of the gradation level, and the gradation level with the gradation distribution is first maximized. Calculate as a value. When the maximum value is calculated, if there is noise in the high gradation part as shown in FIG. 2, the noise part is calculated as the maximum value and malfunction may occur, so that the gradation level goes from 255 to 0. Alternatively, the number of distributions for each gradation may be cumulatively added, and the gradation level when the cumulative value exceeds the threshold value may be calculated as the maximum value. Since the noise part has a small number of distributions, it is possible to calculate the true maximum value P2 by ignoring the noise part by cumulative addition. The threshold value is a numerical value of several percent of the total number of gradation distributions, and is usually a value around 1%, but is not limited to this value.

  The minimum value calculation means 103 calculates the minimum value of the gradation distribution. For the minimum value of the gradation distribution, the number of gradation distributions is examined from 0, which is the minimum value of the gradation level, to 255, which is the maximum value of the gradation level, and the gradation level with the gradation distribution is first minimized. Calculate as a value. When calculating the minimum value, if there is noise in the low gradation part as shown in FIG. 2, the noise part may be calculated as the minimum value and malfunction may occur. Alternatively, the number of distributions for each gradation may be cumulatively added, and the gradation level when the cumulative value exceeds the threshold value may be calculated as the minimum value. Since the noise part has a small number of distributions, the true minimum value P1 can be calculated by ignoring the noise part and ignoring the noise part. The threshold value is a numerical value of several percent of the total number of gradation distributions, and is usually a value around 1%, but is not limited to this value.

  Next, for each color component, the gradation correction maximum value calculation means 104 and the gradation correction minimum value calculation means 105 calculate the gradation correction maximum value and the gradation correction minimum value, which are target values for expanding the gradation distribution. . By gradation correction, the gradation distribution is expanded so as to be distributed between the maximum gradation correction value and the minimum gradation correction value. The method for calculating the maximum gradation correction value and the minimum gradation correction value is the greatest feature of the present invention.

  When the minimum gradation value is P1, the maximum value is P2, the gradation correction minimum value is Q1, and the gradation correction maximum value is Q2, the respective relationships are as shown in FIG. By gradation correction, the gradation distribution is expanded so that the minimum value P1 becomes the gradation correction minimum value Q1 and the maximum value P2 becomes the gradation correction maximum value Q2. In general, when the input image signal is 8 bits, gradation correction is performed by setting the gradation correction minimum value Q1 to 0 and the gradation correction maximum value Q2 to 255. In the present invention, the gradation correction maximum value Q2 is set to The maximum gradation calculation value P2 calculated by the maximum value calculation unit 102 is calculated, and the minimum gradation correction value Q1 is calculated by the minimum gradation value P1 calculated by the minimum value calculation unit 103. Therefore, when the center of the gradation distribution is in the high gradation part as in the prior art (FIG. 14), or when the center of the gradation distribution is in the low gradation part (FIG. 15), the gradation center moves greatly. Or a problem that the distribution shape changes greatly does not occur. A method for calculating the maximum gradation correction value and the minimum gradation correction value will be described below.

  First, a method for calculating the maximum gradation correction value will be described. The maximum gradation correction value is proportional to the maximum gradation value. When the maximum gradation value is close to the maximum value of the dynamic range, the maximum gradation correction value is the maximum value of the dynamic range or a value close to the maximum value, and decreases as the maximum gradation value decreases. Try to take. An example of the relationship between the maximum value and the gradation correction maximum value is shown in FIG.

  In FIG. 4, the horizontal axis represents the maximum gradation value, and the vertical axis represents the maximum gradation correction value. When the maximum value is sufficiently large, the gradation correction maximum value is the maximum value of the dynamic range that can be taken by the input image. In the example of FIG. 4, when the maximum value exceeds 230, the maximum gradation correction value is set to 255. As the maximum value decreases, the gradation correction maximum value also decreases. For example, when the maximum value is 100, the gradation correction maximum value is 175. As shown in the drawing, when the maximum value is less than a certain value, the maximum gradation correction value may be set to a constant value. In this way, excessive gradation correction can be prevented when the maximum value is extremely small. In the example of FIG. 4, when the maximum value is 100 or less, the maximum gradation correction value is a fixed value of 175. In the case of FIG. 4, the maximum gradation correction value with respect to the maximum value changes linearly, but it may be a curve such as a quadratic curve or a cubic curve.

  Next, a method for calculating the gradation correction minimum value will be described. The minimum gradation correction value is proportional to the minimum gradation value. If the minimum value of gradation is close to the minimum value of dynamic range, the minimum value of gradation correction takes the minimum value of dynamic range or a value close to the minimum value, and increases as the minimum value of gradation increases. Try to take. An example of the relationship between the minimum value and the gradation correction minimum value is shown in FIG.

  In FIG. 5, the horizontal axis represents the minimum value, and the vertical axis represents the gradation correction minimum value. When the minimum value is sufficiently small, the gradation correction minimum value takes the minimum value of the dynamic range. In the example of FIG. 5, when the minimum value is smaller than 20, the gradation correction minimum value is set to 0. As the minimum value increases, the gradation correction minimum value also increases. For example, when the minimum value is 150, the gradation correction minimum value is 75. Similar to the calculation of the maximum gradation correction value, the minimum gradation correction value may be a constant value when the minimum value is greater than a certain value. In this way, excessive gradation correction can be prevented when the minimum value is an extremely large value. In the case of FIG. 5, when the minimum value exceeds 150, the gradation correction minimum value is a fixed value of 75. In the case of FIG. 5, the gradation correction minimum value with respect to the minimum value changes linearly, but it may be a curve such as a quadratic curve or a cubic curve.

  As described above, the minimum value P1, the gradation correction target value that is the gradation correction target value Q1, the maximum value P2, and the gradation correction target value that is the gradation correction maximum value Q2 can be calculated. The gradation correction unit 106 performs gradation correction so that gradations distributed from P1 to P2 before correction are distributed from Q1 to Q2. In the gradation correction means 106, gradation correction is performed from the values of the minimum gradation value P1, the maximum value P2, the gradation correction minimum value Q1, and the gradation correction maximum value Q2 by the following equation.

Y = (Q2-Q1) / (P2-P1) * (X-P1) + Q1
Here, X represents the gradation level before correction, Y represents the gradation level after correction, and the relationship between the gradation level before correction and the gradation level after correction is as shown in FIG.

  An example in which gradation correction is performed in this way is shown in FIGS. FIG. 7A shows a gradation distribution before gradation correction of an image whose distribution is biased in the high gradation part, and gradations are distributed from the minimum value P1 to the maximum value P2. FIG. 7B shows the result of the gradation correction according to the present invention. The gradation is distributed from the gradation correction minimum value Q1 to the gradation correction maximum value Q2. The dotted line indicates the center of the gradation, and ΔT is the difference between the center of the gradation before correction and the center of the gradation distribution after correction. As can be seen from FIG. 7, the value of ΔT is smaller than the value shown in FIG. Since expansion to the low gradation is thus limited, it is possible to prevent the center of the distribution from largely moving to the low gradation side and to prevent the distribution shape from changing greatly.

  Further, FIG. 8 shows a correction result in the case of an image whose distribution is biased in the low gradation part, contrary to the example shown in FIG. FIG. 8A shows the tone distribution before tone correction, and FIG. 8B shows the tone distribution after tone correction. As can be seen from the difference ΔT between the center of the gradation before correction and the center of the gradation distribution after correction, it can be seen that the value of ΔT is smaller than the value shown in FIG. . As described above, since the expansion to the high gradation is restricted, it is possible to prevent the center of the distribution from largely moving to the high gradation side and to prevent the distribution shape from changing greatly.

  Although the gradation correction is performed for each color component of the input image by the method described above, when the input image signal is a color signal of red, green, and blue, the maximum value of the gradation for each of red, green, and blue, The minimum value, the maximum gradation correction value, and the minimum gradation correction value are calculated, and the gradation is corrected individually for red, green, and blue. In the printing system, yellow, magenta, and cyan are input image signals as color components, but gradation correction can be performed by the same processing.

  In addition, by making the gradation correction amounts for red, green, and blue the same, gradation correction can be performed while maintaining white balance. The processing method in that case will be described with reference to the flowchart of FIG.

  First, a gradation histogram is calculated for each of the red, green, and blue signals (S902). Next, the maximum value and the minimum value are calculated for each of red, green, and blue (S903, S904). The largest value among the calculated maximum values of red, green, and blue is adopted as the maximum value A (S905). Further, the smallest value of the calculated red, green and blue minimum values is adopted as the minimum value B (S906).

  The maximum gradation correction value is calculated from the calculated maximum value A (S907), and the minimum gradation correction value is calculated from the minimum value B (S908). The gradation correction is performed from the maximum value, gradation correction maximum value, minimum value, and gradation correction minimum value calculated in this way to red, green, and blue (S909). In this way, since the gradation correction amount for each color becomes the same, the gradation correction can be performed without breaking the white balance.

  Also, the average value of the luminance before and after the gradation correction is calculated, and when the luminance of the input image signal is changed by the gradation correction, the average value of the luminance of the input image signal and the average value of the luminance of the output image signal are calculated by the luminance correction means. You may correct | amend so that it may correspond. If the gradation histogram of the input image signal is biased toward the low gradation side or the high gradation side, the average luminance of the image signal may change greatly due to gradation correction. In such a case, the image quality may be reduced due to the gradation correction. Accordingly, when the luminance of the image changes due to the gradation correction, the image quality can be prevented from being lowered by returning to the luminance before the gradation correction. A processing method in this case will be described with reference to FIG.

In FIG. 10, the average luminance value before gradation correction is calculated by the luminance average value calculation means 1001 before gradation correction. Next, the average value of luminance after gradation correction is calculated by the luminance average value calculation means 1002 after gradation correction. The luminance correction amount calculation means 1003 calculates a luminance correction amount such that the average luminance value before gradation correction matches the average luminance value after gradation correction. For example, when the average luminance value before gradation correction is L, the average luminance value after gradation correction is M, and luminance correction is performed by gamma correction,
L = ^Mγ
Then, a gamma correction coefficient γ is calculated, and the brightness is corrected by gamma correction. In this case, the gamma correction coefficient is calculated by the following equation.

γ = logL / logM
The luminance correction unit 1004 extracts a luminance component from the signal subjected to gradation correction by the gradation correction unit 106, and performs gamma correction on the luminance component with the calculated gamma correction coefficient γ.
In this way, by making the luminance value before gradation correction the same as the luminance value after gradation correction, it is possible to prevent a decrease in image quality due to a decrease in luminance.

  Also, as shown in FIG. 11, it is determined whether the input image is an underwater image. If the input image is an underwater image, the tone correction maximum value of the red component is set to a large value to perform tone correction. Good. As described above, in the underwater image, the red component is absorbed by water, and the red component has a small value, resulting in distribution characteristics as shown in FIGS. 16 shows an example of the gradation distribution characteristic of the red component, FIG. 17 shows an example of the gradation distribution characteristic of the green component, and FIG. 18 shows an example of the gradation distribution characteristic of the blue component. Therefore, by setting the maximum gradation correction value of the red component to a larger value, a large gradation correction amount is applied to the red component, so that optimal gradation correction can be performed on the underwater image.

  In FIG. 11, reference numeral 1101 denotes a color difference signal calculation unit, 1102 denotes a color difference average value calculation unit, and 1103 denotes an underwater image determination unit. The underwater image is determined by calculating an average value of the color difference signals of the input image signal and using the average value. The determination method will be described below.

The color difference signal has a red component of the image signal as R, a green component as G, a blue component as B, and a luminance component as Y.
Cr = R−Y
Cb = BY
Cr, Cb represented by In the case of a normal image, both Cr and Cb are distributed around 0, but in the case of underwater, they are distributed in a blue region as shown in FIG. In FIG. 12, the shaded area indicates the distribution of the color difference signal of the input image. Therefore, the average value of Cr and Cb is calculated. If the average value is a value in the blue region, it is determined that the water is underwater. For example, when the average value of the Cr component is smaller than −20 and the average value of the Cb component is a value from −20 to 20, it is determined that the water is underwater. When it is determined to be underwater, processing is performed to bring the maximum gradation correction value of the red component closer to the maximum value of the dynamic range. For example, a value obtained by adding 30 to the maximum gradation correction value calculated in FIG. 4 is used as the maximum gradation correction value to approximate the maximum value of the dynamic range. By doing so, the red component is more greatly corrected in gradation, and thus more optimal gradation correction can be performed.

  As described above, in the first embodiment, the tone correction maximum value calculated from the maximum value and the maximum value of the tone histogram and the tone correction minimum value calculated from the minimum value and the minimum value of the tone histogram are used. In order to perform gradation correction, even if the gradation distribution of the input image is concentrated in a part or the center of the distribution is in the high gradation part or the low gradation part, the shape of the gradation distribution is not destroyed. It is possible to perform tone correction in which the center of the tone distribution does not change greatly.

  An image processing apparatus according to the present invention provides an input image signal based on a maximum tone correction value calculated from a maximum value and a maximum value of a tone histogram, and a tone correction minimum value calculated from a minimum value and a minimum value of the tone histogram. Since it is possible to perform gradation correction in accordance with the characteristics of the image processing device, it is useful as an image processing device for various input devices, and particularly when the gradation distribution of the input image signal is only partially distributed like an underwater image, Suitable when the distribution is biased.

1 is a block diagram of an image processing apparatus according to Embodiment 1 of the present invention. The figure showing the gradation histogram in Embodiment 1 of this invention Explanatory drawing of the gradation correction method in Embodiment 1 of this invention FIG. 6 is a relationship diagram between the maximum gradation value and the maximum gradation correction value in Embodiment 1 of the present invention FIG. 5 is a relationship diagram between the minimum gradation value and the minimum gradation correction value in Embodiment 1 of the present invention The figure which shows the gradation correction characteristic in Embodiment 1 of this invention The figure which shows the gradation correction result in Embodiment 1 of this invention The figure which shows the gradation correction result in Embodiment 1 of this invention Flowchart of gradation correction method using the same correction amount in Embodiment 1 of the present invention Block diagram of an image processing apparatus when performing luminance correction in Embodiment 1 of the present invention 1 is a block diagram of an image processing apparatus when performing gradation correction of an underwater image in Embodiment 1 of the present invention. Color component distribution chart of underwater image in Embodiment 1 of the present invention Flowchart for a conventional image processing apparatus Explanatory drawing about a conventional image processing apparatus Explanatory drawing about a conventional image processing apparatus Gradation characteristics of underwater image Gradation characteristics of underwater image Gradation characteristics of underwater image Gradation characteristics of underwater image

Explanation of symbols

101 gradation histogram calculation means 102 maximum value calculation means 103 minimum value calculation means 104 gradation correction maximum value calculation means 105 gradation correction minimum value calculation means 106 gradation correction means 1001 luminance average value calculation means 1001 before gradation correction After-correction brightness average value calculation means 1003 Brightness correction amount calculation means 1004 Brightness correction means 1101 Color difference signal calculation means 1102 Color difference average value calculation means 1103 Underwater image determination means

Claims (7)

A gradation histogram calculating means for calculating gradation distribution information for each color component of the input image signal;
The maximum value calculating means for calculating the maximum value of the gradation histogram for each color component calculated by the gradation histogram calculating means, and the minimum value of the gradation histogram for each color component calculated by the gradation histogram calculating means. A minimum value calculating means for calculating;
Tone correction maximum value calculation means for calculating the maximum gradation correction value for each color component from the maximum value calculated from the maximum value calculation means, and gradation correction for each color component from the minimum value calculated from the minimum value calculation means Gradation correction minimum value calculating means for calculating a minimum value;
Gradation correction that widens the gradation distribution for each color component so that the maximum value calculated from the maximum value calculation means becomes the maximum value of tone correction and the minimum value calculated from the minimum value calculation means becomes the minimum value of gradation correction Means,
An image processing apparatus comprising: The maximum value calculating means sequentially reduces the gradation from the maximum value of the dynamic range that can be taken by the input image signal in the gradation histogram, and cumulatively adds the number of pixels for each gradation, and the cumulative value is a preset number of pixels. The gradation value when exceeding the maximum value is calculated as the maximum value.
The minimum value calculation means sequentially increases the gradation from the minimum value of the dynamic range that can be taken by the input image signal in the gradation histogram, cumulatively adds the number of pixels for each gradation, and the cumulative value is a pixel in which the cumulative value is set in advance Calculate the gradation value when the number is exceeded as the minimum value,
The image processing apparatus according to claim 1. The maximum value calculating means calculates the maximum value of the gradation histogram for each color component of the input image signal, sets the maximum value of the calculated maximum values of the gradation histogram for each color component as the maximum value, and calculates the minimum value. The means calculates the minimum value of the gradation histogram for each color component of the input image signal, and sets the smallest value among the calculated minimum values of the gradation histogram for each color component as the minimum value. The image processing apparatus described. The gradation correction maximum value calculating means sets the maximum value as the maximum value of the dynamic range when the maximum value of the gradation histogram calculated from the maximum value calculating means is equal to or greater than a predetermined first threshold, In other cases, the maximum gradation correction value is calculated from the maximum value using a predetermined conversion formula,
The gradation correction minimum value calculation means sets the minimum value as the minimum value of the dynamic range when the minimum value of the gradation histogram calculated from the minimum value calculation means is equal to or less than a predetermined second threshold value. In other cases, the gradation correction minimum value is calculated from the minimum value using a predetermined conversion formula.
The image processing apparatus according to claim 1. An average luminance value calculation means before gradation correction for calculating an average luminance value before gradation correction of the input image signal, an average luminance value after gradation correction for calculating an average luminance value after gradation correction, and the gradation Luminance correction amount calculation means for calculating a luminance correction amount from the average luminance value before correction and the average luminance value after gradation correction, and luminance correction for correcting the luminance of the image signal after gradation correction by the luminance correction amount calculation means The image processing apparatus according to claim 1, wherein an average luminance value after gradation correction is made equal to an average luminance value before gradation correction. Underwater image determination means for determining whether the input image is an underwater image, and only when the input image is determined to be an underwater image by the underwater image determination means, the maximum gradation correction value of the red component is set to the maximum of the dynamic range. The image processing apparatus according to claim 1, wherein the image processing apparatus approaches the value. A color difference signal calculating means for calculating a color difference signal from the input image; and a color difference average value calculating means for calculating an average value of the color difference signals. The underwater image determination means determines the underwater image based on the average value of the color difference signals. The image processing apparatus according to claim 6, wherein the image processing apparatus performs the processing.