Naked eye 3D image quality optimization system

Abstract

The invention relates to the technical field of image data processing, solves the problems that traditional 3D display is low in image quality, needs to depend on peripherals such as glasses / helmets, is troublesome to use and operate, and lacks targeted image optimization for the application field, and discloses a naked eye 3D image quality optimization system which comprises a 3D display module serving as a system hardware carrier; the hardware sensing layer is modularly integrated on the 3D display module and comprises an environment sensing array used for collecting environment illumination, color temperature and screen surface reflectivity parameters in real time; the iris tracking unit outputs the interpupillary distance, the three-dimensional space coordinates and the visual angle vector of the viewer in real time; and the data processing layer comprises a multi-parameter processing engine, is connected with the environment sensing array and the iris tracking unit, and comprises a multi-mode color gamut conversion module, a self-adaptive gamma correction module and a dynamic backlight control module. The method has the advantages that the image quality is comprehensively improved through a hardware-software collaborative architecture, the pure naked eye 3D display image is optimized, and the cross-domain adaptability of pure naked eye 3D display is improved.

Description

Technical Field

[0001] The present invention relates to the field of image data processing technology, and specifically to a naked-eye 3D image quality optimization system. Background Art

[0002] Naked-eye 3D display technology is in a rapid development stage and has shown great potential in many fields such as aerospace, scientific research and teaching, and medical imaging. Naked-eye 3D display technology uses optical structures such as gratings, lens arrays, or parallax barriers to enable observers to perceive stereoscopic images without wearing special glasses. However, this technology still faces many challenges in practical applications.

[0003] 1. Traditional 3D displays only address basic imaging issues, ignoring core image quality indicators such as ambient light interference, viewing angle offset, and color distortion.

[0004] 2. Reliance on peripherals such as glasses / helmets makes operation cumbersome: frequent wearing and adjustment reduces efficiency; insufficient precision: optical lenses introduce distortion, resulting in grayscale recognition errors (ΔL>5%); poor comfort: long-term wearing causes dizziness and eye fatigue;

[0005] 3. When the observation angle is offset, the image crosstalk is serious, affecting the three-dimensional effect, and lacks precise color and grayscale optimization for fields such as medicine and aerospace. Summary of the Invention

[0006] The present invention provides a naked-eye 3D image quality optimization system to solve the problems of low image quality in traditional 3D displays, reliance on peripherals such as glasses / helmets, inconvenient operation, and lack of targeted image optimization for application fields.

[0007] To solve the above technical problems, the present invention provides a technical solution: a naked-eye 3D image quality optimization system, comprising: a 3D display module as a system hardware carrier;

[0008] The hardware perception layer is modularly integrated into the 3D display module and includes:

[0009] The environmental sensing array consists of distributed multi-spectral ambient light sensors, a V(λ)-corrected photodetector array, and a spectrum analyzer, which is used to collect ambient illumination, color temperature, and screen surface reflectivity parameters in real time;

[0010] The iris tracking unit integrates a near-infrared imaging module and a 6-DOF pose calculation module to output the viewer's pupil distance, 3D spatial coordinates, and viewing angle vector in real time.

[0011] The data processing layer is integrated into the 3D display module through an embedded architecture and includes:

[0012] A multi-parameter processing engine connects the environmental sensing array and the iris tracking unit and integrates:

[0013] a) Multimodal color gamut conversion module:

[0014] Supports dynamic switching of sRGB / Native / SMPTE-C / DCI-P3 / REC.709 / EBU / User / AdobeRGB color gamuts;

[0015] b) Adaptive gamma correction module, including 512 grayscale adjustment, ambient light responsive gamma curve generator, and personalized learning model;

[0016] Dynamic backlight control module with CNN-based partitioned brightness prediction algorithm, sub-pixel backlight compensation unit and brightness equalization feedback loop;

[0017] The collaborative optimization layer is integrated into the 3D display module in the form of a pluggable module and includes:

[0018] Mode selector, with four preset basic modes: aerospace application, R&D, teaching, and medical imaging;

[0019] Parameter coupling optimizer, which establishes the geometric mapping relationship between the viewer's viewpoint and the screen area, and synchronously adjusts the color gamut, gamma, and backlight parameters;

[0020] DICOM GSDF verification unit, configured to: build luminance calibration curve, support 400cd / m 2 、300cd / m 2 , 200cd / m 2 Three levels of medical display brightness modes; in medical imaging mode, the JND index mapping algorithm is automatically activated to ensure that grayscale identification meets the DICOM consistency requirement of ΔL≤2%; and a verified brightness-contrast report is output.

[0021] Among them, the hardware perception layer, data processing layer, and collaborative optimization layer realize physical-logical double-layer coupling with the 3D display module through a standard communication interface, forming a modular and scalable architecture.

[0022] The multimodal color gamut conversion module performs:

[0023] Resolve the color space identifier of the input signal through the spectral dimension matching algorithm;

[0024] Activate DCI-P3 / REC.709 dual-channel processing for film and television content;

[0025] Prioritize matching AdobeRGB / EBU standards for printed design content;

[0026] When loading user-defined 3D LUT parameters, maintain ΔE2000 ≤ 0.3% color loss.

[0027] The adaptive gamma correction module comprises:

[0028] Ambient light responsive gamma curve generator, dynamically adjusting correction parameters based on 0-100klux illumination;

[0029] The machine learning processor builds a personalized gamma correction model based on historical data and generates a gamma profile associated with the device ID.

[0030] The dynamic backlight control module:

[0031] Adopting sub-pixel backlight compensation algorithm, the crosstalk rate of stereo display is less than 1%;

[0032] Through the brightness balance feedback loop, the brightness difference between each partition is maintained at ≤5nits.

[0033] The environmental sensing array:

[0034] A visible light and near-infrared dual-channel sensor is used to collect ambient spectral data; the ambient light compensation parameters are output through the CIECAM02 color appearance model.

[0035] The parameter coupling optimizer:

[0036] Establish a spatial coordinate transformation matrix based on the viewpoint position; automatically activate the corresponding processing mode according to the input signal metadata; and perform closed-loop collaborative optimization of color gamut, gamma, and backlight parameters.

[0037] The DICOM GSDF Validation Unit:

[0038] At 400cd / m 2 Enable high dynamic range backlight compensation in this mode to maintain black level brightness ≤ 0.05cd / m 2 ;

[0039] At 200cd / m 2 In this mode, the low-light color fidelity algorithm is activated to ensure that the sRGB color gamut coverage is ≥ 99%.

[0040] The advantages of the present invention are: 1. Hardware-software collaborative architecture: integrating environmental perception array (multispectral sensor + iris tracking) and multimodal processing engine to achieve closed-loop optimization from signal input to display output;

[0041] 2. Improved cross-domain adaptability: By pre-setting modes such as aerospace, medicine, and teaching, it automatically matches professional standards such as DICOM GSDF, sRGB / AdobeRGB to meet the needs of different scenarios;

[0042] 3. Pure naked-eye 3D display image optimization: Through sub-pixel backlight compensation (crosstalk rate <0.8%) and viewpoint tracking (positioning accuracy ±0.3mm), an immersive experience without peripherals is achieved;

[0043] 4. Comprehensive improvement in image quality: color richness: supports dynamic switching of 8 color gamuts; three-dimensional layering: 512-level grayscale adjustment plus backlight control; fine details: CNN partition prediction algorithm realizes independent light control in multiple zones with small brightness differences. DETAILED DESCRIPTION

[0044] The present invention is described in further detail below.

[0045] A glasses-free 3D image quality optimization system uses a 3D display module as its core hardware carrier and implements physical-logical dual-layer coupling of the following modules through standardized interfaces (such as HDMI 2.1 and USB-C):

[0046] Hardware perception layer: embedded in the display module frame, including the environmental perception array and iris tracking unit;

[0047] Data processing layer: Implement multi-parameter processing engine and dynamic backlight control module through FPGA and embedded processor;

[0048] Collaborative optimization layer: Integrates the mode selector, parameter coupling optimizer, and DICOM GSDF verification unit with a pluggable hardware module (PCIe interface).

[0049] A layered communication protocol is used between each layer: the physical layer transmits raw data through a high-speed serial bus, and the logical layer implements parameter interaction through a JSON format instruction set, with a response delay of less than 10ms.

[0050] Environmental sensing arrays: including multispectral sensing networks:

[0051] Five sets of multi-spectral ambient light sensors are arranged on the display module, covering the 380-1000nm spectrum range and the sampling rate of 200Hz; an integrated V(λ) corrected photodetector is used to 2 The C-port outputs CIE 1931XYZ color space data; the spectrum analyzer is coupled to the screen surface via optical fiber to monitor the reflectivity in real time (wavelength step size 5nm).

[0052] The data processing method of the multispectral sensor network is as follows: the ambient light compensation parameters are calculated using the CIECAM02 model, and the outputs are: dynamic ambient illumination (0-100klux, ±1% accuracy); color temperature compensation matrix (2000-10000K, ΔCCT ≤ 30K); reflectance spectrum curve (450-650nm, resolution 5nm).

[0053] Iris tracking unit: Near-infrared imaging system: uses a 940nm VCSEL array and global shutter CMOS, with a resolution of 1280×960 and a frame rate of 240fps; binocular baseline distance is 72mm, depth of field is 0.5-5m, and positioning accuracy is ±0.3mm (RMS); 6-DOF pose solution: integrates the PnP algorithm and IMU data (Bosch BMI088), output parameters: interpupillary distance (IPD, 55-75mm, ±0.05mm); 3D coordinates (XYZ axes, ±1mm accuracy); viewing angle vector (pitch angle θ, yaw angle φ, ±0.3° accuracy).

[0054] Multimodal Color Gamut Conversion Module: Dynamic switching process: Parses color space identifiers (such as ICC Profile and HDR10+ metadata) in the input signal metadata; Enables dual-channel processing for DCI-P3 video content: Primary channel: REC.709 gamma curve mapping (γ=2.4); Secondary channel: Histogram equalization based on the DCI-P3 XYZ space (256-level segmentation); When loading a user-defined 3D LUT, a trilinear interpolation algorithm is used to ensure a ΔE2000 color difference of ≤0.3%.

[0055] Adaptive gamma correction module: Ambient light responsive adjustment:

[0056] Dynamically adjust the gamma value according to the real-time illuminance E (unit: klux):

[0057] γ(E)=2.2+0.015×ln(E+1)

[0058] When the illumination range is 0-100klux, γ∈[2.2,2.8].

[0059] Machine learning model: Uses LSTM network to build a personalized gamma curve. Input parameters include:

[0060] Historical usage scenario labels (medicine / research / teaching / aerospace);

[0061] The user's pupil adjustment speed (based on iris tracking data);

[0062] Outputs the gamma configuration file bound to the device ID (stored in encrypted EEPROM).

[0063] Dynamic backlight control module: CNN partition prediction: Input the YUV 4:2:2 data of the current frame and iris tracking coordinates, and use a lightweight CNN (MobileNetV3-Small) to predict the brightness of 1024 partitions, with an inference time of less than 3ms.

[0064] Sub-pixel compensation algorithm:

[0065] Calculate the crosstalk compensation coefficient according to the viewpoint position:

[0066]

[0067] Wherein, k=0.12 is the attenuation coefficient, d is the sub-pixel spacing, and d0=0.25 mm is the reference distance, ensuring that the crosstalk rate is less than 0.8%.

[0068] Mode selection and parameter coupling:

[0069] Medical imaging mode: DICOM GSDF verification unit loads the preset JND index table (12-bit grayscale); at 400cd / m 2 In this mode, the black level compensation algorithm is enabled and the black level brightness is ≤0.05cd / m 2 (Actually measured 0.03cd / m 2 ); Generate a PDF format verification report containing: brightness-contrast curve (DICOM GSDF compliance); ΔL deviation heat map (ΔL ≤ 1.5% at a color temperature of 6500K).

[0070] Space application mode: forced sRGB color gamut (coverage ≥ 99.5%), gamma lock 2.4, blue light band (450-480nm) attenuation rate ≥ 35%.

[0071] Can realize aerospace mode: anti-ambient light interference (gamma value adaptive under 100klux illumination);

[0072] Medical imaging mode: JND index mapping ensures continuous grayscale discrimination, improving diagnostic recognition rate by 30%;

[0073] Educational and teaching mode: Multi-viewing angles are optimized synchronously, and there is no color drift within ±30° viewing angle offset.

[0074] The above description of the present invention and its embodiments is non-limiting, and the actual structure is not limited thereto. In short, if a person skilled in the art is inspired by the above description and, without departing from the purpose of the present invention, designs a structure and embodiment similar to the technical solution without creatively designing, they shall fall within the scope of protection of the present invention.

Claims

1. A naked-eye 3D image quality optimization system, characterized by: include: 3D display module, as the system hardware carrier; The hardware perception layer is modularly integrated into the 3D display module and includes: The environmental sensing array consists of distributed multi-spectral ambient light sensors, a V(λ)-corrected photodetector array, and a spectrum analyzer, which is used to collect ambient illumination, color temperature, and screen surface reflectivity parameters in real time; The iris tracking unit integrates a near-infrared imaging module and a 6-DOF pose calculation module to output the viewer's pupil distance, 3D spatial coordinates, and viewing angle vector in real time. The data processing layer is integrated into the 3D display module through an embedded architecture and includes: A multi-parameter processing engine connects the environmental sensing array and the iris tracking unit and integrates: a) Multimodal color gamut conversion module: supports dynamic switching of sRGB / Native / SMPTE-C / DCI-P3 / REC.709 / EBU / User / AdobeRGB color gamuts; b) Adaptive gamma correction module, including 512 grayscale adjustment, ambient light responsive gamma curve generator, and personalized learning model; c) Dynamic backlight control module, which features a CNN-based partitioned brightness prediction algorithm, a sub-pixel backlight compensation unit, and a brightness equalization feedback loop; The collaborative optimization layer is integrated into the 3D display module in the form of a pluggable module and includes: Mode selector, with four preset basic modes: aerospace application, R&D, teaching, and medical imaging; Parameter coupling optimizer, which establishes the geometric mapping relationship between the viewer's viewpoint and the screen area, and synchronously adjusts the color gamut, gamma, and backlight parameters; DICOM GSDF verification unit, configured to: build luminance calibration curve, support 400cd / m 2 、300cd / m 2 , 200cd / m 2 Three levels of medical display brightness modes; in medical imaging mode, the JND index mapping algorithm is automatically activated to ensure that grayscale identification meets the DICOM consistency requirement of ΔL≤2%; and a verified brightness-contrast report is output. Among them, the hardware perception layer, data processing layer, and collaborative optimization layer realize physical-logical double-layer coupling with the 3D display module through a standard communication interface, forming a modular and scalable architecture.

2. The naked-eye 3D image quality optimization system according to claim 1, characterized in that: The multimodal color gamut conversion module performs: Resolve the color space identifier of the input signal through the spectral dimension matching algorithm; Activate DCI-P3 / REC.709 dual-channel processing for film and television content; Prioritize matching AdobeRGB / EBU standards for printed design content; When loading user-defined 3D LUT parameters, maintain ΔE2000 ≤ 0.3% color loss.

3. The naked-eye 3D image quality optimization system according to claim 1, characterized in that: The adaptive gamma correction module comprises: Ambient light responsive gamma curve generator, dynamically adjusting correction parameters based on 0-100klux illumination; The machine learning processor builds a personalized gamma correction model based on historical data and generates a gamma profile associated with the device ID.

4. The naked-eye 3D image quality optimization system according to claim 1, characterized in that: The dynamic backlight control module: Adopting sub-pixel backlight compensation algorithm, the crosstalk rate of stereo display is less than 1%; Through the brightness balance feedback loop, the brightness difference between each partition is maintained at ≤5nits.

5. The naked-eye 3D image quality optimization system according to claim 1, characterized in that: The environmental sensing array: A visible light and near-infrared dual-channel sensor is used to collect ambient spectral data; the ambient light compensation parameters are output through the CIECAM02 color appearance model.

6. The naked-eye 3D image quality optimization system according to claim 1, characterized in that: The parameter coupling optimizer: Establish a spatial coordinate transformation matrix based on the viewpoint position; automatically activate the corresponding processing mode according to the input signal metadata; and perform closed-loop collaborative optimization of color gamut, gamma, and backlight parameters.

7. The naked-eye 3D image quality optimization system according to claim 1, characterized in that: The DICOM GSDF Validation Unit: At 400cd / m 2 Enable high dynamic range backlight compensation in this mode to maintain black level brightness ≤ 0.05cd / m 2 ; At 200cd / m 2 In this mode, the low-light color fidelity algorithm is activated to ensure that the sRGB color gamut coverage is ≥ 99%.

Images