CN104299596A - Image processing method, method for desaturating image data, and image processing device - Google Patents

Abstract

An image processing method, a method for unsaturated image data and an image processing device are provided. The image processing method comprises the following steps: analyzing the frame to obtain an original gain value corresponding to a plurality of gray-scale values of the frame, wherein the original gain value is higher than 1; obtaining a look-up table, wherein the look-up table records the conversion relation between a plurality of preset gray-scale values and a plurality of preset gain values, and is used when the image data of the picture is subjected to anti-saturation processing; utilizing the look-up table to perform anti-saturation processing on the picture according to the original gain value to obtain an adjusted conversion relation of the picture, wherein the adjusted conversion relation defines the conversion relation between different gray scale values of the picture and corresponding adjusted gain values; and performing gain processing on the image data of the picture according to the adjusted conversion relation.

Description

Image processing method, method for desaturating image data, and image processing device

technical field

The present invention relates to an image processing device, and more particularly to an image processing method, a method for desaturating image data, and an image processing device using the method.

Background technique

FIG. 1 is a diagram illustrating a conventional image processing device. The global gain (global gain) generator 110 determines the gain value G according to the operating requirements of the system, and outputs the gain value G to the calculation circuit 120 . The calculation circuit 120 receives the original image data D ₁ , calculates the product of the gain value G and the image data D ₁ , and then outputs the product (ie G*D ₁ ) to the display circuit 130 as new image data D ₂ . The display circuit 130 drives the display panel according to the image data _D2 to display the corresponding image frame.

Under normal operating conditions, the global gain generator 110 sets the gain value G to 1. FIG. 2 is a diagram illustrating the conversion curve of image data _D1 to image data _D2 when the gain value G in FIG. 1 is 1. FIG. In FIG. 2 , the horizontal axis represents the original image data D ₁ , and the vertical axis represents the new image data D ₂ converted by the calculation circuit 120 . Assume that both the image data _D1 and the image data _D2 are 8-bit data. Under the condition that the gain value G is 1, the value range of the image data D ₁ is 0-255, and the value range of G*D ₁ (ie, the image data D ₂ ) is also 0-255.

When the global gain generator 110 sets the gain value G to be greater than 1 (for example, 1.2), the new image data D ₂ converted by the calculation circuit 120 may be saturated. FIG. 3 is a diagram illustrating the conversion curve of the image data D ₁ to the image data D ₂ when the gain value G in FIG. 1 is 1.2. In FIG. 3 , the horizontal axis represents the original image data D ₁ , and the vertical axis represents the new image data D ₂ converted by the calculation circuit 120 . Assume that both the image data _D1 and the image data _D2 are 8-bit data. Under the condition that the gain value G is 1.2, the value range of the image data D ₁ is 0 to 255, and the value range of G*D ₁ is 0 to 306. However, because the image data D ₂ is 8-bit data, that is, the maximum value of the image data D ₂ is 255, so when the image data D ₁ is in the range of 213 to 255, the image data D ₂ is uniformly 255. This phenomenon is called "saturation". Due to the saturation of the image data _D2 , the details of the high grayscale (high brightness) picture will disappear. The larger the gain value G is, the larger the value range of saturation occurs, and the more serious the loss of high-brightness details is.

Contents of the invention

The invention provides an image processing method, a method for desaturating image data and an image processing device using the method to avoid saturation of image data.

An image processing method of the present invention is used for performing image processing on image data of several frames of an image processing device. Each picture includes several pixel units, and each pixel unit includes several sub-pixels, respectively corresponding to several colors. The image processing method includes: analyzing one of the several pictures to obtain an original gain value corresponding to several gray scale values of the picture, wherein the original gain value is higher than 1; obtaining a preset data pair gain The conversion relationship look-up table, wherein the preset data records several preset gray scale values and several corresponding preset gain values for the gain conversion look-up table, which are used together when the image data of the several pictures are desaturated. Use; according to the original gain value, use the preset data to gain conversion relationship lookup table to perform anti-saturation processing on the picture, and obtain the adjusted data-to-gain conversion relationship of the picture, wherein the adjusted data and gain The conversion relationship defines the conversion relationship between different grayscale values of the picture and the corresponding adjusted gain values; and performs gain processing on the image data of the picture according to the adjusted data-to-gain conversion relationship.

A method for desaturating image data of the present invention is used for an image processing device to perform image processing. The method includes: analyzing a picture among several pictures to obtain the original conversion relationship curve of the picture, the original conversion relationship curve indicates that several grayscale values of the picture are to be processed according to the original gain value, wherein the original The gain value is greater than 1; a preset conversion relationship curve is obtained, wherein the preset conversion relationship curve is used for desaturation processing of the image data of the several pictures; by merging the preset conversion relationship curve and the original conversion relationship curve to perform anti-saturation processing on the picture to obtain an adjusted conversion relationship curve; and perform gain processing on the image data of the picture according to the adjusted conversion relationship curve.

An image processing device of the present invention is used for performing image processing on image data of several frames, wherein each frame includes several pixel units, and each pixel unit includes several sub-pixels respectively corresponding to several colors. The image processing device includes a global gain generator, a conversion relation generator and a conversion circuit. The global gain generator analyzes a frame among the plurality of frames to obtain an original gain value corresponding to several gray scale values of the frame, wherein the original gain value is higher than 1. The conversion relationship generator is coupled to the global gain generator to receive the original gain value. The conversion relationship generator obtains a preset data-to-gain conversion relationship look-up table, wherein the preset data-to-gain conversion look-up table records several preset gray scale values and several corresponding preset gain values for the data It is used together when desaturation processing is performed on the image data of two screens. The conversion relationship generator uses the preset data to look up the gain conversion relationship table according to the original gain value to perform anti-saturation processing on the picture, and obtain the adjusted data-to-gain conversion relationship of the picture, wherein the adjusted data The conversion relationship with gain defines the conversion relationship between different grayscale values of the picture and corresponding adjusted gain values. The conversion circuit is coupled to the conversion relationship generator for performing gain processing on the image data of the frame according to the adjusted data-to-gain conversion relationship.

An image processing device of the present invention is used for image processing. The image processing device includes a global gain generator, a conversion relation generator and a conversion circuit. The global gain generator analyzes one of the several pictures to obtain the original conversion relation curve of the picture, and the original conversion relation curve indicates that several grayscale values of the picture are to be processed according to an original gain value, wherein the original Gain values greater than 1. The conversion relationship generator is coupled to the global gain generator to receive the original gain value. The conversion relationship generator obtains a preset conversion relationship curve, wherein the preset conversion relationship curve is used jointly when the image data of the plurality of frames are desaturated. The conversion relationship generator combines the preset conversion relationship curve and the original conversion relationship curve to perform an anti-saturation process on the image, so as to obtain an adjusted conversion relationship curve. The conversion circuit is coupled to the conversion relationship generator for performing gain processing on the image data of the frame according to the adjusted conversion relationship curve.

Based on the above, the image processing method, the method for desaturating image data, and the image processing device using the method described in the embodiments of the present invention combine at least two conversion curves (or conversion relationships) to determine the conversion of image data. Gain value to avoid saturation of image data and preserve high brightness details in the image.

In order to make the above-mentioned features and advantages of the present invention more comprehensible, the following specific embodiments are described in detail together with the accompanying drawings.

Description of drawings

FIG. 1 is a diagram illustrating a conventional image processing device.

FIG. 2 is a graph illustrating the conversion relationship between image data D 1 and image data _{D 2} when the original gain value G in FIG. 1 is 1. FIG.

FIG. 3 is a graph illustrating the conversion relationship between the image data D 1 and the image data _{D 2 when} the original gain value G in FIG. 1 is 1.2.

FIG. 4 is a schematic flowchart illustrating a method for desaturating image data according to an embodiment of the present invention.

FIG. 5 is a schematic circuit block diagram illustrating an image processing device according to an embodiment of the present invention.

FIG. 6 is a schematic diagram illustrating a transformation relationship curve between original image data D 1 and new image data _{D 3 when} the original gain value G in FIG. 5 is 1.6 according to an embodiment of the present invention.

FIG. 7 is a schematic diagram illustrating the conversion relationship between the original image data D ₁ and the gain value when the original gain value G in FIG. 5 is 1.6 according to another embodiment of the present invention.

FIG. 8 is a schematic circuit block diagram illustrating an image processing device according to another embodiment of the present invention.

FIG. 9 is a schematic flowchart illustrating a method for desaturating image data according to another embodiment of the present invention.

FIG. 10 is a schematic circuit block diagram illustrating an image processing device according to yet another embodiment of the present invention.

[Description of labels]

110, 510, 810: global gain generator 120: calculation circuit

130: display circuit 520, 820: conversion relationship generator

530, 830: conversion circuit 610: original conversion relationship curve

620: preset conversion relationship curve 630, 940: intersection point

640: Intersection range 821: Lookup table

822: Curve merging unit 831: Comparison circuit

832: Gain Determiner 833: Calculator

910: Preset conversion relationship curve 920: Original conversion relationship curve

930: Area 950: Adjusted Conversion Relationship Curve

D ₁ : Raw image data D ₂ , D ₃ , D ₄ : Image data

G: Original gain value G’: Adjusted gain value

M: Represents the grayscale value

S410～S440, S710～S740: steps

Detailed ways

The term "coupling" used throughout the specification of this case (including the scope of claims) may refer to any direct or indirect means of connection. For example, if it is described that a first device is coupled to a second device, it should be interpreted that the first device can be directly connected to the second device, or the first device can be connected through other devices or some kind of connection means. indirectly connected to the second device. In addition, wherever possible, elements/components/steps using the same reference numerals in the drawings and embodiments represent the same or similar parts. Elements/components/steps using the same symbols or using the same terms in different embodiments can refer to related descriptions.

FIG. 4 is a schematic flowchart illustrating a method for desaturating image data according to an embodiment of the present invention. FIG. 5 is a schematic circuit block diagram illustrating an image processing device according to an embodiment of the present invention. Referring to FIG. 4 and FIG. 5 , the image processing device includes a global gain generator 510 , a conversion relationship generator 520 and a conversion circuit 530 . The image processing device may perform image processing on image data of multiple frames, wherein these frames respectively include multiple pixel units, each pixel unit includes multiple sub-pixels, and these sub-pixels respectively correspond to multiple colors.

The conversion circuit 530 is coupled to the global gain generator 510 and the conversion relationship generator 520 . The global gain generator 510 can be analogized with reference to the related description of the global gain generator 110 in FIG. 1 . In step S410 , the global gain generator 510 may analyze one frame among the multiple frames in the image data to obtain an original transformation relation curve of the frame. The original conversion relation curve indicates that multiple grayscale values of the picture are to be processed according to the original gain value G. In some embodiments, the original conversion relationship curve may be a data-to-data curve, respectively defining a corresponding relationship between gray scale values before gain conversion and gray scale values after gain conversion. For example, FIG. 2 or FIG. 3 can be regarded as a special example of the original transformation relationship curve. The global gain generator 510 may provide the original gain value G representing the original conversion relationship curve to the conversion circuit 530 .

When the original gain value G is not greater than 1, the original gain value G will not saturate the image data _D3 , so the conversion circuit 530 can choose to use the original conversion curve representing the original gain value G to convert the original image data _D1 to new image data D ₃ . This embodiment does not limit the conversion method of the conversion circuit 530 to convert the original image data _D1 into the new image data _D3 . For example, the conversion circuit 530 can select/make a corresponding lookup table (lookup table, LUT) according to the original gain value G (the original conversion relationship curve), and then use the corresponding lookup table to convert the original image data _D1 into a new image Data D ₃ . For another example, the conversion circuit 530 may calculate the product of the original gain value G and the image data D ₁ (ie, G*D ₁ ) as the image data D ₃ . The conversion circuit 530 can output the image data _D3 to a display circuit (not shown). The display circuit can drive a display panel (not shown) according to the image data _D3 to display corresponding image frames.

In step S420 , the conversion relationship generator 520 can obtain a preset conversion relationship curve, wherein the preset conversion relationship curve can be commonly used when the image data of multiple frames are desaturated. The preset conversion relationship curve may be determined according to design requirements. For example, in some embodiments, if the original conversion relationship curve is a data-to-data curve, the preset conversion relationship curve may be a data-to-data curve, which at least covers the highest grayscale applied by the image processing device In the high grayscale (high brightness) range of values, the slope of the preset conversion relationship curve is greater than 0 and less than or equal to 1. Therefore, when the original gain value G provided by the global gain generator 510 is greater than 1, that is, when the slope of the original conversion relationship curve is greater than 1, there is a certain relationship between the original conversion relationship curve and the preset conversion relationship curve. intersection. For example, FIG. 6 is a schematic diagram illustrating a transformation relationship curve between the original image data D 1 and the new image data _{D 3 when} the original gain value G in FIG. 5 is 1.6 according to an embodiment of the present invention. In FIG. 6 , the horizontal axis represents the original image data D ₁ , and the vertical axis represents the new image data D ₃ converted by the conversion circuit 530 . Assume that the image data _D1 and the image data _D3 are both 8-bit data, so the value range of the image data _D1 is 0 to 255, and the value range of the image data _D3 is also 0 to 255. Curve 610 represents the original conversion curve corresponding to the original gain value G. As shown in FIG. 6 , the original conversion relationship curve 610 includes: an oblique straight line with a slope equal to the original gain value G and a horizontal straight line with a slope equal to zero. Obviously, under the condition that the original gain value G is 1.6, the original conversion relation curve 610 representing the original gain value G is “saturated” within the range of 160 to 255 of the image data D ₁ .

FIG. 6 also shows a preset conversion relationship curve 620 provided by the conversion relationship generator 520, wherein the slope of the preset conversion relationship curve 620 is greater than 0 in the high gray scale (high brightness) range, even the preset conversion relationship curve 620 It is assumed that all slopes of the conversion relationship curve 620 are greater than zero. Therefore, the preset conversion relationship curve 620 will not be "saturated". The preset conversion relationship curve 620 can be regarded as a set of maximum gain values acceptable to the system. The preset conversion relationship curve 620 can be determined according to the actual product design requirements. In this embodiment, different preset conversion relationship curves 620 may be used for image data of different colors in the frame. The preset conversion relationship curve 620 can be realized by a look-up table or other methods. Since the original gain value G provided by the global gain generator 510 is greater than 1 (eg, G=1.6), there is an intersection 630 between the original conversion relationship curve 610 and the preset conversion relationship curve 620 . Among them, in the range of low grayscale values where the grayscale value before conversion (image data D ₁ ) is smaller than the grayscale value of the intersection point corresponding to the intersection point 630, the grayscale value after conversion on the original conversion relationship curve 610 (image data D ₃ ) All are less than or equal to the converted gain value (image data D ₃ ) on the preset conversion relationship curve 620; and the gray scale value before conversion (image data D ₁ ) is higher than the high gray scale value corresponding to the intersection point 630 Within the value range, the converted grayscale values (image data D ₃ ) on the original conversion relationship curve 610 are greater than or equal to the converted gain values (image data D ₃ ) on the preset conversion relationship curve 620 .

Please refer to FIG. 4 and FIG. 5 , when the original gain value G provided by the global gain generator 510 is greater than 1, the original gain value G may saturate the image data _D3 , so the conversion circuit 530 can dynamically select a representative original The original conversion relationship curve of the gain value G or the preset conversion relationship curve provided by the conversion relationship generator 520 is used to convert the original image data D ₁ into the new image data D ₃ . The conversion circuit 530 may perform step S430 to perform anti-saturation processing on the frame by merging the preset conversion relationship curve and the original conversion relationship curve, so as to obtain an adjusted conversion relationship curve. In some embodiments, the adjusted conversion relationship curve may be a data-to-data curve, so as to respectively define the corresponding relationship between the gray scale value before gain conversion and the gray scale value after gain conversion.

For example, as shown in FIG. 6 , when the image data D ₁ is smaller than the intersection point 630 , the conversion circuit 530 may choose to use the original conversion relationship curve 610 as the adjusted conversion relationship curve to convert the image data D ₁ into the image data D ₃ . When the image data D ₁ is greater than the intersection point 630 , the conversion circuit 530 may choose to use the preset conversion relationship curve 620 as the adjusted conversion relationship curve to convert the image data D ₁ into the image data D ₃ . That is to say, the conversion circuit 530 can select the gray scale value (image data D ₁ ) in the original conversion relationship curve 610 that is lower than the intersection gray level value corresponding to the intersection point 630 of the preset conversion relationship curve 620 and the original conversion relationship curve 610 A line segment is used as the first part of the adjusted conversion relationship curve; and a part of the line segment in the preset conversion relationship curve 620 whose gray scale value is higher than the intersection point gray scale value is selected as the second part of the adjusted conversion relationship curve.

For another example, the conversion circuit 530 may perform "smooth processing" on the adjusted conversion relationship curve, so that part of the line segments covering the original conversion relationship curve 610 and the preset conversion relationship curve 620 on the adjusted conversion relationship curve have a Smooth connection. The conversion circuit 530 can define an "intersection point range", wherein the intersection point of the original conversion relationship curve and the default conversion relationship curve is located in the intersection point range. The conversion circuit 530 can compare the image data D ₁ with the intersection range (such as the intersection range 640 shown in FIG. 6 ) in step S430 . In the embodiment shown in FIG. 6, the conversion circuit 530 can also perform smoothing processing on the original conversion relationship curve 610 and the preset conversion relationship curve 620 located in the intersection range 640, so that the difference between the original conversion relationship curve 610 and the preset conversion relationship curve 620 There is a smooth connection between them. In some embodiments, the rounding process includes low pass filter (Low pass filter) processing. The conversion circuit 530 can also convert the original conversion relationship curve 610 and the preset conversion relationship curve 620 in the intersection range 640 into a smooth connecting line by low-pass filtering, as shown in FIG. 6 . Therefore, when the image data D ₁ falls within the intersection point range 640 , the conversion circuit 530 may choose to use this rounded connection line to convert the image data D ₁ into the image data D ₃ .

After step S430 is completed, the conversion circuit 530 may proceed to step S440 to perform gain processing on the image data D ₁ of the frame according to the adjusted conversion relationship curve in step S430 . For example, when the image data D ₁ is smaller than the intersection range 640 , the conversion circuit 530 may choose to use the original conversion relationship curve 610 to convert the image data D ₁ into the image data D ₃ . When the image data D ₁ is larger than the intersection point range 640 , the conversion circuit 530 may choose to use the preset conversion relationship curve 620 to convert the image data D ₁ into the image data D ₃ . When the image data D ₁ falls within the intersection range 640 , the conversion circuit 530 may choose to use the rounded connection line of the intersection range 640 to convert the image data D ₁ into the image data D ₃ .

The original conversion relationship curve, the preset conversion relationship curve and the adjusted conversion relationship curve mentioned in the present invention should not be limited to the relevant description of the implementation example shown in FIG. 6 . For example, in some other embodiments, the original conversion relationship curve, the preset conversion relationship curve and the adjusted conversion relationship curve may be data-to-gain curves. Wherein, at least one corresponding gain value on the preset conversion relationship curve is greater than 0 and less than or equal to 1 in the high gray scale range covering at least the highest gray scale value applied by the image processing device.

FIG. 7 is a schematic diagram illustrating the conversion relationship between the original image data D ₁ and the gain value when the original gain value G in FIG. 5 is 1.6 according to an embodiment of the present invention. In FIG. 7 , the horizontal axis represents the original image data D ₁ , and the vertical axis represents the gain value. The curve 910 shown in FIG. 7 represents the preset conversion relationship curve provided by the conversion relationship generator 520 . In this embodiment, different preset conversion relationship curves 910 may be used for image data of different colors in the frame. Curve 920 shown in FIG. 7 represents the original conversion curve corresponding to the original gain value G (eg, G=1.6). In this embodiment, the original conversion relationship curve 920 is a horizontal straight line passing through the original gain value G. As shown in FIG.

The lower part of FIG. 7 shows an enlarged schematic view of the area 930 shown in the upper part of FIG. 7 . It is assumed that the image data D ₁ and the image data D ₃ are both 8-bit data, so the value range of the image data D ₁ and the image data D ₃ is 0 to 255. Where the original image data D ₁ is the maximum value (ie, 255), the preset gain value of the preset conversion relationship curve 910 is 1. In addition, in the low grayscale range covering at least the lowest grayscale value applied by the image processing device, that is, in the low grayscale (low brightness) range of the image data _D1 , at least A corresponding gain value is greater than the original gain value G provided by the global gain generator 510 . Therefore, when the original gain value G provided by the global gain generator 510 is greater than 1 (for example, G=1.6), there is an intersection point 940 between the preset conversion relationship curve 910 and the original conversion relationship curve 920 representing the original gain value G.

Because the original gain value G is greater than 1, the original conversion relationship curve 920 may be "saturated" in the high grayscale (high brightness) range of the image data _D1 . Since the gain value of the default conversion relationship curve 910 converges to 1 in the high grayscale (high brightness) range of the image data _D1 , the default conversion relationship curve 910 will not be "saturated". The preset conversion relationship curve 910 can be regarded as a set of maximum gain values acceptable to the system. The preset conversion relationship curve 910 can be determined according to the actual product design requirements. The preset conversion relationship curve 910 can be realized by a look-up table or other methods.

Please refer to FIG. 5 and FIG. 7, when the original gain value G provided by the global gain generator 510 is greater than 1, the original gain value G may saturate the image data _D3 , so the conversion circuit 530 can be based on the image data _D1 . The original image data D ₁ is converted to a gain value by dynamically selecting to use the original conversion relationship curve 920 or the preset conversion relationship curve 910 . For example, as shown in FIG. 7, when the image data _D1 is smaller than the intersection point 940, the conversion circuit 530 may choose to use the original conversion relationship curve 920 as the first part of the adjusted conversion relationship curve 950, so as to convert the image data _D1 into the gain value. When the image data _D1 is greater than the intersection point 940, the conversion circuit 530 may choose to use the preset conversion relationship curve 910 as the second part of the adjusted conversion relationship curve 950 to convert the image data _D1 into the gain value. The conversion circuit 530 can also use the gain value of the adjusted conversion relationship curve 950 to convert the image data D ₁ into the image data D ₃ . Since the adjusted conversion relationship curve 950 converges to 1 in the high-grayscale (high-brightness) range of the image data _D1 , it can ensure that the gained image data _D3 will not be "saturated".

In the embodiment shown in FIG. 7 , the conversion circuit 530 can select a part of the preset conversion relationship curve 910 that is smaller than the original gain value G (for example, 1.6), and select a part of the original conversion relationship curve 920 where the image data _D1 is smaller than the intersection point 940 A part of the curve is used as the adjusted conversion relationship curve 950 . In the embodiment shown in FIG. 7 , the conversion circuit 530 can also perform smoothing processing on the adjusted conversion relationship curve 950 so that the adjusted conversion relationship curve 950 has a smooth connection, as shown in FIG. 7 . For example, in some embodiments, the rounding includes low-pass filtering.

The implementation of the image processing device should not be limited to the related description of FIG. 5 . For example, FIG. 8 is a schematic circuit block diagram illustrating an image processing device according to another embodiment of the present invention. The embodiment shown in FIG. 8 can be deduced by referring to the related description of FIG. 5 . Referring to FIG. 4 and FIG. 8 , the image processing device includes a global gain generator 810 , a conversion relationship generator 820 and a conversion circuit 830 . The conversion relationship generator 820 is coupled to the global gain generator 810 to receive the original gain value G, that is, an original conversion relationship curve (such as the original conversion relationship curve 610 shown in FIG. 6 or the original conversion relationship curve 920 shown in FIG. 7 ).

The global gain generator 810 can be analogized with reference to the related description of the global gain generator 110 in FIG. 1 or the global gain generator 510 in FIG. 5 . In step S410 , the global gain generator 810 determines the original gain value G according to the operating requirements of the system, and provides the original gain value G to the transformation relationship generator 820 .

The conversion relationship generator 820 is also coupled to the conversion circuit 830 . When the original gain value G is not greater than 1, the original gain value G will not saturate the image data _D4 , so the conversion relationship generator 820 can choose to transmit the original gain value G provided by the global gain generator 810 to the conversion circuit 830 . The conversion circuit 830 can convert the original image data D ₁ into new image data D ₄ according to the original gain value G transmitted by the conversion relationship generator 820 . This embodiment does not limit the conversion method of the conversion circuit 830 to convert the original image data _D1 into the new image data _D4 . For example, the conversion circuit 830 can select/create a corresponding lookup table (LUT) according to the original gain value G, and then use the corresponding lookup table to convert the original image data D ₁ into the new image data D ₄ . For another example, the conversion circuit 830 can calculate the product of the original gain value G and the image data D ₁ (ie, G*D ₁ ) as the image data D ₄ . The conversion circuit 830 can output the image data _D4 to a display circuit (not shown). The display circuit can drive a display panel (not shown) according to the image data _D4 to display corresponding image frames.

In step S420, the conversion relationship generator 820 obtains a preset conversion relationship curve (such as the preset conversion relationship curve 620 shown in FIG. 6 or the preset conversion relationship curve 910 shown in FIG. 7 ), wherein the preset conversion relationship curve can be used for It is commonly used when desaturation processing is performed on image data of multiple screens.

In step S430, the conversion relationship generator 820 combines the preset conversion relationship curve in step S420 and the original conversion relationship curve in step S410 to desaturate the image, thereby obtaining an adjusted conversion relationship curve. For example, taking FIG. 6 as an example, please refer to FIG. 6 and FIG. 8 . In step S430 , the conversion relationship generator 820 can desaturate the picture through the preset conversion relationship curve 620 and the original conversion relationship curve 610 , so as to obtain the adjusted conversion relationship curve. For another example, take FIG. 7 as an example, please refer to FIG. 7 and FIG. 8 . In step S430 , the conversion relationship generator 820 can desaturate the picture through the preset conversion relationship curve 910 and the original conversion relationship curve 920 , so as to obtain the adjusted conversion relationship curve 950 .

After step S430 is completed, the conversion circuit 830 may perform step S440 to perform gain processing on the image data of the frame according to the adjusted conversion relationship curve.

The implementation of the image processing method should not be limited to the description in FIG. 4 . For example, FIG. 9 is a schematic flowchart illustrating a method for desaturating image data according to another embodiment of the present invention. The embodiment shown in FIG. 9 can be deduced by referring to the relevant description of FIG. 4 . The method shown in FIG. 9 can be applied to the image processing device shown in FIG. 5 or the image processing device shown in FIG. 8 .

Referring to FIG. 5 and FIG. 9 , the image processing device may perform image processing on image data of multiple frames, wherein each frame includes multiple pixel units, and each pixel unit includes multiple sub-pixels corresponding to multiple colors. The image processing device includes a global gain generator 510 , a conversion relation generator 520 and a conversion circuit 530 . The global gain generator 510 may perform step S710 to analyze a frame among the multiple frames to obtain an original gain value G corresponding to multiple grayscale values of the frame. For example, taking FIG. 7 as an example, the original gain value G in step S710 may be the original conversion relation curve 920 .

In step S720, the conversion circuit 530 can obtain a preset data-to-gain conversion relationship look-up table provided by the conversion relationship generator 520, wherein the preset data-to-gain conversion look-up table records a plurality of preset gray scale values and respective corresponding Corresponding multiple preset gain values. The preset data-to-gain conversion look-up table can be used in common when desaturation processing is performed on the image data of multiple frames. For example, taking FIG. 7 as an example, the preset data-to-gain conversion relationship look-up table in step S720 may record the preset conversion relationship curve 910 .

In step S730, the conversion circuit 530 can use the preset data in step S720 to look up the gain conversion relationship table according to the original gain value in step S710 to perform anti-saturation processing on the picture, and obtain the adjusted data pair of the picture Gain conversion relationship, wherein the adjusted data and gain conversion relationship define the conversion relationship between different grayscale values of the image and corresponding adjusted gain values. For example, taking FIG. 7 as an example, the adjusted data versus gain conversion relationship in step S730 may be an adjusted conversion relationship curve 950 .

In step S740, the conversion circuit 530 may perform gain processing on the image data of the frame according to the adjusted data-to-gain conversion relationship in step S730. The implementation manner of step S740 is not limited here. For example, the conversion circuit 530 may obtain multiple grayscale values of the picture, and obtain multiple adjusted gain values corresponding to the multiple grayscale values of the picture according to the adjusted data-to-gain conversion relationship in step S730 and using the multiple adjusted gain values, the conversion circuit 530 performs gain processing for the multiple gray scale values of the picture respectively.

The method shown in FIG. 9 can also be applied to the image processing device shown in FIG. 8 . Please refer to FIG. 8 and FIG. 9 , the global gain generator 810 analyzes the frame at step S710 to obtain an original gain value G corresponding to multiple grayscale values of the frame. The conversion relation generator 820 is coupled to the global gain generator 810 to receive the original gain value G. As shown in FIG. For example, taking FIG. 7 as an example, the original gain value G in step S710 may be the original conversion relation curve 920 .

In step S720, the conversion relationship generator 820 further obtains a preset data-to-gain conversion relationship look-up table. For example, taking FIG. 7 as an example, the preset data-to-gain conversion relationship look-up table in step S720 may record the preset conversion relationship curve 910 .

In step S730, the conversion relationship generator 820 can perform anti-saturation processing on the picture by using the preset data-to-gain conversion relationship look-up table in step S720 according to the original gain value G, and obtain the adjusted data-to-gain conversion relationship of the picture. . In this embodiment, step S730 may set the adjusted gain values corresponding to at least one higher gray-scale value among the plurality of gray-scale values of the picture to be equal to the preset gain values, and set the picture The adjusted gain values corresponding to at least one lower gray-scale value among the plurality of gray-scale values are set to be equal to the original gain value. For example, the conversion relationship generator 820 can directly use the preset data to generate the conversion relationship between the preset gray scale values and the corresponding preset gain values that belong to a lower gain range in the gain conversion look-up table. The unit 820 may set the adjusted gain values corresponding to all the grayscale values in the higher gain range to be equal to the original gain value, so as to obtain the higher gain range part of the adjusted data-to-gain conversion relationship. The lower gain range includes at least one gain value, and the lower boundary gain value of the lower boundary range corresponds to the highest gray scale value applied by the image processing device. The higher gain range includes at least one gain value, and a higher boundary gain value of the higher gain range corresponds to the lowest gray scale value used by the image processing device.

The conversion relationship generator 820 can also obtain a specific gain value according to at least one of the original gain value G and the preset data-to-gain conversion relationship look-up table, and determine the higher gain range and the lower gain range according to the specific gain value. gain range. Wherein, a higher boundary gain value of the lower gain range is lower than or equal to the specific gain value, and a lower boundary gain value of the higher gain range is higher than or equal to the specific gain value. In some embodiments, the conversion relationship generator 820 can regard the original gain value G as the specific gain value.

In the high grayscale range covering at least the highest grayscale value applied by the image processing device, at least one corresponding preset gain value in the gain conversion relationship look-up table of the preset data is 1. In the low grayscale range covering at least the lowest grayscale value applied by the image processing device, at least one corresponding preset gain value on the gain conversion relationship look-up table of the preset data pair is higher than the original gain value g.

Taking FIG. 7 as an example, step S730 can directly use the part (lower gain boundary range) of the preset conversion relationship curve 910 with a higher gray scale value (for example, greater than the intersection point 940) to be set as the lower part of the adjusted conversion relationship curve 950. The part of the low gain range, wherein the lower boundary gain value (eg gain value 1) of the lower boundary range corresponds to the highest grayscale value (eg grayscale value 255) applied by the image processing device. Step S730 can also use the part (higher gain boundary range) of the original conversion relationship curve 920 (ie, the original gain value) shown in FIG. The upper gain range part of 950, wherein the upper boundary gain value of the upper boundary range corresponds to the lowest grayscale value (eg grayscale value 0) applied by the image processing device. In the high gray scale range covering the highest gray scale value (for example, the gray scale value is 255), at least one corresponding preset gain value on the preset conversion relationship curve 910 is 1. In the low gray scale range covering the lowest gray scale value (for example, the gray scale value is 0), the corresponding preset gain values on the preset conversion relationship curve 910 are all higher than the original gain value G. As shown in FIG.

The conversion circuit 830 is coupled to the conversion relationship generator 820 . In step S740, the conversion circuit 830 may perform gain processing on the image data of the frame according to the adjusted data-to-gain conversion relationship in step S730.

FIG. 10 is a schematic circuit block diagram illustrating an image processing device according to yet another embodiment of the present invention. The embodiment shown in FIG. 10 can be deduced by referring to the relevant description of FIG. 8 . The difference from the embodiment shown in FIG. 8 is that the conversion relationship generator 820 shown in FIG. 8 includes a lookup table 821 and a curve merging unit 822 . Referring to FIG. 7 , FIG. 9 and FIG. 10 , the lookup table 821 can record and provide the preset conversion relationship curve 910 to the curve merging unit 822 (step S720 ). The curve combining unit 822 is coupled to the look-up table 821 and the global gain generator 810 . The curve merging unit 822 selects in step S730 the part of the preset conversion relationship curve 910 that is smaller than the original value G (for example, 1.6, determined by the global gain generator 810), and selects the image data _D1 in the original conversion relationship curve 920 that is smaller than the intersection point 940 part, as the adjusted conversion relationship curve 950 (ie, the adjusted data-to-gain conversion relationship), so as to provide the adjusted conversion relationship curve 950 to the conversion circuit 830 .

In some embodiments, the conversion circuit 830 includes a gain determiner 832 and a calculator 833 . The gain determining circuit 832 is coupled to the conversion relationship generator 820 . The gain determination circuit 832 obtains multiple grayscale values of the frame, and obtains multiple adjusted gain values G' corresponding to the multiple grayscale values of the frame according to the adjusted data-to-gain conversion relationship in step S730. The calculator 833 is coupled to the gain determining circuit 832 . The calculator 833 uses the adjusted gain value G' to perform gain processing for the multiple grayscale values of the frame (step S740).

In other embodiments, the conversion circuit 830 shown in FIG. 8 may further include a comparison circuit 831 coupled to the gain determination circuit 832 . The comparison circuit 831 receives the original image data D ₁ , and obtains a representative gray scale of each pixel unit according to a plurality of individual gray scale values of a plurality of sub-pixels (sub-pixels) of each pixel unit in the original image data D ₁ The value M is used as one of the plurality of grayscale values. The gain determination circuit 832 obtains the representative grayscale value M provided by the comparison circuit 831, and according to the adjusted data-to-gain conversion relationship in step S730, obtains a plurality of adjusted gain values corresponding to the representative grayscale values M of these pixel units respectively G'. The calculator 833 uses the adjusted gain value G′ to perform gain processing for the corresponding gray scale values in the original image data _D1 (step S740 ).

The implementation of the comparison circuit 831 is not limited here. For example, the comparison circuit 831 can find the highest gray-scale value from among the multiple individual gray-scale values of the multiple sub-pixels of each pixel unit as the representative gray-scale value M of the pixel unit. For example, assuming that the original image data _D1 is a pixel unit data, and the pixel unit data includes sub-pixel data such as red (R), green (G), blue (B), etc., the comparison circuit 831 can obtain from the same pixel The maximum value (highest grayscale value) is selected from the three sub-pixel data of R, G, and B of the unit as the representative grayscale value M of the pixel unit. The gain determiner 832 is coupled to the comparison circuit 831 to receive the representative gray scale value M. As shown in FIG. The gain determiner 832 finds the adjusted gain value G′ from the adjusted data-to-gain conversion relationship provided by the conversion relationship generator 820 (eg, the adjusted conversion relationship curve 950 shown in FIG. 7 ) according to the representative grayscale value M.

The calculator 833 is coupled to the gain determiner 832 to receive the adjusted gain value G′. The calculator 833 can use the adjusted gain value G′ provided by the gain determiner 832 and the original image data D ₁ to calculate the gained image data D ₄ . This embodiment does not limit the manner in which the calculator 833 calculates the image data _D4 after gain. For example, the calculator 833 may include a multiplier, wherein the multiplier may calculate the product of the adjusted gain value G′ and the original image data D ₁ (ie, G′*D ₁ ) as the image data D ₄ . The calculator 833 can output the image data _D4 to a display circuit (not shown). The display circuit can drive a display panel (not shown) according to the image data _D4 to display corresponding image frames.

The method shown in FIG. 4 can also be applied to the image processing device shown in FIG. 10 . Referring to FIG. 4 , FIG. 8 and FIG. 9 , the global gain generator 810 analyzes the picture in step S410 to obtain the original gain value G (ie, the original conversion relationship curve) corresponding to the multiple grayscale values of the picture. For example, taking FIG. 6 as an example, the original gain value G in step S410 may be the original conversion relationship curve 610 . Taking FIG. 7 as an example, the original gain value G in step S410 may be the original conversion relationship curve 920 .

The conversion relationship generator 820 further obtains a preset conversion relationship curve in step S420 . For example, taking FIG. 6 as an example, the preset conversion relationship curve in step S420 may be the curve 620 shown in FIG. 6 . Taking FIG. 7 as an example, the preset conversion relationship curve in step S420 may be the curve 910 shown in FIG. 7 . The lookup table 821 can record and provide the preset conversion relationship curve.

The curve combining unit 822 is coupled to the look-up table 821 and the global gain generator 810 . In step S430, the curve merging unit 822 selects a partial line segment in the original conversion relationship curve whose gray scale value is lower than the gray level value of the intersection point corresponding to the intersection point of the preset conversion relationship curve and the original conversion relationship curve, as the adjusted conversion relationship The first part of the curve. In step S430, the curve merging unit 822 selects a part of the preset conversion relationship curve whose grayscale value is higher than the intersection point grayscale value as the second part of the adjusted conversion relationship curve. For example, if Fig. 6 is taken as an example, in step S430, the curve merging unit 822 selects a part of the line segment whose grayscale value is lower than the grayscale value of the intersection point corresponding to the intersection point 630 in the original conversion relationship curve 610 shown in FIG. The first part of the curve, and the part of the line segment whose grayscale value is higher than the grayscale value of the intersection point corresponding to the intersection point 630 in the preset conversion relationship curve 620 is selected as the second part of the adjusted conversion relationship curve. Taking FIG. 7 as an example, in step S430, the curve merging unit 822 selects a part of the line segment whose gray scale value is lower than the gray scale value of the intersection point corresponding to the intersection point 940 in the original conversion relationship curve 920 shown in FIG. The first part of the preset conversion relationship curve 910 is selected as the second part of the adjusted conversion relationship curve 950 .

To sum up, this embodiment adjusts the gain, and sets a preset conversion relationship curve (such as 620 shown in FIG. 6 or 910 shown in FIG. 7 ) as the maximum acceptable gain value of the system. The conversion relationship generator 820 can combine the preset conversion relationship curve with the original conversion relationship curve (such as 610 shown in FIG. 6 or 920 shown in FIG. 7 ) representing the original gain value G to obtain an adjusted conversion relationship curve. According to the maximum value of the three sub-pixel data of R, G, and B in the image data D ₁ , the conversion circuit 830 finds out the adjusted gain value G′ from the preset conversion relationship curve 950 . The adjusted gain value G' is applicable to the three sub-pixel data of R, G, and B in the image data _D1 . Taking FIG. 7 as an example, since the adjusted conversion relationship curve 950 converges to 1 in the high grayscale (high brightness) range of the image data _D1 , it can ensure that the three values of R, G, and B in the image data _D4 after gain are The sub-pixel data will not be "saturated", achieving the effect of preserving high-brightness details.

Although the present invention has been disclosed as above with the embodiments, it is not intended to limit the present invention. Anyone with ordinary knowledge in the technical field can make some changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention should be defined by the appended claims.

Claims (70)

1. An image processing method for performing image processing on image data of a plurality of frames of an image processing device, wherein each of the frames includes a plurality of pixel units, and each of the plurality of pixel units includes a plurality of sub-pixels corresponding to a plurality of sub-pixels respectively color, the image processing method includes: analyzing a picture among the multiple pictures to obtain an original gain value corresponding to multiple grayscale values of the picture, wherein the original gain value is higher than 1; Obtaining a preset data-to-gain conversion relationship look-up table, wherein the preset data-to-gain conversion look-up table records a plurality of preset gray scale values and a plurality of corresponding preset gain values for the image data of the plurality of screens to perform Commonly used in anti-saturation processing; According to the original gain value, use the preset data to look up the gain conversion relationship table to perform anti-saturation processing on the picture, and obtain the adjusted data-to-gain conversion relationship of the picture, wherein the adjusted data and the gain conversion relationship defining a conversion relationship between different grayscale values of the picture and corresponding adjusted gain values; and Gain processing is performed on the image data of the frame according to the adjusted data-to-gain conversion relationship. 2. The image processing method according to claim 1, wherein according to the adjusted data-to-gain conversion relationship, the step of performing gain processing on the image data of the picture comprises: Obtaining multiple grayscale values of the picture, and obtaining multiple adjusted gain values corresponding to the multiple grayscale values of the picture according to the adjusted data-to-gain conversion relationship; and Using the multiple adjusted gain values, gain processing is performed for the multiple gray scale values of the frame respectively. 3. The image processing method according to claim 2, wherein the plurality of grayscale values of the picture are obtained, and according to the adjusted data-to-gain conversion relationship, the corresponding multiplicity of grayscale values of the picture are respectively obtained. The steps for obtaining an adjusted gain value include: Obtaining a representative grayscale value of each pixel unit as one of the plurality of grayscale values according to a plurality of individual grayscale values of the plurality of sub-pixels of each pixel unit; and According to the adjusted data-to-gain conversion relationship, the gain value corresponding to the representative gray scale value of each pixel unit is obtained as one of the plurality of adjusted gain values. 4. The image processing method according to claim 3, wherein the step of obtaining the representative gray-scale value of each pixel unit according to a plurality of individual gray-scale values of each sub-pixel comprises: Find the highest gray-scale value from among the individual gray-scale values of the plurality of sub-pixels of each pixel unit as the representative gray-scale value of the pixel unit. 5. The image processing method according to claim 1, wherein according to the original gain value, the preset data is used to look up the gain conversion relationship table to perform the anti-saturation process on the picture to obtain the adjustment of the picture Post-data-to-gain conversion relationships include: setting the adjusted gain values corresponding to at least one higher gray-scale value in the plurality of gray-scale values of the frame to be equal to the preset gain values; and The adjusted gain values corresponding to at least one lower gray-scale value of the plurality of gray-scale values of the frame are set equal to the original gain value. 6. The image processing method according to claim 5, wherein the adjusted gain values corresponding to at least one higher gray-scale value in the gray-scale values of the picture are set equal to the preset The steps to gain value include: The preset data is directly used to convert the preset gray scale values in the lower gain range and the corresponding preset gain values in the gain conversion look-up table. 7. The image processing method according to claim 6, wherein the lower gain range at least includes at least one gain value, and the lower boundary gain value of the lower boundary range corresponds to the highest gray scale applied by the image processing device value. 8. The image processing method according to claim 7, wherein the lower boundary gain value is equal to 1. 9. The image processing method according to claim 6, further comprising: Obtaining a specific gain value according to at least one of the original gain value and the preset data-to-gain conversion relationship look-up table; and The lower gain range is determined according to the specific gain value. 10. The image processing method according to claim 9, wherein a higher boundary gain value of the lower gain range is lower than or equal to the specific gain value. 11. The image processing method according to claim 9, wherein according to at least one of the original gain value and the preset data-to-gain conversion relationship look-up table, obtaining the specific grayscale value comprises: Treat the original gain value as the specific gain value. 12. The image processing method according to claim 5, wherein the adjusted gain values corresponding to at least one lower gray-scale value among the gray-scale values of the picture are set equal to the original gain value include: The adjusted gain values corresponding to all the grayscale values in the higher gain range are set equal to the original gain value, so as to obtain the part of the higher gain range of the adjusted data-to-gain conversion relationship of the image. 13. The image processing method according to claim 12, wherein the higher gain range at least includes at least one gain value, and the upper boundary gain value of the higher gain range corresponds to the lowest gray scale applied by the image processing device value. 14. The image processing method according to claim 12, further comprising: According to the original gain value and the preset data pair gain conversion table lookup, to obtain a specific gain value; and The higher gain range is determined according to the specific gain value. 15. The image processing method according to claim 14, wherein a lower boundary gain value of the higher gain range is higher than or equal to the specific gain value. 16. The image processing method according to claim 14, wherein according to at least one of the original gain value and the preset data-to-gain conversion relationship look-up table, obtaining the specific grayscale value comprises: Treat the original gain value as the specific gain value. 17. The image processing method according to claim 1, further comprising: A rounding process is performed so that the adjusted gain values corresponding to the multiple gray-scale values adjacent to the original gray-scale value have a smooth distribution. 18. The image processing method according to claim 17, wherein the rounding process comprises low-pass filtering process. 19. The image processing method according to claim 1, wherein in the high grayscale range covering at least the highest grayscale value applied by the image processing device, at least one correspondence between the preset data and the gain conversion relationship look-up table The default gain value is 1. 20. The image processing method according to claim 1, wherein in the low grayscale range covering at least the lowest grayscale value applied by the image processing device, the preset data has at least one value on the gain conversion relationship look-up table. A corresponding preset gain value is higher than the original gain value. 21. A method of desaturating image data for image processing by an image processing apparatus, comprising: Analyzing one of the multiple pictures to obtain an original conversion relationship curve of the picture, the original conversion relationship curve indicates that multiple gray scale values of the picture are to be processed according to an original gain value, wherein the original gain value is greater than 1; Obtaining a preset conversion relationship curve, wherein the preset conversion relationship curve is used jointly when performing desaturation processing on the image data of the plurality of frames; Desaturation processing is performed on the image by merging the preset conversion relationship curve with the original conversion relationship curve, so as to obtain an adjusted conversion relationship curve; and Gain processing is performed on the image data of the frame according to the adjusted conversion relationship curve. 22. The method for desaturating image data according to claim 21, wherein each of the original conversion relationship curve, the preset conversion relationship curve and the adjusted conversion relationship curve is a data versus gain curve. 23. The method for desaturating image data according to claim 22, wherein in the high grayscale range covering at least the highest grayscale value applied by the image processing device, at least one corresponding on the preset conversion relationship curve The gain values for are all greater than 0 and less than or equal to 1. 24. The method for desaturating image data according to claim 22, wherein in the low grayscale range covering at least the lowest grayscale value applied by the image processing device, at least one corresponding on the preset conversion relationship curve Gain values are higher than the original gain value. 25. The method for desaturating image data as claimed in claim 22, wherein the original conversion relationship curve is a horizontal straight line passing through the original gain value. 26. The method for desaturating image data according to claim 22, wherein image data of different colors of the frame use different preset conversion relationship curves. 27. The method for desaturating image data according to claim 21, wherein each of the original conversion relationship curve, the preset conversion relationship curve and the adjusted conversion relationship curve is a data-to-data curve, respectively defined The corresponding relationship between the gray scale value before gain conversion and the gray scale value after gain conversion. 28. The method for desaturating image data according to claim 27, wherein in the high grayscale range covering at least the highest grayscale value applied by the image processing device, the slope of the preset conversion relationship curve is greater than 0 and less than or equal to 1. 29. The method of desaturating image data according to claim 27, wherein The original conversion relationship curve has an intersection point with the default conversion relationship curve, In the low gray-scale value range where the gray-scale value before conversion is less than the gray-scale value of the intersection point corresponding to the intersection point, the converted gray-scale values on the original conversion relationship curve are all less than or equal to the converted gray-scale values on the preset conversion relationship curve. gain value, and In the high gray-scale value range where the gray-scale value before conversion is greater than the gray-scale value of the intersection point, the converted gray-scale values on the original conversion relationship curve are all greater than or equal to the converted gain value on the preset conversion relationship curve. 30. The method for desaturating image data according to claim 27, wherein the original conversion relationship curve comprises: an oblique straight line with a slope equal to the original gain value and a horizontal straight line with a slope equal to zero. 31. The method for desaturating image data according to claim 27, wherein image data of different colors of the frame use the same preset conversion relationship curve. 32. The method for desaturating image data according to claim 21, wherein the adjusted conversion relationship is obtained by combining the preset conversion relationship curve and the individual conversion relationship curve to perform the desaturation process on the picture curves, including: Selecting a part of the original conversion relationship curve whose grayscale value is lower than the grayscale value of the intersection point corresponding to the intersection point of the preset conversion relationship curve and the original conversion relationship curve as the first part of the adjusted conversion relationship curve; and A part of the line segment in the preset conversion relationship curve whose gray scale value is higher than the gray level value of the intersection point is selected as the second part of the adjusted conversion relationship curve. 33. The method of desaturating image data of claim 21, further comprising: The adjusted conversion relationship curve is smoothed so that a part of the adjusted conversion relationship curve covering the original conversion relationship curve and the preset conversion relationship curve has a smooth connection. 34. The method of desaturating image data according to claim 33, wherein the smoothing process comprises a low pass filtering process. 35. An image processing device for performing image processing on image data of a plurality of frames, wherein each of the frames includes a plurality of pixel units, and each of the plurality of pixel units includes a plurality of sub-pixels, respectively corresponding to a plurality of colors, the Image processing devices include: a global gain generator, analyzing a picture among the multiple pictures to obtain an original gain value corresponding to multiple gray scale values of the picture, wherein the original gain value is higher than 1; A conversion relationship generator, coupled to the global gain generator to receive the original gain value, and obtain a preset data-to-gain conversion relationship look-up table, wherein the preset data-to-gain conversion look-up table records a plurality of preset gray scale values A plurality of preset gain values corresponding to each are used together for desaturation processing of the image data of the plurality of frames, and the conversion relationship generator utilizes the preset data to modify the gain conversion relationship according to the original gain value Look up the table to perform anti-saturation processing on the picture to obtain the adjusted data-to-gain conversion relationship of the picture, wherein the adjusted data-to-gain conversion relationship defines the different grayscale values of the picture and the corresponding adjusted gain values conversion relationship between; and The conversion circuit is coupled to the conversion relationship generator to perform gain processing on the image data of the frame according to the adjusted data-to-gain conversion relationship. 36. The image processing device according to claim 35, wherein the conversion circuit comprises: The gain determining circuit is coupled to the conversion relationship generator, wherein the gain determining circuit obtains multiple gray scale values of the picture, and obtains the multiple gray scale values of the picture according to the adjusted data-to-gain conversion relationship. a corresponding plurality of adjusted gain values; and The calculator is coupled to the gain determining circuit, wherein the calculator uses the adjusted gain values to respectively perform gain processing for the plurality of gray scale values of the frame. 37. The image processing device according to claim 35, wherein the conversion circuit comprises: A comparison circuit, according to a plurality of individual gray-scale values of the plurality of sub-pixels of each pixel unit, to obtain a representative gray-scale value of each pixel unit as one of the plurality of gray-scale values; a gain determining circuit, coupled to the conversion relationship generator and the comparison circuit, wherein the gain determining circuit obtains the representative gray scale value, and obtains the representative gray scale values of the pixel units according to the adjusted data-to-gain conversion relationship a plurality of adjusted gain values corresponding to the respective values; and The calculator is coupled to the gain determining circuit, wherein the calculator uses the adjusted gain values to respectively perform gain processing for the plurality of gray scale values of the frame. 38. The image processing device according to claim 37, wherein the comparison circuit finds the highest grayscale value from among the plurality of individual grayscale values of the plurality of sub-pixels of each pixel unit as the pixel unit Represents the grayscale value. 39. The image processing device according to claim 35, wherein the conversion relationship generator sets the adjusted gain values corresponding to some higher gray-scale values among the gray-scale values of the picture to be equal to the a preset gain value; and the conversion relationship generator sets the adjusted gain values corresponding to some of the lower gray-scale values of the image to be equal to the original gain value. 40. The image processing device according to claim 39, wherein the conversion relationship generator directly uses the preset data to pair the preset gray scale values in the gain conversion look-up table that belong to the lower gain range with the corresponding some preset gain values. 41. The image processing device according to claim 40, wherein the lower gain range at least includes at least one gain value, and the lower boundary gain value of the lower boundary range corresponds to the highest gray scale applied by the image processing device value. 42. The image processing apparatus according to claim 41, wherein the lower boundary gain value is equal to 1. 43. The image processing device according to claim 40, wherein the conversion relationship generator obtains a specific gain value according to at least one of the original gain value and the preset data-to-gain conversion relationship look-up table; and the conversion relationship The generator determines the lower gain range according to the specific gain value. 44. The image processing device according to claim 43, wherein a higher boundary gain value of the lower gain range is lower than or equal to the specific gain value. 45. The image processing device according to claim 43, wherein the conversion relation generator regards the original gain value as the specific gain value. 46. The image processing device according to claim 39, wherein the conversion relationship generator sets the adjustment gain values corresponding to all gray scale values in the higher gain range to be equal to the original gain value to obtain the The portion of the higher gain range of the adjusted data-to-gain conversion relationship. 47. The image processing device according to claim 46, wherein the higher gain range at least includes at least one gain value, and the higher boundary gain value of the higher gain range corresponds to the lowest gray scale applied by the image processing device value. 48. The image processing device according to claim 46, wherein the conversion relationship generator obtains a specific gain value according to the original gain value and the preset data to a gain conversion relationship look-up table; and the conversion relationship generator obtains a specific gain value according to the specific The gain value determines this upper gain range. 49. The image processing device according to claim 48, wherein a lower boundary gain value of the higher gain range is higher than or equal to the specific gain value. 50. The image processing device according to claim 48, wherein the conversion relationship generator regards the original gain value as the specific gain value. 51. The image processing device according to claim 35, wherein the conversion relationship generator performs a smoothing process, so that the adjusted gain values corresponding to the gray scale values adjacent to the original gray scale value have a smooth distribution. 52. The image processing apparatus according to claim 51, wherein the rounding process comprises a low-pass filter process. 53. The image processing device according to claim 35, wherein in the high gray scale range covering at least the highest gray scale value applied by the image processing device, at least one correspondence between the preset data and the gain conversion relationship look-up table The default gain value is 1. 54. The image processing device according to claim 35, wherein in the low grayscale range covering at least the lowest grayscale value applied by the image processing device, at least one correspondence between the preset data and the gain conversion relationship look-up table preset gain values are all higher than the original gain value. 55. An image processing device for performing image processing, the image processing device comprising: The global gain generator analyzes a picture among multiple pictures to obtain the original conversion relationship curve of the picture, and the original conversion relationship curve indicates that multiple grayscale values of the picture are to be processed according to the gain value of the original gain value, wherein the original gain value greater than 1; A conversion relationship generator, coupled to the global gain generator to receive the original gain value, and obtain a preset conversion relationship curve, wherein the preset conversion relationship curve is used for desaturation processing of the image data of the plurality of frames. using, and the conversion relationship generator merges the preset conversion relationship curve and the original conversion relationship curve to perform desaturation processing on the picture, so as to obtain an adjusted conversion relationship curve; and The conversion circuit is coupled to the conversion relationship generator to perform gain processing on the image data of the frame according to the adjusted conversion relationship curve. 56. The image processing device according to claim 55, wherein each of the original conversion relationship curve, the preset conversion relationship curve, and the adjusted conversion relationship curve is a data versus gain curve. 57. The image processing device according to claim 56, wherein in the high gray scale range covering at least the highest gray scale value applied by the image processing device, at least one corresponding gain value on the preset conversion curve is Greater than 0 and less than or equal to 1. 58. The image processing device according to claim 56, wherein in the low grayscale range covering at least the lowest grayscale value applied by the image processing device, at least one corresponding gain value on the preset conversion curve is higher than the original gain value. 59. The image processing device according to claim 56, wherein the original conversion relationship curve is a horizontal straight line passing through the original gain value. 60. The image processing device according to claim 56, wherein image data of different colors of the frame use different preset conversion relationship curves. 61. The image processing device according to claim 55, wherein each of the original conversion relationship curve, the preset conversion relationship curve, and the adjusted conversion relationship curve is a data-to-data curve, respectively defining gray values before gain conversion The corresponding relationship between the scale value and the gray scale value after gain conversion. 62. The image processing device according to claim 61, wherein in the high grayscale range covering at least the highest grayscale value applied by the image processing device, the slope of the preset conversion relationship curve is greater than 0 and less than or equal to 1 . 63. The image processing apparatus according to claim 61, wherein The original conversion relationship curve has an intersection point with the preset conversion relationship curve, In the low gray-scale value range where the gray-scale value before conversion is less than the gray-scale value of the intersection point corresponding to the intersection point, the converted gray-scale values on the original conversion relationship curve are all less than or equal to the converted gray-scale values on the preset conversion relationship curve. gain value, and In the high gray-scale value range where the gray-scale value before conversion is greater than the gray-scale value of the intersection point, the converted gray-scale values on the original conversion relationship curve are all greater than or equal to the converted gain value on the preset conversion relationship curve. 64. The image processing device according to claim 61, wherein the original conversion relationship curve system comprises: an oblique straight line with a slope equal to the original gain value and a horizontal straight line with a slope equal to zero. 65. The image processing device according to claim 61, wherein image data of different colors in the frame use the same preset conversion relationship curve. 66. The image processing device according to claim 55, wherein the transformation relationship generator comprises: a lookup table providing the preset conversion curve; and a curve merging unit, coupled to the look-up table and the global gain generator, wherein the curve merging unit selects the grayscale value in the original conversion relationship curve lower than the intersection point corresponding to the preset conversion relationship curve and the original conversion relationship curve Part of the line segment of the intersection gray level value is used as the first part of the adjusted conversion relationship curve; and the curve merging unit selects a part of the line segment whose gray level value is higher than the intersection point gray level value in the preset conversion relationship curve as the adjusted conversion The second part of the relationship curve. 67. The image processing device according to claim 55, wherein the conversion relationship generator performs smoothing processing on the adjusted conversion relationship curve, so that the adjusted conversion relationship curve covers the original conversion relationship curve and the preset conversion relationship Some segments of the curve have smooth connections. 68. The image processing apparatus according to claim 67, wherein the rounding process comprises a low-pass filtering process. 69. The image processing device according to claim 55, wherein the conversion circuit comprises: A comparison circuit, according to a plurality of individual gray-scale values of the plurality of sub-pixels of each pixel unit, to obtain a representative gray-scale value of each pixel unit as one of the plurality of gray-scale values; a gain determining circuit, coupled to the conversion relationship generator and the comparison circuit, wherein the gain determining circuit obtains the representative gray scale value, and obtains the representative gray scale values of the pixel units respectively according to the adjusted conversion relationship curve a corresponding plurality of adjusted gain values; and The calculator is coupled to the gain determining circuit, wherein the calculator uses the adjusted gain values to respectively perform gain processing for the plurality of gray scale values of the frame. 70. The image processing device according to claim 69, wherein the comparison circuit finds the highest grayscale value among the plurality of individual grayscale values of the plurality of sub-pixels of each pixel unit as the pixel unit Represents the grayscale value.