CN-122027828-A - PQ coding-based SDR-to-HDR image processing method and device

Abstract

The invention provides an SDR-to-HDR image processing method and device based on PQ coding. The method adopts a plurality of processing steps, firstly, the self-adaptive brightness dynamic range adjustment is carried out on the SDR image, the mapping curve is dynamically adjusted according to the image scene, the details of the dark area are enhanced, and the details of the highlight area are kept. Secondly, after the color gamut of the original image is expanded from BT.709 to BT.2020 through color gamut conversion, the color coverage of the image is further expanded through a color coordinate gain expansion algorithm to adapt to the HDR display to support colors. Finally, the first linear value is mapped to the luminance range of the HDR display using the PQ encoding standard. The whole flow of the method can be modularized and designed into a high-efficiency pipeline structure to be deployed on an FPGA platform, so that the brightness, color and detail performance of the image can be obviously enhanced, the method is suitable for the fields of video playing, image displaying, real-time image processing and the like, and has strong practicability and wide application prospect.

Inventors

• ZHANG JIAN
• LIU JIYUAN
• ZHANG YUNING
• NIU BAOLI
• WU MEI

Assignees

• 新型显示与视觉感知石城实验室

Dates

Publication Date

20260512

Application Date

20260127

Claims (10)

1. 1. An SDR to HDR image processing method based on PQ coding, comprising the steps of: (1) Performing sRGB degamma decoding on an input SDR image, and converting the image from a nonlinear color space into a linear light value; (2) Performing brightness optimization on the linear light value obtained in the step (1) through a self-adaptive brightness adjustment algorithm; (3) Performing color gamut conversion and color gamut expansion processing on the linear light value optimized in the step (2); (4) And (3) performing PQ coding on the image processed in the step (3), and adapting the PQ coding to the brightness range of the HDR display device.
2. 2. The method of claim 1, wherein the adaptive brightness adjustment algorithm of step (2) comprises smoothing the image brightness using an exponentially weighted average algorithm and dynamically adjusting based on the contrast map and the brightness adjustment factor.
3. 3. The method of claim 2, wherein the dynamically adjusting comprises setting a reference luminance to control an overall luminance distribution and enhancing local detail of the image by adjusting contrast mapping parameters.
4. 4. The method of claim 1, wherein the gamut conversion and gamut expansion process in step (3) comprises: (3.1) mapping linear RGB values from bt.709 gamut to bt.2020 gamut using a gamut conversion matrix; (3.2) determining an extended gain by calculating an intersection point of the ray and the color gamut boundary based on CIE 1931 color coordinates, and dynamically adjusting color information of each pixel by applying the gain.
5. 5. The method of claim 4, wherein the gamut conversion matrix used in step (3.1) is:
6. 6. The method of claim 4, wherein in step (3.2) the extended gain is a pixel-by-pixel gain factor determined by calculating the intersection of rays with bt.2020 gamut boundaries and bt.709 gamut boundaries.
7. 7. The method of claim 1 wherein the PQ encoding in step (4) uses SMPTE ST 2084 standard to non-linearly compress the image brightness.
8. 8. The method of any one of claims 1 to 7, further comprising the step (5) of generating and encapsulating dynamic metadata containing luminance and color gamut information for optimizing display effects on an HDR display device.
9. 9. An image processing apparatus, comprising: The decoding module is used for performing sRGB degamma decoding on the input SDR image so as to output a linear light value; the brightness adjustment module is connected to the decoding module and is used for carrying out brightness smoothing on the linear light value by adopting an exponential weighted average algorithm and carrying out dynamic adjustment based on contrast mapping and brightness adjustment factors; A gamut processing module, coupled to the brightness adjustment module, for mapping and expanding the image from bt.709 gamut to bt.2020 gamut; And the coding module is connected with the color gamut processing module and is used for carrying out nonlinear brightness compression on the image by adopting PQ coding based on SMPTE ST 2084 standard.
10. 10. The apparatus of claim 9, wherein the modules of the apparatus are deployed in a pipeline configuration on an FPGA platform.

Description

PQ coding-based SDR-to-HDR image processing method and device Technical Field The invention relates to the technical field of image processing, in particular to a SDR (Standard Dynamic Range) -HDR (High Dynamic Range) image processing method based on PQ (Perceptual Quantizer) codes and a device thereof. Background With the development of high resolution display technology, users' demands for high quality media content are increasing, making consumer devices supporting high resolution and high dynamic range extremely competitive in the market. The HDR technology can provide higher visual effect and more real image experience by expanding the dynamic range and the color gamut of brightness, and compared with an SDR image, the HDR technology has the advantages of being capable of presenting richer brightness and wider colors, so that the viewing experience is more vivid. However, the production and transmission of current video content is still largely based on the SDR standard, resulting in that most of the existing video content does not have native HDR characteristics. Currently, there are three main methods for obtaining HDR content: and (one), combining brightness and color information of the images under a plurality of different exposure conditions to generate a high dynamic range image. (II) capturing an HDR image using a high dynamic range sensor, generating native HDR video by a camera and display device supporting the HDR standard. (III) for most SDR image video sources, the SDR source is converted to an HDR source by Tone Mapping (Tone Mapping) techniques. For the three technical routes, the first method has a certain defect in real-time performance through multi-frame synthesis, and the second method has the advantage that the cost for obtaining the HDR display effect by a user is increased by using an expensive HDR camera and display equipment. The third technical route has stronger flexibility, and for the existing SDR material, the HDR effect can be generated through professional post-production and color correction software, the HDR version of classical film and television works can be produced, and the HDR content can be adjusted in detail during post-production so as to ensure that the desired display effect is met. The prior art can achieve SDR to HDR conversion, but many challenges remain. For example, existing tone mapping techniques often have difficulty in achieving both luminance detail and color performance, or excessive color distortion during the mapping process. In addition, existing conversion methods generally require analysis of metadata of video content from frame to frame or scene to scene, increasing computational complexity and increasing difficulty of application in real-time video processing. The invention aims to provide a more efficient, accurate and highly-adaptive SDR-to-HDR conversion technology, which can improve the performance of an HDR display device and bring higher-quality image display effect to users. Disclosure of Invention Object of the Invention The invention aims to provide a high-efficiency SDR (digital video recorder) to HDR (high-definition digital video recorder) conversion method and device, and aims to solve the problem that the potential of the existing SDR content cannot be fully displayed on an HDR display device. Standard Dynamic Range (SDR) video signals are converted to a High Dynamic Range (HDR) format by utilizing a tone mapping algorithm to present richer brightness, color, and contrast on HDR-enabled devices. The technology can be compatible with the existing SDR content, and can display a clearer visual effect on the HDR equipment through effective algorithm optimization, so that the viewing experience of a user is improved. Technical proposal In order to achieve the above object, the present invention provides an SDR to HDR image processing method based on PQ encoding, the core flow of which includes the following steps, which are sequentially executed: 1. and (3) linearization decoding, namely performing sRGB degamma decoding on the input SDR image, converting the image from a nonlinear color space into a linear light value, and laying a foundation for the follow-up accurate processing. The adopted piecewise transfer function is as follows: Wherein, the To input normalized channel values in the sRGB color space,Is the converted linear light value. 2. Adaptive luminance adjustment-luminance optimization of linear light value images to fit the wider dynamic range of HDR displays. This step includes two key sub-processes: And (3) brightness smoothing, namely adopting an exponential weighted average (EMA) algorithm to carry out smoothing treatment on the brightness of the image, and avoiding severe brightness fluctuation. The recursive implementation formula is as follows: dynamic mapping, namely dynamically adjusting the final brightness mapping relation by combining a basic mapping curve, a middle gray pushing item and a highlight inhibiti