JP5297897B2 - Image signal processing device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To achieve high contrast in high-luminance and low-luminance portions while suppressing noise in the luminance portions, and to achieve high contrast in feature portions of an image. <P>SOLUTION: An image signal processing apparatus includes: a first region input/output characteristic control means 3 for controlling variation in input/output characteristics for each optional region while being interlocked with a value of an input luminance signal; a first region luminance signal correcting means 1 for varying input/output characteristics of a luminance signal in accordance with output of the first region input/output characteristic control means; a second region input/output characteristic control means 2 for controlling variation in input/output characteristics for each optional region while being interlocked with output of the first region luminance signal correcting means; a second region luminance signal correcting means 2 for varying input/output characteristics of a luminance signal in accordance with output of the second region input/output characteristic control means; a color signal input/output characteristic control means 5 for controlling variation in input/output characteristics of an input color signal by utilizing the output of the first region input/output characteristic control means and the output of the second region input/output characteristic control means; and a color signal correcting means 6 for varying input/output characteristics of a color signal in accordance with output of the color signal input/output characteristic control means. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to an image signal processing apparatus.

  As background art in this technical field, for example, there is JP-A-2007-318256 (Patent Document 1). This publication provides “an image correction circuit capable of easily realizing adaptive color correction in consideration of the luminance correction amount. The γ correction circuit 21 performs luminance correction (contrast improvement) on the input luminance signal Yin. The luminance correction amount ΔY color correction unit 3 performs color correction based on a predetermined expression, and adaptive color correction linked to the luminance correction amount ΔY in the γ correction circuit 21 is realized. Further, the denominator of the luminance correction amount ΔY is set to a fixed value L. This does not complicate the circuit configuration, and further, for example, when the luminance correction amount ΔY and the user color gain coefficient K are larger than necessary. Even so, it is avoided that the color difference signals Uout and Vout after color correction become excessive due to the fixed value L that is the denominator of ΔY and the adaptive color correction degree M that satisfies 0 <M ≦ 3. (Summary).

  Moreover, as background art of this technical field, there exists Unexamined-Japanese-Patent No. 2000-156871 (patent document 2), for example. In this publication, “the present invention relates to an image processing apparatus and an image processing method, and is applied to an image processing apparatus such as a television receiver, a video tape recorder, a television camera, a printer, etc. The gradation data can be corrected by effectively avoiding a decrease in contrast, after the luminance data y (i, j) is separated from the image data x (i, j). , J), a correction coefficient g (i, j) is generated based on the determination result r (i, j) of the region to which the image belongs, and the pixel value is corrected. ”(Summary).

  Moreover, as background art of this technical field, for example, there is JP-A-2008-263475 (Patent Document 3). According to this publication, “contrast correction is appropriately performed in accordance with each image. The input / output characteristic setting unit monotonously increases the output contrast component with respect to the input contrast component, and the value of the input contrast component is inclined near zero. The S-shaped input / output characteristic curve CL21 having the maximum contrastGain / γ_comp is generated, and the gain characteristic calculation unit generates a contrast gain curve representing the gain value characteristic with respect to the input contrast component based on the input / output characteristic curve CL21. The contrast correction unit corrects the contrast component of each pixel based on the contrast gain curve and the gain value obtained based on either the input contrast component or the luminance value. It can be applied to cameras "(summary).

  Moreover, as background art of this technical field, for example, there is JP-A-2004-23522 (Patent Document 4). The gazette discloses that the present invention provides a contrast correction circuit that can reproduce a high-contrast video regardless of the video content. The level for correcting the gradation of an input video signal. Tone correction means, means for generating a luminance signal from the input video signal, means for dividing one screen into a plurality of areas, calculating an average value of the luminance signal for each area, based on the luminance average value for each area, Means for generating frequency distribution information and control means for controlling the input / output characteristics of the tone correction means based on the frequency distribution information of luminance are provided. "(Summary).

JP 2007-318256 A JP 2000-156871 A JP 2008-263475 A JP 2004-23522 A

  In the method of Patent Document 1, “a luminance histogram distribution (not shown) of the luminance signal Yin is acquired for each image frame, and the luminance gain of the γ curve is adapted for each image frame based on the acquired luminance histogram distribution. In this way, the gradation correction processing according to the conventional method is processing for correcting the entire gradation at the expense of the contrast at both ends, as can be seen from the γ curve. Therefore, the conventional method eventually has a problem that the contrast is partially lowered in the processed image. In Patent Documents 3 and 4, the same problem can be said.

  In addition, in the method of Patent Document 2, it is necessary to store images of two frames, “imaging result VN by normal exposure” and “imaging result VS by short exposure”, in a memory, which increases the circuit scale. There was a problem.

  All or a part of the above problems are solved or improved by the invention described in the claims.

According to the present invention, all or part of the above problems can be solved or improved, and high image quality can be realized.
For example, in the first area luminance signal correcting means and the first area input / output characteristic control means (hereinafter referred to as first area means), high contrast of the luminance portion is realized while suppressing noise in the low luminance and high luminance portions. In the second area luminance signal correcting means and the second area input / output characteristic control means (hereinafter referred to as second area means), a portion having a small signal distribution is effectively used to realize a high contrast of a characteristic portion of an image. . Further, for example, the point that the dynamic range of the entire screen, which is an adverse effect of the first area means, can be improved by the second area means, and the contrast is partially reduced, which is an adverse effect of the second area means. It can be improved by the area means, and furthermore, the phenomenon that the hue becomes unnatural due to the change of the luminance and the ratio of the color signals which are harmful to the first area means and the second area means can be suppressed.

  Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

Concept of Embodiment 1 in the present invention Concept of Embodiment 2 of the present invention First region processing concept in the present invention An example of a method for calculating a blending rate for luminance correction in the present invention Example of Blending Data Generation Method for Brightness Correction in the Present Invention Second region processing concept in the present invention An example showing the slope calculation method in the present invention An example showing a rounding correction data calculation method in the present invention An example showing a correction data calculation method in the present invention

  Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a conceptual diagram of Embodiment 1 in the present invention.
The first area input / output characteristic control unit 003 performs control for changing input / output characteristics in an arbitrary small area in an image by using information on a local area including peripheral pixels of the area. is there.

  The first area luminance signal correction means 001 corrects the input / output characteristics of the input luminance signal for each small area in accordance with the output of the first area input / output characteristic control means 003.

  FIG. 3 is an example showing the concept of the first area means, and the left side shows a dark area that is crushed black and a bright area that is overexposed, and the gradation is clear from the beginning. It is an input image having both of the normal areas. The right side shows input / output characteristics for improving the contrast of each region.

  The first area input / output characteristic control means 003 (FIG. 1) determines the brightness of the small area from the local area information, for example, 3 such as a dark area, a normal area, and a bright area as shown on the left side of FIG. The determination is made for one region, and individual input / output characteristics that improve the contrast are selected as shown on the right side of FIG. In this manner, by using the local area information for brightness determination, it is possible to suppress a rapid change in input / output characteristics due to the influence of noise. Further, if the output of the tone curve is simply used, the input / output characteristics become discontinuous on the boundary between regions having different brightness, and the image will be broken. Therefore, the blending rate is calculated according to the brightness of the small area, and a signal in which the input and the output are mixed as shown in the gamma characteristic graph of FIG. 3 is output. The local area information is an index value calculated by applying a predetermined weight to the small area and its surrounding pixels. The predetermined weight may be a weight for calculating an average value. In the case of a low luminance region, the blending rate is increased as the index value is smaller, and the blending rate is decreased as the index value is larger. In the case of a high luminance region, the blending rate is increased as the index value is larger, and the blending rate is decreased as the index value is smaller. If it is not a high luminance or low luminance portion, blending is not performed. As described above, the low-brightness and high-brightness part is means for realizing high contrast of the brightness part while suppressing noise.

  The second area input / output characteristic control means 004 (FIG. 1) performs control for uniformly changing the input / output characteristics for each arbitrary image area from the output of the first area luminance signal correction means 001 (FIG. 1). Do.

  The second area luminance signal correction means 002 (FIG. 1) corrects the input / output characteristics of the luminance signal according to the output of the second area input / output characteristic control means 004 (FIG. 1).

  FIG. 6 is an example showing the concept of the second region means. This is a means for capturing the features of an image from an arbitrary image area using a luminance histogram or the like and changing the input / output characteristics uniformly so that the gradation of the feature portion is given. As a feature of the image, when the luminance is largely distributed in the center, it is determined that the feature is distributed at the intermediate luminance level, and the feature distribution is low at the low luminance and the high luminance. The input / output characteristics include an S-shaped characteristic that suppresses low-brightness and high-brightness gradations and broadens medium-brightness gradations. If the luminance is distributed at low luminance, it is determined that the feature is distributed in the low luminance portion and the distribution is small in the high luminance portion. The input / output characteristics change so as to widen the gradation of the low-brightness part and suppress the gradation of the high-brightness part that originally has a small distribution. In this way, the input luminance distribution is used to determine the parts with low and high distribution, and adjust the shape and slope of the tone curve according to the characteristics of the distribution to effectively use the part with low signal distribution. The high contrast of the characteristic part is realized.

  The color signal input / output characteristic control unit 005 (FIG. 1) changes the luminance signal from the output of the first region luminance signal correction unit 001 (FIG. 1) and the output of the second region luminance signal correction unit 002 (FIG. 1). To control the change in the input / output characteristics of the color signal.

  The color signal correction means 006 (FIG. 1) changes the input / output characteristics of the color signal in accordance with the output of the color signal input / output characteristic control means 005 (FIG. 1).

  As an example, the hue is determined from the ratio of the luminance signal and the color signal, but the ratio may change due to the change in the input / output characteristics of the luminance signal, and the hue may become unnatural. For example, when the luminance signal is higher than the color signal in a low luminance region or the like, the color becomes light. Considering this point, the color input / output characteristics are changed with respect to the change in luminance. Here, when the input / output characteristics of the color signal are simply changed to the same as the luminance change ratio, the range of colors that can be expressed differs between Y, C and R, G, B, so the colors of the inputs Y, C are output. It may not be possible to reproduce with Y and C. In consideration of this point, it is a means for realizing the correction of the color signal by changing the input / output characteristics of the color signal in consideration of the relationship between Y, C and R, G, B.

  According to the above form, the first area means achieves high contrast of the luminance part while suppressing noise in the low luminance and high luminance parts, and the second area means effectively uses the part with a small signal distribution, The high contrast of the characteristic part is realized. Further, the point that the dynamic range of the entire screen, which is an adverse effect of the first area means, can be improved by the second area means, and the first area means is that the contrast is partially lowered which is an adverse effect of the second area means. In addition, it is possible to suppress a phenomenon in which the hue becomes unnatural due to a change in the ratio of the luminance and color signals which are harmful to the first area means and the second area means.

FIG. 2 is a conceptual diagram of Embodiment 2 in the present invention.
The index value calculation unit 101 outputs an index value calculated by performing predetermined weighting on the central pixel and its peripheral pixels of the input luminance. The predetermined weight may be a weight for calculating an average value.

  The luminance correction blending rate calculation unit 102 calculates and outputs a mixing ratio for blending the output of the memory unit 103 and the input luminance.

  FIG. 4 is an example of a method for calculating a blending rate for luminance correction. x indicates an index value that is an output of the index value calculation unit 101 (FIG. 2). x1 represents a threshold value indicating a low luminance region. x2 represents a threshold value indicating a high luminance region. y1 represents the intercept of the low luminance region. y2 represents the intercept of the high luminance region. a1 indicates the inclination of the low luminance region. a2 indicates the inclination of the high luminance region. bitFull indicates a bit length representing the maximum value of the input luminance signal. Y_α indicates the mixing ratio. Here, the formula for obtaining the mixing ratio Y_α is Y_α = a1 × (x / bitFull) + y1 when the index value is less than x1. When the index value is greater than or equal to x2, Y_α = a2 × ((bitFull−x) / bitFull) + y2. In other conditions, Y_α = 1. For the set values other than x, it is possible to reduce the amount of calculation by performing a quantitative evaluation using an actual image in advance and calculating an optimum value and using the prepared set value.

  The memory unit 103 (FIG. 2) outputs preset memory data using the input luminance as a read address.

  FIG. 5 shows an example of a method for calculating the correction pixel. Here, the memory data is output with a characteristic that gives a gradation to the input luminance. As the code coefficient, a positive coefficient is output when it is determined as a low luminance area, and a negative coefficient is output when it is determined as a high luminance area. Further, the memory data value can be quantitatively evaluated using an actual image in advance to calculate an optimal value, and the amount of calculation can be reduced by using the set value prepared in advance. Y_α is the output of the luminance correction blending rate calculation unit 102 (FIG. 2). The memory data and code coefficient are output from the memory unit 103 (FIG. 2). Here, the equation for obtaining the output Y as the correction pixel is output Y = Y_α × input luminance + (1−Y_α) × memory data × code coefficient.

  As described above, the luminance correction blending data generation unit 104 (FIG. 2) generates a low value from the input luminance signal, the output of the luminance correction blending rate calculation unit 102 (FIG. 2), and the output of the memory unit 103 (FIG. 2). This is a means for realizing a high contrast of the luminance portion while suppressing noise with respect to the luminance portion and the high luminance portion.

  The threshold value calculation unit 105 (FIG. 2) uses a luminance histogram or the like from any region to capture image features, and to change the input / output characteristics uniformly on the screen so as to have the gradation of the feature portion. Output a control signal.

  FIG. 6 is an example showing the concept of the second area processing. The control signal output is a contrast correction threshold (dark), a contrast correction threshold (bright), a rounding correction gain (dark), and a rounding correction gain (bright). An example of a luminance histogram in the case where a large amount of luminance is distributed in the center will be described. The contrast correction threshold (dark) is the luminance level at the boundary where the distribution increases in the low luminance portion from the input luminance histogram. Sets the luminance level at the boundary where the distribution decreases in the high luminance portion from the input luminance histogram. Further, the rounding correction calculation unit 107 (FIG. 2) performs a process for giving a smooth characteristic such that a step or the like is not conspicuous to signals near the contrast correction threshold values (dark) and (bright). ) And (bright) are control signals for adjusting the degree of rounding correction gain (dark) and (bright). These set values are configured to perform quantitative evaluation using an actual image in advance, calculate an optimum value for each histogram pattern, and select a set value prepared in advance for each input luminance histogram pattern. By doing so, the amount of calculation can be reduced.

  The gain calculation unit 106 (FIG. 2) calculates and outputs a gradient by which the input luminance is multiplied from the output of the threshold value calculation unit 105 (FIG. 2).

  FIG. 7 shows an example of a method for calculating the slope in the gain calculation unit 106 (FIG. 2). Here, the equation for obtaining the inclination is inclination = bitFull / (contrast correction value (bright) −contrast correction value (dark)).

  The rounding correction calculation unit 107 (FIG. 2) is such that a step or the like is not conspicuous with respect to signals near the contrast correction threshold (dark) and (bright) from the input luminance signal and the output of the threshold calculation unit 105 (FIG. 2). Rounding correction data for giving a special characteristic is output.

  FIG. 8 shows an example of a rounding correction data calculation method in the rounding correction calculation unit 107 (FIG. 2). Here, the equation for obtaining the rounding correction data is ((contrast correction threshold (dark) − ((memory address size−1) / 2))) <= correction target pixel <= (contrast correction threshold (dark) + ((memory address In the case of size-1) / 2)), rounding correction memory address = correction target pixel-contrast correction threshold (dark)-((memory address size-1) / 2), and rounding correction data = memory data [rounding correction memory. Address] × rounding correction gain (dark) and (contrast correction threshold (bright) − ((memory address size−1) / 2)) <= correction target pixel <= ((contrast correction threshold (bright) + ((Memory address size-1) / 2)), rounding correction memory address = correction target pixel-contrast correction threshold (bright)-((memory address size-1) / 2) Rounding correction data = −1 × memory data [rounding correction memory address] × rounding correction gain (bright) Under other conditions, rounding correction data = 0. , With a characteristic that smoothes the step near the threshold, performs quantitative evaluation using an actual image in advance, calculates an optimal value, and reduces the amount of calculation by using the preset setting value I can plan.

  The correction unit 108 (FIG. 2) calculates correction data from the output of the threshold calculation unit 105 (FIG. 2), the output of the gain calculation unit 106 (FIG. 2), and the output of the rounding correction calculation unit 107 (FIG. 2). Output.

  FIG. 9 shows an example of a correction data calculation method in the correction unit 108 (FIG. 2). Here, the equation for obtaining the correction data is correction data = (input luminance signal−contrast correction (dark)) × slope + rounding correction data. However, when (input luminance signal−contrast correction (dark)) <0, correction data = 0.

  The color correction index value calculation unit 109 (FIG. 2) calculates the luminance signal change from the input luminance and the output of the correction unit 108 (FIG. 2), and changes the luminance change rate and color correction to change the input / output characteristics of the color signal. Output data.

  As an example, the formula for obtaining the luminance change rate is obtained by: input luminance / output luminance of the correction unit 108 (FIG. 2). The color correction data is data for correcting a phenomenon in which the color tone becomes unnatural when output color signal = color signal × luminance change rate. As a concept, when the input / output characteristics of the color signal are changed to be equivalent to the luminance change ratio, the range of colors that can be expressed is different between Y, C and R, G, B, so the shades of the inputs Y, C are output. It may not be possible to reproduce with Y and C. Based on this point, the correction data changes the input / output characteristics of the color signal in consideration of the relationship between Y, C and R, G, B. The correction data is changed according to the relationship between the luminance and color signal components Cr and Cb.

  The color correction unit 110 (FIG. 2) calculates and outputs corrected color data from the input color signal and the output of the color correction index value calculation unit 109 (FIG. 2). As an example, the corrected color data can be obtained by luminance change rate × input color signal + color correction data.

  With the above configuration, the first area means achieves high contrast in the luminance part while suppressing noise in the low luminance and high luminance parts while reducing the amount of calculation, and the signal distribution is reduced in the second area means. Effective use of a small part realizes high contrast of a characteristic part of an image. Further, the point that the dynamic range of the entire screen, which is an adverse effect of the first area means, can be improved by the second area means, and the first area means is that the contrast is partially lowered which is an adverse effect of the second area means. In addition, it is possible to suppress a phenomenon in which the hue becomes unnatural due to a change in the ratio of the luminance and color signals which are harmful to the first area means and the second area means.

  As mentioned above, although preferred embodiment concerning this invention was described referring an accompanying drawing, it cannot be overemphasized that this invention is not limited to this example. It is obvious for those skilled in the art that various changes or modifications can be conceived within the scope of the technical idea described in the claims. It is understood that it belongs to.

  For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described. Further, a part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of one embodiment.

001 First area luminance signal correction means 002 Second area luminance signal correction means 003 First area input / output characteristic control means 004 Second area input / output characteristic control means 005 Color signal input / output characteristic control means 006 Color signal correction means 101 Index value Calculation unit 102 Luminance correction blending rate calculation unit 103 Memory unit 104 Brightness correction blending data generation unit 105 Threshold calculation unit 106 Gain calculation unit 107 Rounding correction calculation unit 108 Correction unit 109 Color correction index value calculation unit 110 Color correction unit

Claims (2)

A luminance signal input means for inputting a luminance signal;
First correction means for correcting the luminance signal input from the luminance signal input means with a plurality of input / output characteristics different at least in a low luminance portion and a high luminance portion;
Second correction means for correcting the luminance signal corrected by the first correction means with one input / output characteristic that is uniform within the screen;
Luminance signal output means for outputting the luminance signal corrected by the first and second correction means;
A color signal input means for inputting a color signal;
Third correction means for correcting the color signal input from the color signal input means with input / output characteristics corresponding to the input / output characteristics of the first correction means and the input / output characteristics of the second correction means. When,
Color signal output means for outputting the color signal corrected by the third correction means;
An image signal processing apparatus. The image signal processing apparatus according to claim 1,
The second correction unit calculates a histogram of the luminance signal corrected by the first correction unit within a predetermined area of the image, and uses a uniform input / output characteristic in the screen based on the calculated histogram. Determining an image signal processing apparatus.

Images