Intelligent painting display method based on AR (Augmented Reality)

Abstract

The invention relates to the technical field of AR (Augmented Reality), and particularly discloses an intelligent painting display method based on AR, which comprises the following steps of: acquiring humidity data of a current environment through an environment sensor; generating texture adjustment parameters of the painting work based on the humidity data, wherein the texture adjustment parameters are used for quantifying the influence degree of humidity on the texture of the painting work; dynamically modifying the digital texture of the painting work by using the texture adjustment parameters; and fusing the modified digital texture with the augmented reality scene, and displaying the fused painting on augmented reality equipment in real time. According to the invention, through the dynamic adjustment of the digital texture of the painting work and the fusion of the augmented reality scene, the intelligent self-adaptive display effect of the visual texture of the painting work is realized.

Description

Technical Field

[0001] This invention relates to the field of AR technology, and more specifically to an AR-based intelligent display method for paintings. Background Technology

[0002] AR (Augmented Reality) is a technology that overlays digital content onto a user's real-world environment in real time. It uses devices such as smartphones, tablets, or AR glasses to combine virtual information (such as 3D models, images, and audio) with the real world to expand a user's perception and cognition.

[0003] The physical texture of a painting subtly changes visually with variations in ambient humidity. For example, in humid air, the edges of a watercolor painting naturally diffuse, the texture of layered oil paint becomes more defined in a dry environment, and charcoal particles in a sketch create unique light and shadow reflections under specific humidity levels. However, existing AR display systems use fixed digital textures, typically created in standard laboratory environments, which cannot respond to the constantly changing humidity conditions in real-world settings. When users view a digital watercolor painting through an AR device during the rainy season, although the surrounding humidity has significantly increased, the painting retains the sharp edges of its dry state. This discrepancy in physical properties creates a subtle sense of unease. While users may not be able to pinpoint the problem, it is precisely this subtle difference that affects the immersion and credibility of the AR experience. Summary of the Invention

[0004] The purpose of this invention is to provide an AR-based intelligent display method for paintings, thereby solving the aforementioned technical problems.

[0005] The objective of this invention can be achieved through the following technical solutions: The AR-based intelligent display method for paintings includes the following steps: The humidity data of the current environment is collected through environmental sensors; Texture adjustment parameters for paintings are generated based on humidity data, where the texture adjustment parameters are used to quantify the degree of influence of humidity on the texture of the paintings. The digital texture of a painting is dynamically modified using texture adjustment parameters, including adjusting the blur and contrast properties of the texture. The modified digital textures are blended with the augmented reality scene, and the blended artwork is displayed in real time on the augmented reality device.

[0006] As a further aspect of the present invention, it includes the following steps: The humidity data of the current environment is collected through environmental sensors; Texture adjustment parameters for paintings are generated based on humidity data, where the texture adjustment parameters are used to quantify the degree of influence of humidity on the texture of the paintings. The digital texture of a painting is dynamically modified using texture adjustment parameters, including adjusting the blur and contrast properties of the texture. The modified digital textures are blended with the augmented reality scene, and the blended artwork is displayed in real time on the augmented reality device.

[0007] As a further aspect of the present invention: the extraction of texture adjustment parameters includes: The Sobel operator is used to perform convolution on the texture image to generate gradient components in the horizontal and vertical directions. The gradient magnitude of each pixel is calculated by combining the gradient components in the horizontal and vertical directions, and the gradient magnitudes of all pixels are formed into a gradient magnitude map. Calculate the average value of the gradient magnitude map. The average value reflects the sharpness of the texture edges. Normalize the average value to the range of zero to one. The normalized value is used as the texture blur factor. The texture image is converted to a grayscale image. The standard deviation of the pixel values ​​in the grayscale image is calculated and mapped to a range of zero to one. The mapped value is used as a contrast correction factor.

[0008] As a further aspect of the present invention: dynamically modifying the digital texture of a painting includes: Load the digital texture of the painting as the original texture; The original texture is decomposed into a high-frequency detail layer and a low-frequency base layer. The decomposition process uses a Gaussian filtering algorithm to separate texture components of different frequencies. The high-frequency detail layer is processed by convolution using a texture blur factor. The convolution process uses a Gaussian kernel function to adjust the kernel size according to the texture blur factor. The low-frequency base layer applies a contrast correction factor to perform a linear transformation, which modifies the distribution range of pixel values. The processed high-frequency detail layer and low-frequency base layer are then recombined into a modified digital texture.

[0009] As a further aspect of the present invention: dynamically modifying the digital texture of a painting includes: Load the original digital texture image of the painting; The original digital texture image is decomposed into a high-frequency detail layer and a low-frequency base layer. The decomposition process uses a Gaussian filtering algorithm to separate texture components of different frequencies. The high-frequency detail layer is processed by convolution using a texture blur factor. The convolution process uses a Gaussian kernel function to adjust the kernel size according to the texture blur factor. The low-frequency base layer applies a contrast correction factor to perform a linear transformation, which modifies the distribution range of pixel values. The processed high-frequency detail layer and low-frequency base layer are then recombined into a modified digital texture.

[0010] As a further aspect of the present invention: applying a contrast correction factor to perform a linear transformation on the low-frequency base layer includes: The scaling factor and offset of the pixel value are determined based on the contrast correction factor; The scaling factor is used to expand or compress the dynamic range of pixel values, while the offset is used to adjust the overall brightness of pixel values. The linear transformation formula is that the modified pixel value equals the original pixel value multiplied by the scaling factor plus the offset.

[0011] As a further aspect of the present invention: determining the scaling factor and offset of the pixel value includes: The scaling factor is obtained through linear interpolation. The linear interpolation uses the contrast correction factor as the interpolation parameter and performs interpolation based on the predefined lower limit and upper limit of the scaling factor. The lower limit of the scaling factor corresponds to the scaling factor when the contrast correction factor is zero, and the upper limit of the scaling factor corresponds to the scaling factor when the contrast correction factor is 1. The value obtained by interpolation is used as the scaling factor. The offset is obtained through linear interpolation. The linear interpolation uses the contrast correction factor as the interpolation parameter and performs interpolation based on the predefined lower and upper offset values. The lower offset value corresponds to the offset when the contrast correction factor is zero, and the upper offset value corresponds to the offset when the contrast correction factor is 1. The value obtained by the interpolation calculation is used as the offset.

[0012] As a further aspect of the present invention: merging the modified digital texture with the augmented reality scene includes: Acquire real-time video streams of augmented reality scenes; The background area in the live video stream was identified as the display location of the artwork; The modified digital texture is projected onto the identified background area using texture mapping technology. The mapping process uses perspective transformation to align the geometric relationship between the digital texture and the real environment.

[0013] The beneficial effects of this invention compared to the prior art are as follows: This invention constructs a closed-loop display mechanism integrating environmental perception and dynamic response of digital textures. This allows paintings to exhibit corresponding visual texture changes in augmented reality based on variations in the humidity of the real environment, thus achieving a natural synergy between digital art and the physical environment. The method enables paintings to display subtle changes in edge softening and color diffusion when humidity increases, while maintaining clear brushstrokes and distinct textures in dry environments. This allows the digital image to exhibit visual response characteristics similar to real paintings under different humidity conditions. Through this dynamic change, the AR display effect more closely conforms to actual physical laws, the light and shadow relationship between the painting and the real scene is more harmonious, and the overall visual experience is more natural and realistic. It effectively eliminates the incongruity caused by static textures in traditional AR painting displays, significantly improving visual consistency and immersive experience, and giving digital paintings vivid environmental adaptability and artistic expression in augmented reality environments. Attached Figure Description

[0014] The invention will now be further described with reference to the accompanying drawings.

[0015] Figure 1 This is a flowchart illustrating the AR-based intelligent display method for paintings according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Please see Figure 1 As shown, this invention is an AR-based intelligent display method for paintings, comprising the following steps: The humidity data of the current environment is collected through environmental sensors; Texture adjustment parameters for paintings are generated based on humidity data, where the texture adjustment parameters are used to quantify the degree of influence of humidity on the texture of the paintings. In a preferred embodiment of the present invention, the texture adjustment parameters for generating a painting based on humidity data include: Simulate multiple discrete humidity levels using environmental control equipment; For each humidity level, texture images of the paintings are captured using an image acquisition device, and these texture images encompass the typical features of the paintings' texture. Texture adjustment parameters are extracted from the texture image. These parameters include a texture blur factor and a contrast correction factor. The texture blur factor is obtained by calculating the gradient change of pixels at the texture edge, while the contrast correction factor is obtained by analyzing the distribution range of image pixel values. Each humidity level is associated with its corresponding texture feature parameter, which is used as a data entry in the lookup table, and the index of this data entry is the humidity level. Based on the humidity data of the current environment, the corresponding texture adjustment parameters are obtained from the lookup table.

[0018] In a preferred embodiment, the extraction of texture adjustment parameters includes: The Sobel operator is used to perform convolution on the texture image to generate gradient components in the horizontal and vertical directions. The gradient magnitude of each pixel is calculated by combining the gradient components in the horizontal and vertical directions, and the gradient magnitudes of all pixels are formed into a gradient magnitude map. Calculate the average value of the gradient magnitude map. The average value reflects the sharpness of the texture edges. Normalize the average value to the range of zero to one. The normalized value is used as the texture blur factor. The texture image is converted to a grayscale image. The standard deviation of the pixel values ​​in the grayscale image is calculated and mapped to a range of zero to one. The mapped value is used as a contrast correction factor.

[0019] It should be noted that by acquiring texture images of paintings under different humidity conditions and extracting their edge gradients and pixel distribution features, a correspondence between humidity changes and texture visual features can be established, thus expressing the influence of the physical environment on the image texture in a parametric way. Using gradient amplitude to reflect texture edge sharpness captures the impact of humidity changes on brushstroke sharpness; using pixel value distribution range to reflect contrast changes demonstrates the effect of humidity on overall brightness and darkness. By creating a lookup table of these features at different humidity levels, the system can directly obtain the corresponding texture adjustment parameters after acquiring real-time humidity data, realizing a mapping from physical environmental changes to digital visual responses.

[0020] The digital texture of a painting is dynamically modified using texture adjustment parameters. The digital texture reflects the surface visual characteristics of the painting. The dynamic modification process includes adjusting the blur and contrast attributes of the texture. Another preferred embodiment of the present invention includes dynamically modifying the digital texture of a painting, comprising: Load the digital texture of the painting as the original texture; The original texture is decomposed into a high-frequency detail layer and a low-frequency base layer. The decomposition process uses a Gaussian filtering algorithm to separate texture components of different frequencies. The high-frequency detail layer is processed by convolution using a texture blur factor. The convolution process uses a Gaussian kernel function to adjust the kernel size according to the texture blur factor. The low-frequency base layer applies a contrast correction factor to perform a linear transformation, which modifies the distribution range of pixel values. The processed high-frequency detail layer and low-frequency base layer are then recombined into a modified digital texture.

[0021] One optional implementation process involves reading image data stored in digital media into memory, parsing and standardizing the texture file through an image data interface, converting it into a two-dimensional matrix form that can be recognized by subsequent image processing algorithms. The original texture contains the overall color distribution and surface detail information of the artwork, which together visually determine the texture of the painting.

[0022] To facilitate targeted processing of texture components at different levels, it is necessary to perform frequency separation on the original texture to distinguish the subtle brushstroke information in the high-frequency region from the overall tonal structure in the low-frequency region.

[0023] The decomposition process employs a Gaussian filtering algorithm. By performing a smooth convolution operation on the pixel array of the original texture, local pixel values ​​are weighted by the pixels within their neighborhood, resulting in a low-frequency base layer after detail removal. The Gaussian filter's weight distribution exhibits a characteristic of high weights at the center and low weights at the edges based on spatial distance, which weakens edge intensity while maintaining the overall brightness trend, extracting slowly changing color and brightness information from the image. The difference image between the original texture and the low-frequency base layer constitutes the high-frequency detail layer, which mainly contains visual details such as brushstroke texture and material graininess. During this process, the smoothness can be flexibly controlled by adjusting the spatial size and standard deviation parameters of the filter kernel to ensure that the division between details and basic structure conforms to the natural layers of painting texture. After decomposition, the digital texture is represented as two independent but complementary data layers: the high-frequency detail layer preserves sharp local texture features, while the low-frequency base layer carries the overall color and lighting distribution, providing a structured data foundation for subsequent dynamic texture adjustments.

[0024] In a preferred embodiment, applying a texture blurring factor to the high-frequency detail layer during convolution processing includes: The standard deviation of the Gaussian kernel is calculated based on the texture blur factor. The standard deviation is directly proportional to the texture blur factor; as the texture blur factor increases, the standard deviation increases accordingly. The calculated Gaussian kernel is convolved with the high-frequency detail layer. The convolution operation gradually traverses each pixel region of the high-frequency detail layer to modify the edge regions of the high-frequency detail layer and make them softer.

[0025] In another preferred case of this embodiment, the linear transformation of the low-frequency base layer using the contrast correction factor includes: Determining a scaling factor and an offset for the pixel values according to the contrast correction factor; The scaling factor is used to expand or compress the dynamic range of the pixel values, and the offset is used to adjust the overall brightness of the pixel values; The linear transformation formula is that the modified pixel value is equal to the original pixel value multiplied by the scaling factor plus the offset.

[0026] It should be noted that determining the scaling factor and the offset for the pixel values includes: The scaling factor is obtained through linear interpolation calculation. The linear interpolation calculation uses the contrast correction factor as the interpolation parameter and performs interpolation based on the predefined lower limit value and upper limit value of the scaling factor. The lower limit value of the scaling factor corresponds to the scaling factor when the contrast correction factor is zero, and the upper limit value of the scaling factor corresponds to the scaling factor when the contrast correction factor is 1. The value obtained by the interpolation calculation is used as the scaling factor; The offset is obtained through linear interpolation calculation. The linear interpolation calculation uses the contrast correction factor as the interpolation parameter and performs interpolation based on the predefined lower limit value and upper limit value of the offset. The lower limit value of the offset corresponds to the offset when the contrast correction factor is zero, and the upper limit value of the offset corresponds to the offset when the contrast correction factor is 1. The value obtained by the interpolation calculation is used as the offset.

[0027] Optionally, the gray level range is obtained by statistically calculating the minimum value and the maximum value of the pixels in the low-frequency base layer, and this range is mapped to the standard display interval, and overall proportional scaling is performed under the condition of keeping the relative brightness relationship between pixels unchanged. The global proportional factor used in the normalization mapping reflects the stretching and shrinking relationship between the original gray level range and the standard interval. This proportion is the lower limit value of the scaling factor, which is used to maintain the unchanged bright-dark relationship of the image when the contrast correction factor is zero. The lower limit value of the scaling factor is obtained based on the gray level range of the original image, and the basic proportion of the overall brightness distribution is determined through normalization, so as to enable the system to faithfully maintain the original light and dark relationship without any artificial gain or compression when the contrast is not enhanced, and make the output image in the zero correction state equivalent to the input image.

[0028] Without altering the image's brightness distribution, a baseline value representing the overall brightness balance of the image is determined by statistically analyzing the median brightness of the low-frequency base layer grayscale histogram. This median brightness is set as the brightness reference point in a zero-offset state, ensuring the image's brightness center remains stable without contrast adjustment. This brightness baseline serves as the lower limit of the offset. The determination of the lower limit depends on the median brightness of the grayscale distribution because the median stably reflects the overall brightness balance point, unaffected by extreme highlights or shadows. Using the median brightness as the zero-offset baseline ensures that the image's brightness center does not shift without contrast adjustment, maintaining the original visual balance.

[0029] The grayscale distribution of the low-frequency base layer is expanded to bring the pixel values ​​in bright and dark areas closer to the boundaries of the display range, achieving the maximum brightness difference without oversaturation or overflow. By comparing the brightness range ratios before and after expansion, the resulting magnification factor serves as the upper limit of the scaling factor, used to achieve dynamic range expansion under maximum contrast. Setting the upper limit of the scaling factor, by expanding the grayscale range to obtain the maximum brightness difference, establishes a reference boundary under maximum contrast correction conditions. The magnification ratio at this point represents the maximum brightness scaling the image can withstand, serving as the upper limit endpoint of linear interpolation, ensuring a clear range of contrast variation from the original state to the enhanced state.

[0030] The average brightness of the image is recalculated at maximum contrast and compared with the target neutral brightness reference. The difference between the two is used as a brightness compensation amount to balance the overall brightness shift after expansion. This compensation value is the upper limit of the offset, used to maintain the overall brightness stability of the image when the contrast correction factor is one. The upper limit of the offset is obtained by recalculating the average brightness after maximum contrast expansion and comparing it with the neutral brightness reference. This brightness compensation amount is used to correct the brightness shift caused by dynamic range stretching, so that the image still maintains a natural overall brightness balance under strong contrast.

[0031] It is worth noting that when performing convolution processing on the high-frequency detail layer, the shape of the Gaussian kernel used for convolution is determined based on the obtained texture blur factor. The standard deviation of the Gaussian kernel changes proportionally to the texture blur factor (the ratio is set empirically). When the blur factor is large, the standard deviation increases accordingly, allowing the convolution operation to perform weighted averaging of pixels over a wider neighborhood. During the generation of the Gaussian kernel, the weight values ​​at each position are calculated based on the two-dimensional spatial distribution function, so that the weights are maximized at the center and smoothly decrease towards the periphery, thereby constructing a weighted template with spatial continuity.

[0032] After the Gaussian kernel is established, it is convolved point-by-point with the pixel matrix of the high-frequency detail layer. During the operation, the output value of each pixel is obtained by weighted summation of its neighboring pixels according to Gaussian weights. This traversal process covers the entire image area, gradually smoothing strong gradient transitions at the edges and achieving the effect of softening the edges of high-frequency parts. After the convolution operation, the resulting high-frequency layer, while maintaining the overall detail structure, weakens overly sharp stroke variations, providing a smooth detail foundation for subsequent recombination with the low-frequency layer.

[0033] When performing a linear transformation on the low-frequency base layer, two key parameters are determined based on the contrast correction factor: a scaling factor that controls the range of brightness and darkness, and an offset that adjusts the overall brightness. The scaling factor is determined within a preset upper and lower limit range through interpolation; the larger the value, the wider the dynamic range of the pixel. The offset is calculated using the same interpolation method and is used to change the overall brightness level. Each pixel value of the original low-frequency layer is sequentially multiplied by the scaling factor and then added to the offset to obtain a new pixel value distribution. This linear transformation is performed point-by-point at each pixel location in the image, orderly adjusting the relative differences between dark and bright areas. After processing, the low-frequency base layer is redistributed in terms of overall brightness and contrast, forming a harmonious visual hierarchy with the high-frequency layer processed by convolution. This lays a balanced foundation of light and hierarchy for the final synthesis of digital textures with environmental responsiveness.

[0034] The modified digital textures are blended with the augmented reality scene, and the blended artwork is displayed in real time on the augmented reality device.

[0035] Another preferred embodiment of the present invention includes fusing the modified digital texture with the augmented reality scene, comprising: Acquire real-time video streams of augmented reality scenes; The background area in the live video stream was identified as the display location of the artwork; The modified digital texture is projected onto the identified background area using texture mapping technology. The mapping process uses perspective transformation to align the geometric relationship between the digital texture and the real environment.

[0036] Specifically, during the fusion of the modified digital texture with the augmented reality scene, the augmented reality device continuously captures real-time video streams of the user's environment through its built-in camera components. Each frame of the video stream is input into the image recognition module as the spatial carrier for displaying the artwork. The image recognition module uses feature detection algorithms to analyze the geometric shape, edge structure, and lighting distribution in the image to determine the planar area suitable for projecting the artwork. This area is typically determined based on stable and regularly shaped background objects in the scene. Key points are extracted using matching algorithms, and their spatial positional relationships are calculated to establish the correspondence between the environmental coordinate system and the camera coordinate system, thereby obtaining the spatial pose information of the display plane.

[0037] After obtaining the display area, the modified digital texture is loaded into the rendering module as input for the virtual image to be mapped. Texture mapping technology is used during the mapping process to map the pixel coordinates of the digital texture to the corresponding 3D coordinates of the identified display area. To ensure that the geometric proportions and viewpoint direction of the artwork are consistent with the real environment when presented in the augmented reality scene, perspective transformation calculations are applied during the mapping operation. The perspective transformation calculation is based on camera parameters, the position of the display plane, and the user's viewpoint direction. By establishing a projection matrix, the virtual image is aligned with the perspective of the real space, so that the digital texture produces the same perspective distortion effect as real objects on the camera's imaging plane. During rendering, the texture brightness is adaptively adjusted according to the scene's lighting conditions to ensure that the virtual texture and the real-time video stream maintain coordination in terms of lighting levels, avoiding discontinuous boundaries or abrupt color differences.

[0038] The final merged image is output in real time as a video overlay on the display of the augmented reality device. When users observe it through the device view, they can see that the modified painting is precisely attached to the display area of ​​the real environment, forming a continuous, stable and spatially consistent visual fusion effect.

[0039] This solution fundamentally solves the problem of digital textures in existing AR displays failing to respond to changes in real-world humidity by constructing a dynamic closed-loop system of "perception-mapping-adjustment-rendering." Its core principle lies in: real-time collection of humidity data from the user's surroundings using environmental sensors, enabling the system to continuously perceive changes in external physical conditions; mapping the humidity data to a pre-established humidity-texture parameter lookup table, achieving a precise mapping from environmental physical quantities to image visual attributes—this mapping relationship allows humidity changes to be transformed into controllable texture adjustment factors; the system performs frequency domain decomposition on the digital texture of the artwork, processing the high-frequency detail layer and the low-frequency base layer separately, and dynamically modifying them based on the mapped parameters, thereby simulating the influence of humidity on pigment diffusion, brightness levels, and edge sharpness; and merging the modified digital texture with the augmented reality scene through perspective transformation, ensuring that the visual effect remains consistent with the real environment in terms of spatial geometry and lighting relationships.

[0040] Through this complete process, changes in environmental humidity are no longer merely external information, but are transformed into an internal mechanism driving dynamic changes in the texture of the painting. This allows digital paintings in AR to "physically respond" to the environment like real paintings, effectively eliminating the incongruity caused by static textures and enhancing the naturalness and immersion of the display. This principle of dynamic texture adjustment based on physical perception enables AR paintings to reproduce the artistic texture of a real environment at the visual level, achieving a sensory unity between the virtual and the real.

[0041] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. An AR-based intelligent display method for paintings, characterized in that, Includes the following steps: The humidity data of the current environment is collected through environmental sensors; Texture adjustment parameters for paintings are generated based on humidity data, where the texture adjustment parameters are used to quantify the degree of influence of humidity on the texture of the paintings. The digital texture of a painting is dynamically modified using texture adjustment parameters. The digital texture reflects the surface visual characteristics of the painting. The dynamic modification process includes adjusting the blur and contrast attributes of the texture. The modified digital textures are blended with the augmented reality scene, and the blended artwork is displayed in real time on the augmented reality device.

2. The AR-based intelligent display method for paintings according to claim 1, characterized in that, Texture adjustment parameters for generating paintings based on humidity data include: Simulate multiple discrete humidity levels using environmental control equipment; For each humidity level, texture images of the paintings are captured using an image acquisition device, and these texture images encompass the typical features of the paintings' texture. Texture adjustment parameters are extracted from the texture image. These parameters include a texture blur factor and a contrast correction factor. The texture blur factor is obtained by calculating the gradient change of pixels at the texture edge, while the contrast correction factor is obtained by analyzing the distribution range of image pixel values. Each humidity level is associated with its corresponding texture feature parameter, which is used as a data entry in the lookup table, and the index of this data entry is the humidity level. Based on the humidity data of the current environment, the corresponding texture adjustment parameters are obtained from the lookup table.

3. The AR-based intelligent display method for paintings according to claim 2, characterized in that, Extracted texture adjustment parameters include: The Sobel operator is used to perform convolution on the texture image to generate gradient components in the horizontal and vertical directions. The gradient magnitude of each pixel is calculated by combining the gradient components in the horizontal and vertical directions, and the gradient magnitudes of all pixels are formed into a gradient magnitude map. Calculate the average value of the gradient magnitude map. The average value reflects the sharpness of the texture edges. Normalize the average value to the range of zero to one. The normalized value is used as the texture blur factor. The texture image is converted to a grayscale image. The standard deviation of the pixel values ​​in the grayscale image is calculated and mapped to a range of zero to one. The mapped value is used as a contrast correction factor.

4. The AR-based intelligent display method for paintings according to claim 1, characterized in that, Dynamically modifying the digital textures of paintings includes: Load the digital texture of the painting as the original texture; The original texture is decomposed into a high-frequency detail layer and a low-frequency base layer. The decomposition process uses a Gaussian filtering algorithm to separate texture components of different frequencies. The high-frequency detail layer is processed by convolution using a texture blur factor. The convolution process uses a Gaussian kernel function to adjust the kernel size according to the texture blur factor. The low-frequency base layer applies a contrast correction factor to perform a linear transformation, which modifies the distribution range of pixel values. The processed high-frequency detail layer and low-frequency base layer are then recombined into a modified digital texture.

5. The AR-based intelligent display method for paintings according to claim 4, characterized in that, Applying a texture blur factor to the convolution process of the high-frequency detail layer includes: The standard deviation of the Gaussian kernel is calculated based on the texture blur factor. The standard deviation is directly proportional to the texture blur factor; as the texture blur factor increases, the standard deviation increases accordingly. The calculated Gaussian kernel is convolved with the high-frequency detail layer. The convolution operation gradually traverses each pixel region of the high-frequency detail layer to modify the edge regions of the high-frequency detail layer and make them softer.

6. The AR-based intelligent display method for paintings according to claim 4, characterized in that, Applying a contrast correction factor to perform a linear transformation on the low-frequency base layer includes: The scaling factor and offset of the pixel value are determined based on the contrast correction factor; The scaling factor is used to expand or compress the dynamic range of pixel values, while the offset is used to adjust the overall brightness of pixel values. The linear transformation formula is: the modified pixel value equals the original pixel value multiplied by the scaling factor plus the offset.

7. The AR-based intelligent display method for paintings according to claim 6, characterized in that, Determining the scaling factor and offset of pixel values ​​includes: The scaling factor is obtained through linear interpolation. The linear interpolation uses the contrast correction factor as the interpolation parameter and performs interpolation based on the predefined lower limit and upper limit of the scaling factor. The lower limit of the scaling factor corresponds to the scaling factor when the contrast correction factor is zero, and the upper limit of the scaling factor corresponds to the scaling factor when the contrast correction factor is 1. The value obtained by interpolation is used as the scaling factor. The offset is obtained through linear interpolation. The linear interpolation uses the contrast correction factor as the interpolation parameter and performs interpolation based on the predefined lower and upper offset values. The lower offset value corresponds to the offset when the contrast correction factor is zero, and the upper offset value corresponds to the offset when the contrast correction factor is 1. The value obtained by the interpolation calculation is used as the offset.

8. The AR-based intelligent display method for paintings according to claim 1, characterized in that, Integrating modified digital textures with augmented reality scenes includes: Acquire real-time video streams of augmented reality scenes; The background area in the live video stream was identified as the display location of the artwork; The modified digital texture is projected onto the identified background area using texture mapping technology. The mapping process uses perspective transformation to align the geometric relationship between the digital texture and the real environment.

Images