TV image processing method and device based on scene recognition, equipment and medium

Abstract

The invention relates to the technical field of television image optimization processing. The invention relates to a TV television image processing method and device based on scene recognition, equipment and a medium, and the method comprises the steps: obtaining scene types in a television image, enabling the scene types to comprise a dynamic scene type and a static scene type, and carrying out the optimization of the scene types in the television image based on preset optimization parameters according to the scene types in the television image, and dynamically adjusting the brightness, the contrast ratio and the sharpening degree of the television image to obtain a processed television image, and carrying out noise reduction and detail enhancement on the processed television image to generate a final optimized television image. The method and the device have the effect of improving the adaptive processing capability of the television image.

Description

Technical Field

[0001] The present invention relates to the technical field of TV image optimization processing, and in particular to a TV image optimization processing method based on scene recognition. Background Art

[0002] With the continuous development of technology, high-definition TV and smart TV have become an important part of modern family entertainment life. The quality of TV images directly affects the audience's viewing experience. At present, when TV displays different types of scenes, such as dynamic scenes and static scenes, it usually uses unified image processing parameters, resulting in poor display effects in different scenes. Summary of the invention

[0003] In order to improve the adaptive processing capability of television images, the present application provides a TV image processing method, device, equipment and medium based on scene recognition.

[0004] The above-mentioned invention objective of the present application is achieved through the following technical solutions: A TV image processing method based on scene recognition, the TV image processing method based on scene recognition comprising: Acquire a scene type in a television picture, wherein the scene type includes a dynamic scene type and a static scene type; According to the scene type in the TV picture, based on preset optimization parameters, dynamically adjust the brightness, contrast and sharpness of the TV image to obtain a processed TV image; The processed television image is subjected to noise reduction and detail enhancement to generate a final optimized television image.

[0005] By adopting the above technical solutions, by distinguishing between dynamic scenes and static scenes, the system can adopt appropriate processing strategies according to the different characteristics of the scenes. For example, dynamic scenes may require higher motion compensation, while static scenes can pay more attention to detail clarity and contrast. Dynamic analysis of the color, brightness and texture of the TV picture ensures that the system can respond to changes in the picture in real time and optimize the image processing effect. According to the scene type, the brightness, contrast and sharpness are dynamically adjusted so that the image processing can adapt to different types of content requirements. For example, objects in motion may require higher brightness and clarity, while static images require enhanced contrast and details. The system can automatically adjust the processing parameters according to different scene types (dynamic or static) to enhance the visual effect of the image. The noise reduction algorithm is applied to the processed image to reduce noise and improve the clarity of the image. Through the detail enhancement algorithm, the details and texture in the image are enhanced to make the image more layered and clear.

[0006] In a preferred example, the present application may be further configured as follows: the scene type in the TV picture is obtained, and the scene type includes a dynamic scene type and a static scene type, including: By performing real-time feature analysis on the color, brightness and texture of the TV screen image, preliminary scene features are obtained; Based on the preliminary scene features, the television screen image is classified to identify the dynamically changing area and the static area in the television screen image; A motion feature analysis is performed on the dynamically changing area to identify the dynamic scene type, and a static feature analysis is performed on the static area to identify the static scene type.

[0007] By adopting the above technical solution, the color, brightness and texture of the TV screen image are analyzed in real time to obtain preliminary scene features, so that the dynamically changing areas and static areas in the picture can be identified. Secondly, by performing motion feature analysis on the dynamically changing areas and static feature analysis on the static areas, the dynamic scene types and static scene types can be accurately identified. This detailed classification and analysis enables the TV to make more accurate image adjustments based on the specific scene type, ensuring that the image processing algorithm can adapt to different picture features, thereby improving the image clarity and visual experience.

[0008] In a preferred example, the present application may be further configured as follows: performing motion feature analysis on the dynamically changing area to identify the dynamic scene type, and performing static feature analysis on the static area to identify the static scene type, including: By performing the motion feature analysis on the dynamically changing area, the motion feature of the dynamically changing area is obtained, and the motion feature of the dynamically changing area is compared with known dynamic scene types in a preset scene detection library to identify the dynamic scene type; The static feature analysis is performed on the static area to obtain the features of the static scene, and the features of the static scene are compared with the known static scene types in the preset scene detection library to identify the static scene type.

[0009] By adopting the above technical solution, by accurately identifying the dynamic and static scenes in the TV picture, it is ensured that each scene type can be accurately classified. By identifying the specific dynamic scene type and static scene type, the most appropriate image processing algorithm can be applied to different scenes. For example, dynamic scenes may need to optimize motion blur and clarity, while static scenes focus on enhancing details and contrast. Accurate scene recognition enables the TV image processing system to more effectively apply preset optimization parameters, reduce errors and unnecessary processing, and improve the efficiency and effect of image processing.

[0010] In a preferred example, the present application can be further configured as follows: dynamically adjusting the brightness, contrast and sharpness of the TV image based on the preset optimization parameters according to the scene type in the TV screen to obtain the processed TV image, including: According to the dynamic scene type, based on the preset optimization parameters, the brightness of the area where the moving object is located is increased; according to the static scene type, based on the preset optimization parameters, the contrast of the area where the moving object is located is maintained; according to the static scene type, based on the preset optimization parameters, the area where the moving object is located is sharpened; According to the static scene type, based on the preset optimization parameters, the global brightness of the TV image is increased; according to the static scene type, based on the preset optimization parameters, the contrast between the dark and bright parts of the TV image is enhanced; according to the static scene type, based on the preset optimization parameters, the global sharpness of the TV image is increased.

[0011] By adopting the above technical solutions, increasing the brightness and sharpening of moving objects in dynamic scenes can effectively reduce motion blur, enhance detail performance, and make the motion picture clearer. Increasing the global brightness and contrast of static scenes, as well as global sharpening, can enhance the overall details and visual effects of static images, making the images more vivid and delicate. By applying appropriate processing parameters according to different scene types, it can better adapt to various viewing environments, provide more appropriate image quality, and enhance the user's visual experience. Accurately adjusting the brightness, contrast, and sharpening parameters can avoid image distortion caused by excessive or inappropriate processing, ensuring that the processed images remain natural and realistic.

[0012] In a preferred example, the present application may be further configured as follows: the noise reduction and detail enhancement of the processed television image to generate a final optimized television image includes: Preliminarily removing noise points in the processed television image by an image noise reduction algorithm to obtain a noise-reduced image; Performing texture and edge feature analysis on the denoised image to obtain an image analysis result, and enhancing details of the denoised image based on the image analysis result to generate a detail-enhanced image; The detail-enhanced image is subjected to edge-preserving filtering to obtain a sharpened image, and the sharpened image is subjected to rendering to generate the final optimized television image.

[0013] By adopting the above technical solution, the image noise reduction algorithm is used to remove the noise in the processed image, which helps to reduce interference and ensure that the subsequent image processing effect is clearer and more accurate. Through texture and edge feature analysis, the noise-reduced image is enhanced in detail, making the details and edges in the image more obvious, thereby improving the visual effect and information content of the image. Edge-preserving filtering is applied to sharpen the image while maintaining the naturalness and details of the edge, avoiding image distortion caused by over-sharpening. The sharpened image is rendered to further enhance the visual effect of the image, making the final optimized TV image clearer, more realistic and vivid.

[0014] In a preferred example, the present application can be further configured as follows: the TV image processing method based on scene recognition further includes: Real-time detection of changes in brightness and color temperature of the ambient light to obtain ambient light information; Based on the ambient light information, the final optimized television image is adjusted to compensate for the color cast caused by the ambient light change.

[0015] By adopting the above technical solution, the lighting conditions of the current viewing environment can be dynamically captured by detecting the changes in the brightness and color temperature of the ambient light. This allows the TV image processing system to understand and respond to changes in the ambient light in a timely manner. The final optimized TV image is adjusted based on the ambient light information to compensate for the color shift caused by the changes in ambient light. This ensures that the color of the TV image remains accurate and consistent under different lighting conditions, thereby improving the visual quality and viewing experience of the image.

[0016] In a preferred example, the present application can be further configured as follows: the TV image processing method based on scene recognition further includes: constructing a user preference model according to the operation information of the user adjusting the television screen image; Based on the user preference model, image processing parameters are automatically adjusted by analyzing the user's preferences for different scenes.

[0017] By adopting the above technical solution, the user's adjustment behavior on the TV screen image is recorded to generate a user preference model that can accurately reflect each user's image processing needs and visual preferences. Using the user preference model, the system can automatically adjust the image processing parameters to meet the user's personalized needs. This adjustment makes the TV image more in line with the user's specific preferences, thereby improving the viewing experience.

[0018] The second object of the invention is achieved by the following technical solutions: A TV image processing device based on scene recognition, the TV image processing device based on scene recognition comprising: A scene type acquisition module, used to acquire the scene type in the TV picture, wherein the scene type includes a dynamic scene type and a static scene type; A television image processing module, used to dynamically adjust the brightness, contrast and sharpness of the television image according to the scene type in the television picture based on preset optimization parameters to obtain a processed television image; The television image generation module is used to perform noise reduction and detail enhancement on the processed television image to generate a final optimized television image.

[0019] By adopting the above technical solutions, by distinguishing between dynamic scenes and static scenes, the system can adopt appropriate processing strategies according to the different characteristics of the scenes. For example, dynamic scenes may require higher motion compensation, while static scenes can pay more attention to detail clarity and contrast. Dynamic analysis of the color, brightness and texture of the TV picture ensures that the system can respond to changes in the picture in real time and optimize the image processing effect. According to the scene type, the brightness, contrast and sharpness are dynamically adjusted so that the image processing can adapt to different types of content requirements. For example, objects in motion may require higher brightness and clarity, while static images require enhanced contrast and details. The system can automatically adjust the processing parameters according to different scene types (dynamic or static) to enhance the visual effect of the image. The noise reduction algorithm is applied to the processed image to reduce noise and improve the clarity of the image. Through the detail enhancement algorithm, the details and texture in the image are enhanced to make the image more layered and clear.

[0020] The third objective of the present application is achieved through the following technical solutions: A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the TV image processing method based on scene recognition are implemented.

[0021] The fourth objective of the present application is achieved through the following technical solutions: A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the TV image processing method based on scene recognition are implemented.

[0022] In summary, the present application includes at least one of the following beneficial technical effects: 1. By distinguishing between dynamic scenes and static scenes, the system can adopt appropriate processing strategies according to the different characteristics of the scenes. For example, dynamic scenes may require higher motion compensation, while static scenes can pay more attention to detail clarity and contrast. Dynamically analyze the color, brightness and texture of the TV picture to ensure that the system can respond to changes in the picture in real time and optimize the image processing effect. According to the scene type, dynamically adjust the brightness, contrast and sharpness so that image processing can adapt to different types of content requirements. For example, objects in motion may require higher brightness and clarity, while static images require enhanced contrast and details. The system can automatically adjust processing parameters according to different scene types (dynamic or static) to enhance the visual effect of the image. Apply noise reduction algorithms to the processed images to reduce noise and improve image clarity. Through the detail enhancement algorithm, the details and texture in the image are enhanced to make the image more layered and clear; 2. Ensure that each scene type is accurately classified by accurately identifying dynamic and static scenes in the TV screen. By identifying specific dynamic scene types and static scene types, the most appropriate image processing algorithm can be applied to different scenes. For example, dynamic scenes may need to optimize motion blur and clarity, while static scenes focus on enhancing details and contrast. Accurate scene recognition enables the TV image processing system to more effectively apply preset optimization parameters, reduce errors and unnecessary processing, and improve the efficiency and effectiveness of image processing; 3. Record the user's adjustment behavior on the TV screen image and generate a user preference model that can accurately reflect each user's image processing needs and visual preferences. Using the user preference model, the system can automatically adjust the image processing parameters to meet the user's personalized needs. This adjustment makes the TV image more in line with the user's specific preferences, thereby improving the viewing experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flow chart of a TV image processing method based on scene recognition in one embodiment of the present application; Figure 2 is a flowchart of the implementation of step S10 of the TV image processing method based on scene recognition in one embodiment of the present application; Figure 3 is a flowchart for implementing step S12 of a TV image processing method based on scene recognition in an embodiment of the present application; Figure 4 is a flowchart of the implementation of step S20 of the TV image processing method based on scene recognition in one embodiment of the present application; Figure 5 is a flowchart of the implementation of step S30 of the TV image processing method based on scene recognition in one embodiment of the present application; Figure 6 This is a flowchart of the implementation of the TV image processing method based on scene recognition after step S10 in one embodiment of the present application; Figure 7 This is a flowchart of the implementation of the TV image processing method based on scene recognition after step S10 in one embodiment of the present application; Figure 8 It is a principle block diagram of a TV image processing method based on scene recognition in one embodiment of the present application; Fig. 9 It is a schematic diagram of a device in an embodiment of the present application. DETAILED DESCRIPTION

[0024] The present application is further described in detail below in conjunction with the accompanying drawings.

[0025] In one embodiment, if Figure 1 As shown, the present application discloses a TV image processing method based on scene recognition, which specifically includes the following steps: S10: Acquire a scene type in the TV picture, where the scene type includes a dynamic scene type and a static scene type.

[0026] In the present embodiment, the scene type in the TV picture refers to the classification of scenes according to the dynamic features and static features of the picture content, so as to optimize the image processing parameters. The scene type is identified by analyzing the image content in the TV picture, and classifying the scene into a dynamic scene type or a static scene type according to the features such as motion change, color change, brightness change, etc. The dynamic scene type refers to a scene containing a large amount of motion or change in the TV picture. For example, objects in motion, fast-switching pictures, action scenes in movies or videos, sports games, etc., the picture content of these scenes changes. The static scene type refers to a scene in the TV picture that is mainly static or has little change. For example, landscape pictures, close-ups of people, text displays, news broadcast scenes, etc., the picture content of these scenes is stable and there is no drastic change.

[0027] Specifically, by collecting image data of continuous frames in the TV screen and extracting image features using methods such as motion detection, color histogram analysis, texture analysis and edge detection, the system can identify the scene type in the TV screen in real time.

[0028] Furthermore, the system refers to an image processing system integrated into a TV. The system includes hardware and software components, such as image sensors, processors, and program modules for executing image analysis and optimization algorithms. Through the collaborative work of these components, the system can collect and process TV picture data in real time, identify the scene type in the picture, and automatically adjust the image display parameters according to different scene types.

[0029] S20: According to the scene type in the TV picture, based on preset optimization parameters, dynamically adjust the brightness, contrast and sharpness of the TV image to obtain a processed TV image.

[0030] In this embodiment, the preset optimization parameters refer to a set of image processing parameters pre-set for different scene types (such as dynamic scenes and static scenes). Brightness refers to the overall light intensity displayed in a TV image, which affects the visibility of the darkest and brightest parts of the TV image. Contrast refers to the difference between the bright and dark parts of a TV image. Sharpness refers to the clarity of the edges of a TV image and the prominence of details. A processed TV image refers to an image generated after brightness, contrast, and sharpening adjustments.

[0031] Specifically, when the system recognizes the current scene type, it will automatically call the corresponding preset optimization parameters. For example, in a dynamic scene, the system will choose parameters that increase brightness, enhance contrast, and sharpen appropriately, while in a static scene, it may choose parameters that reduce sharpness and moderate contrast. For LCD TVs, the system can change the brightness of the overall picture by adjusting the brightness of the TV backlight. At the same time, local dimming can be performed for different areas of the picture to make the brightness more in line with viewing needs. By adjusting the brightness value of the image, the light intensity of the entire picture is adjusted. The system increases or decreases the brightness value of each pixel according to the preset brightness parameters. The system adjusts the brightness histogram of the image to make the brightness distribution more uniform and improve the contrast of the image. For example, in a dynamic scene, the system may stretch the brightness range of the histogram to enhance the contrast between light and dark. The contrast of different areas in the image is refined and adjusted to make the image more vivid by enhancing the contrast between dark and bright areas. The system uses edge detection algorithms to identify contours and edges in the image, and focuses on these parts to improve the sharpness of the image. Image details are enhanced by applying a sharpening filter (such as a Laplacian filter or a nonlinear sharpening filter) to make edges clearer.

[0032] Furthermore, the above processing is performed while the television receives the video signal, ensuring that the output image is a dynamically adjusted effect.

[0033] S30: performing noise reduction and detail enhancement on the processed television image to generate a final optimized television image.

[0034] In this embodiment, the final optimized television image refers to a high-quality image generated after undergoing noise reduction and detail enhancement processing.

[0035] Specifically, firstly, the processed TV image is analyzed by the spatial domain method to identify the noise part in the processed TV image. Through the mean filtering algorithm, for each pixel in the image, the average value of the surrounding neighborhood pixels of the processed TV image is calculated, and the original pixel value is replaced by this average value, so as to smooth the image, reduce the noise, and obtain the denoised TV image. The Laplace operator algorithm is used on the denoised TV image to identify the edge part of the denoised TV image. By performing a deconvolution operation on the image, the local contrast of the image is enhanced, so as to highlight the texture and details of the image. While enhancing the details, the image is edge-preserving filtered to ensure that the edges and details of the image are clear and sharp without introducing artifacts or over-sharpening. The image after detail enhancement and edge preservation processing is finally rendered to adapt to the display characteristics of the TV screen and ensure the quality of the image when it is displayed.

[0036] Furthermore, the steps of noise reduction and detail enhancement are combined to generate an optimized image with low noise, rich details and good visual effects. This optimized image will be displayed on the TV screen as the final output image. The mean filter algorithm refers to an image processing technique that smoothes the image by taking the average of each pixel in the image and the pixels in the surrounding neighborhood. The Laplace operator algorithm refers to an image edge detection technique based on the second-order derivative. It detects edges by calculating the second-order derivative of the change in pixel brightness in the image, highlighting the details and edges in the image.

[0037] In one embodiment, if Figure 2 As shown, in step S10, that is, obtaining the scene type in the TV picture, the scene type includes a dynamic scene type and a static scene type, including: S11: Preliminary scene features are obtained by performing real-time feature analysis on the color, brightness and texture of the TV screen image.

[0038] In this embodiment, the preliminary scene features refer to key image features extracted through real-time analysis of basic attributes of the television screen image, such as color, brightness, and texture.

[0039] Specifically, the tone of the scene is judged by extracting the distribution and saturation of the main colors in the picture. For example, a bright and colorful picture may represent a dynamic scene, while a picture with a single tone and low saturation may represent a static scene. By detecting the brightness values ​​of the pixels in the picture, the distribution of highlight areas and low-brightness areas in the picture is identified. Rapid changes in high brightness usually indicate dynamic scenes, such as fast switching shots in action movies, while pictures with stable brightness are more likely to be static scenes. The dynamic or static nature of the scene is distinguished by calculating the texture complexity of different areas in the picture, such as edge density and detail richness. Complex textures and high edge density usually mean dynamic scenes, such as people or scenery in motion, while smooth, simple-textured areas are called static scenes.

[0040] S12: Based on the preliminary scene features, the television screen image is classified to identify dynamically changing areas and static areas in the television screen image.

[0041] In this embodiment, the dynamically changing area refers to the area in the TV picture where there is significant movement or frequent changes. These areas usually contain moving objects, rapidly changing light and shadow or other dynamic elements, such as people in motion, traffic, swaying trees, etc. The characteristics of the dynamically changing area are that the image pixels change frequently and the color, brightness or texture change greatly. The static area refers to the part of the TV picture that changes less or almost does not change. These areas usually contain static backgrounds, fixed objects or scenes that remain in the same state for a long time, such as walls, skies, distant buildings, etc. The characteristics of the static area are that the pixels change less frequently and the color, brightness and texture remain relatively stable.

[0042] Specifically, by comparing the pixel values ​​in the picture frame by frame, areas with significant changes are detected. These changes can be sudden changes in color, drastic changes in brightness, or rapid changes in texture. Using the frame difference method, the motion vector of each pixel in the image is calculated to identify which areas have significant motion. By comparing the pixel changes in adjacent frames, if the color, brightness, or texture of the pixels do not change between multiple frames, these areas are marked as static areas. If the color, brightness, or texture of the pixels change between multiple frames, these areas are marked as dynamically changing areas. By observing the change trends of specific areas over a long period of time, stable background areas are identified and regarded as static areas.

[0043] S13: Performing motion feature analysis on the dynamically changing area to identify the dynamic scene type, and performing static feature analysis on the static area to identify the static scene type.

[0044] Specifically, the motion direction and speed of the object are determined by calculating the motion vector of the pixels in the image. The optical flow method can capture the motion information of all pixels in the image and generate a motion vector field to identify the characteristics of the motion. The motion vectors in the dynamic area are clustered to identify the main motion mode. Different types of dynamic scenes may have different motion characteristics, such as translation, rotation, scaling, etc. Through clustering, these motion patterns can be identified and matched with the preset dynamic scene types. The analyzed motion features are compared with the known dynamic scene types in the preset scene detection library. For example, scenes such as running, driving, and playing ball have their own unique motion characteristics. Through comparison, the type of the current scene can be determined. The goal of static feature analysis is to identify the static scene types corresponding to these areas by extracting specific features in static areas. Texture feature extraction is performed on static areas, and the texture pattern in the area is analyzed using methods such as gray level co-occurrence matrix (GLCM). Different types of static scenes usually have different texture features. For example, a forest scene may have a delicate and complex texture, while a sky scene may have a smooth texture. The color distribution in the static area is statistically analyzed to generate a color histogram. The color histogram can reflect the main color tone and color distribution characteristics of the area, and is used to distinguish different static scenes, such as grass, ocean, desert, etc. The extracted static features are matched with the known static scene types in the preset scene detection library. For example, some indoor scenes may have specific color combinations and texture patterns, while outdoor scenes may have different color and texture combinations. Through matching, the scene type of the static area can be identified. After identifying the scene types of the dynamically changing area and the static area, they are combined to form a complete understanding of the entire picture. For example, in a picture, the dynamic area in the foreground may be "pedestrians crossing the road", while the static area in the background may be "city streets". Through this combination, a more accurate and detailed scene description can be generated. For example, in a TV picture, the foreground is a running person and the background is a static city landscape. The system first identifies the dynamic area in the foreground and determines it as the "running" scene type through motion feature analysis. Then, the system performs static feature analysis on the static area in the background and determines it as the "city street" scene type. Finally, the system combines these two analysis results and concludes that the scene type of the current TV picture is "running person on a city street".

[0045] Furthermore. For example, the gray-level co-occurrence matrix (GLCM) method refers to a statistical method for describing image texture. It generates a matrix by calculating the gray-level relationship between pixel pairs in the image, which represents the gray-level co-occurrence of pixel pairs at a specific distance and direction. The known dynamic scene types in the preset scene detection library refer to a set of pre-stored dynamic scene features and labels that have been analyzed and classified. These scene types cover common sports scenes, such as running, jumping, driving, etc., and each type corresponds to a specific set of sports features. The known static scene types in the preset scene detection library refer to pre-stored, analyzed and classified static scene features and labels. These types include common static scenes, such as urban landscapes, indoor environments, natural scenery, etc., and each type corresponds to a specific set of color and texture features.

[0046] In one embodiment, if Figure 3 As shown, in step S13, that is, performing motion feature analysis on the dynamically changing area to identify the dynamic scene type, and performing static feature analysis on the static area to identify the static scene type, includes: S131: obtaining motion features of the dynamically changing region by performing the motion feature analysis on the dynamically changing region, and comparing the motion features of the dynamically changing region with known dynamic scene types in a preset scene detection library to identify the dynamic scene type.

[0047] In this embodiment, the motion characteristics of the dynamically changing area refer to the motion state of the pixels in the TV picture, including the motion direction, speed, trajectory and motion mode.

[0048] Specifically, for each frame of image, by calculating the difference between the frame image and the previous frame image of the current frame image, if the difference exceeds a preset threshold, the pixel point is marked as a dynamic change point to form a dynamic change area. The optical flow method is used to calculate the movement direction of the pixel points in the dynamic change area. The optical flow method obtains the main movement direction in the area by analyzing the motion vectors of the pixel points in adjacent frames. The movement speed of the pixel points in the dynamic change area is calculated according to the optical flow method. The speed is estimated by comparing the change distance of the position of the same pixel point in adjacent frames. The motion trajectory of the dynamic change area is generated by the motion direction and speed information of consecutive frames. The change law of the motion direction, speed and trajectory in the dynamic change area is analyzed to extract the motion mode of the area. For example, the detection of periodic changes can be determined as swinging, and the linear motion may be the movement of an object. The motion features of the extracted dynamic change area are compared and matched with the known dynamic scene types in the scene detection library through cross-validation. If the motion features of the dynamic change area have the highest matching degree with the features of a known dynamic scene type, the dynamic change area is identified as a dynamic scene type.

[0049] S132: Obtaining features of a static scene by performing the static feature analysis on the static area, and comparing the features of the static scene with known static scene types in the preset scene detection library to identify the static scene type.

[0050] In this embodiment, the features of a static scene refer to characteristics that can describe the spatial distribution, texture pattern, color information, etc. of a static area in an image. These features help distinguish and identify different types of static scenes, such as landscapes, buildings, and indoor scenes.

[0051] Specifically, feature extraction is performed on the image blocks in the static area, and texture features, spatial distribution features, color features, and frequency domain features are calculated respectively. The extracted static scene features are compared and matched with the features of known static scene types in the preset scene detection library. The nearest neighbor matching method is used for feature matching to obtain matching results. Based on the matching results, the static scene type corresponding to the current static area is identified and marked as a static scene type.

[0052] Furthermore, the gray-level co-occurrence matrix extracts the texture features of the image, such as contrast, uniformity, and entropy, by calculating the co-occurrence frequency between different gray values. Texture features distinguish static scenes with different texture types, such as rough rock surfaces and smooth walls. LBP is an operator used for texture analysis that generates a binary pattern code by comparing the gray values ​​of the central pixel with the surrounding pixels. This feature can effectively capture the texture information of static areas and is used to identify features such as walls and floors. In static scenes, the density of edges can reflect the complexity of the scene. Through methods such as the Sobel operator or Canny edge detection, the edge information in the image is extracted and the edge density is calculated to distinguish scenes with clear building outlines from natural scenery. The geometric shape features of objects in the image are extracted through shape descriptors to identify scenes with obvious geometric shapes, such as buildings and bridges. A color histogram is generated by counting the distribution of pixels of different colors in the image. Color features can be used to identify the main color tone and color distribution pattern of the scene, such as scenes with blue sky, white clouds, and green trees. The frequency components of the image are analyzed by converting the image from the spatial domain to the frequency domain. Static scenes usually have the characteristics of being dominated by low-frequency components. These frequency domain features can be extracted using Fourier transform to identify static scenes with stability and regularity.

[0053] In one embodiment, if Figure 4 As shown, in step S20, that is, according to the scene type in the TV screen, based on the preset optimization parameters, the brightness, contrast and sharpness of the TV image are dynamically adjusted to obtain the processed TV image, including: S21: According to the dynamic scene type, based on the preset optimization parameters, the brightness of the area where the moving object is located is increased; according to the static scene type, based on the preset optimization parameters, the contrast of the area where the moving object is located is maintained; according to the static scene type, based on the preset optimization parameters, the area where the moving object is located is sharpened.

[0054] Specifically, for dynamic scene types, based on the preset optimization parameters, the brightness of the area where the moving object is located is enhanced. This is achieved by adjusting the brightness value in the RGB channel to make the moving object more prominent in the picture. The brightness adjustment algorithm can adopt a linear gain method, that is, according to the preset brightness gain parameters, the pixel values ​​in the area are linearly amplified. According to the preset optimization parameters, the contrast of the area where the moving object is located is kept unchanged. In order to avoid the contrast imbalance caused by brightness adjustment, a contrast adjustment algorithm is used, and an adaptive contrast enhancement algorithm is used to maintain or slightly increase the contrast in the area, so that the moving object area can maintain a clear outline and layering while enhancing the brightness. After the brightness and contrast adjustments are completed, the area where the moving object is located is sharpened. Sharpening is to improve the clarity of the picture by enhancing the high-frequency components (i.e., edges and details) in the image. The sharpening algorithm can adopt the Laplace algorithm.

[0055] Furthermore, RGB channels refer to the three basic color channels used to represent the color of digital images, namely red, green, and blue. In each pixel, the color is a combination of the different intensities of these three channels. The value range of each channel is usually between 0 and 255, indicating the intensity of the color. For example, the pixel value of pure red is (255,0,0). By adjusting the RGB channels separately, the brightness, contrast, and color performance of the image can be changed. The preset brightness gain parameter refers to the gain coefficient applied to the brightness adjustment of the image during the image processing process. This parameter determines how the pixel value is amplified or reduced when the brightness is adjusted. The gain parameter is usually a value greater than 1, which is used to amplify the brightness and make the image brighter; if the gain parameter is less than 1, the brightness is reduced. The adaptive contrast enhancement algorithm refers to an image processing technology based on histogram equalization, which is used to enhance the contrast of the image. The traditional histogram equalization method evenly distributes the pixel values ​​across the entire image, while the adaptive contrast enhancement algorithm performs equalization processing separately according to the local area of ​​the image.

[0056] S22: According to the static scene type, based on the preset optimization parameters, the global brightness of the TV image is increased; according to the static scene type, based on the preset optimization parameters, the contrast between the dark and bright parts of the TV image is enhanced; according to the static scene type, based on the preset optimization parameters, the global sharpness of the TV image is increased.

[0057] Specifically, according to the detected static scene type, it is determined whether the global brightness of the image needs to be increased. For example, in a low-light scene, the image may be darker overall, and the brightness needs to be increased. The preset brightness gain parameter G is used. 亮度 , amplify the RGB value of each pixel in the image according to the formula: New brightness = original brightness × G 亮度 . Divide the image into several small areas, and perform histogram equalization on each small area separately. This will increase the contrast between dark and bright areas, making the details of the image more obvious. After processing each area, they are recombined into a complete image to ensure that the global contrast is enhanced. Apply the Laplacian operator edge detection algorithm to extract the edge information in the image. Then, the original image is fused with the edge information to generate a sharpened image. The implementation method is usually: sharpened image = original image + λ × edge information. Among them, λ is a sharpening coefficient used to control the degree of sharpening. By adjusting λ, the intensity of the sharpening effect can be dynamically controlled. The brightness, contrast and sharpening of the image are adjusted in turn to finally generate a processed TV image.

[0058] Furthermore, the Laplacian operator edge detection algorithm refers to highlighting edge information in an image by calculating the second-order derivative of pixel values ​​in the image.

[0059] In one embodiment, if Figure 5 As shown, in step S30, that is, the processed television image is subjected to noise reduction and detail enhancement to generate a final optimized television image, including: S31: Preliminarily removing noise points in the processed television image through an image noise reduction algorithm to obtain a noise-reduced image.

[0060] In this embodiment, the image noise reduction algorithm refers to an algorithm that improves the clarity and quality of an image by processing noise points in a television image.

[0061] Specifically, a two-dimensional Gaussian kernel matrix is ​​generated according to the specified standard deviation σ. The kernel size is usually selected as an odd size such as 3x3, 5x5 or 7x7, and the standard deviation determines the distribution of weights in the kernel. Each element value of the kernel matrix is ​​calculated by the Gaussian function: Where x and y are the horizontal and vertical distances from the center of the kernel. Convolve the Gaussian kernel with the processed TV image. Convolution is achieved by moving the kernel matrix point by point on the image and taking a weighted average of the pixels in the area covered by the kernel matrix. For each pixel, the weighted sum of the pixels in its neighborhood is calculated, and the result is used as the new value of the pixel, thereby achieving the effect of smoothing the image and reducing noise. At the edge of the image, filling processing is usually required due to the lack of sufficient neighborhood pixels. Common filling methods include zero filling or mirror filling to ensure that the Gaussian filter can cover the entire image. After Gaussian filtering, the noise in the image is effectively smoothed to obtain a denoised image. S32: Performing texture and edge feature analysis on the denoised image to obtain an image analysis result, and enhancing details of the denoised image based on the image analysis result to generate a detail-enhanced image.

[0062] In this embodiment, texture and edge feature analysis refers to extracting texture details and edge information in the image by analyzing the local area of ​​each pixel in the denoised image through an algorithm. Texture features mainly reflect the surface details and repetitive patterns of the image, while edge features represent the contours and mutation areas of objects in the image.

[0063] Specifically, each pixel is compared with the surrounding pixels. If the surrounding pixel value is greater than or equal to the central pixel value, it is marked as 1, otherwise it is marked as 0. Then, these binarization results are combined into a binary number as the texture feature of the pixel. The edge detection algorithm calculates the gradient change of the grayscale value of the image to identify the area where the grayscale value changes dramatically in the image, that is, the edge. Through the texture and edge feature analysis method, an image analysis result is generated to describe the texture density, edge strength and other characteristics of each part of the image. Using the image analysis results, adjust the local contrast, sharpness and other parameters of the denoised image to enhance the details of the image. In areas with smoother textures, the denoising effect is maintained to avoid noise recovery, while in areas with more obvious texture and edge features, the contrast and sharpness are appropriately increased to enhance the clarity and details of the image.

[0064] S33: performing edge-preserving filtering processing on the detail-enhanced image to obtain a sharpened image, and performing rendering processing on the sharpened image to generate the final optimized television image.

[0065] Specifically, edge-preserving filtering refers to an image processing technology that tries to preserve edge details of objects in an image while smoothing or reducing noise on the image.

[0066] Specifically, the edge and details of the image are retained by performing edge-preserving filtering on the detail-enhanced image. Subsequently, the overall clarity and visual effect of the image are further improved by rendering the sharpened image, and finally an optimized image suitable for TV display is generated.

[0067] In one embodiment, if Figure 6 As shown, after step S10, the TV image processing method based on scene recognition further includes: S40: Real-time detection of the brightness and color temperature changes of the ambient light to obtain ambient light information. In this embodiment, the ambient light information refers to the brightness and color temperature parameters of the ambient light detected in real time by the ambient light sensor.

[0068] Specifically, an ambient light sensor is installed on the front panel or around the TV. The light sensor can detect the brightness and color temperature changes of the current ambient light in real time. Brightness can be expressed in terms of light intensity (such as lux), while color temperature is expressed in Kelvin (K). The light sensor transmits the detected ambient light brightness and color temperature data to the image processing system. The image processing system analyzes the ambient light brightness and color temperature data to extract the current ambient light information.

[0069] S50: Based on the ambient light information, adjust the final optimized television image to compensate for the color cast caused by the ambient light change.

[0070] Specifically, the global brightness of the TV image is adjusted according to the brightness value of the ambient light. For example, when the ambient light is strong, the brightness of the image is increased to maintain the consistency of the display effect. The color temperature of the TV image is adjusted according to the color temperature of the ambient light. When the ambient light color temperature is warm, the warm tones of the image are reduced, and when the ambient light color temperature is cold, the warm tones of the image are increased. The contrast and color of the image are adjusted to compensate for the impact of changes in ambient light. Ensure that the image still maintains a natural and realistic visual effect under different ambient light conditions.

[0071] In one embodiment, if Figure 7 As shown, after step S10, the TV image processing method based on scene recognition further includes: S60: constructing a user preference model according to the user's operation information for adjusting the television screen image.

[0072] In this embodiment, the user preference model refers to a model established based on the user's adjustment operation information on the TV screen image, which is used to describe and predict the user's preference for image parameters (such as brightness, contrast, color temperature, etc.).

[0073] Specifically, record various adjustment operations performed by users on the TV screen image, such as manual adjustment of brightness, contrast, and color temperature. Record the time and ambient light conditions of each adjustment operation to analyze user preferences at different times and conditions. Extract the specific parameter values ​​adjusted by the user (for example, how much the brightness has increased, how much the contrast has decreased). Extract contextual information related to the user adjustment, such as the current program type, ambient light conditions, viewing time, etc. Analyze the collected data using statistical analysis to identify the user's adjustment patterns and preference trends. Train the user preference model by inputting the extracted features and the user's adjustment records into the user preference model.

[0074] S70: Automatically adjusting image processing parameters by analyzing user preferences for different scenes based on the user preference model.

[0075] In this embodiment, the image processing parameters refer to various image setting values ​​for optimizing and adjusting the display effect of a television picture.

[0076] Specifically, during the operation of the TV, the user's operating behavior is monitored in real time, especially the adjustment behavior in the image setting menu. When the user manually adjusts parameters such as brightness, contrast and color temperature, the system records the current scene type and the adjusted parameter values. All adjustment behaviors and scene information are recorded in the scene information database, and the system regularly analyzes these data to extract common user preference patterns. The system can identify the user's preferred parameters in a specific scene, and these preferences are modeled as part of the user preference model. When the user enters the same or similar scene again, the system calls the user preference model based on the current scene type to predict the user's possible parameter requirements. The system automatically applies these predicted parameters and adjusts the picture settings in real time. After each automatic adjustment, if the user manually adjusts the parameters again, the system will record these new adjustment behaviors and update the user preference model so that the model more accurately reflects the user's latest preferences.

[0077] Furthermore, the scene information database refers to a database used to store and manage various scene information and feature information related to television images.

[0078] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0079] In one embodiment, a TV image processing device based on scene recognition is provided, and the TV image processing device based on scene recognition corresponds to the TV image processing method based on scene recognition in the above embodiment. Figure 8As shown, the TV image processing device based on scene recognition includes a scene type acquisition module, a TV image processing module and a TV image generation module. The detailed description of each functional module is as follows: A scene type acquisition module, used to acquire the scene type in the TV picture, wherein the scene type includes a dynamic scene type and a static scene type; A television image processing module, used to dynamically adjust the brightness, contrast and sharpness of the television image according to the scene type in the television picture based on preset optimization parameters to obtain a processed television image; The television image generation module is used to perform noise reduction and detail enhancement on the processed television image to generate a final optimized television image.

[0080] Optionally, the scenario type acquisition module includes: A preliminary scene feature submodule is used to obtain preliminary scene features by performing real-time feature analysis on the color, brightness and texture of the TV screen image; An image classification submodule, for classifying the television screen image based on the preliminary scene features, and identifying dynamically changing areas and static areas in the television screen image; The scene type recognition module is used to perform motion feature analysis on the dynamically changing area to identify the dynamic scene type, and to perform static feature analysis on the static area to identify the static scene type.

[0081] Optionally, the scene type recognition module includes: A dynamic scene type unit, configured to obtain a motion feature of the dynamically changing region by performing the motion feature analysis on the dynamically changing region, and compare the motion feature of the dynamically changing region with known dynamic scene types in a preset scene detection library to identify the dynamic scene type; The static scene type unit is used to obtain the characteristics of the static scene by performing the static feature analysis on the static area, and compare the characteristics of the static scene with the known static scene types in the preset scene detection library to identify the static scene type.

[0082] Optionally, the television image processing module includes: A dynamic scene type adjustment submodule, configured to increase the brightness of the region where the moving object is located based on the preset optimization parameters according to the dynamic scene type, maintain the contrast of the region where the moving object is located based on the preset optimization parameters according to the static scene type, and perform sharpening processing on the region where the moving object is located based on the preset optimization parameters according to the static scene type; The static scene type adjustment submodule is used to increase the global brightness of the television image based on the preset optimization parameters according to the static scene type, enhance the contrast between the dark and bright parts of the television image based on the preset optimization parameters according to the static scene type, and increase the global sharpness of the television image based on the preset optimization parameters according to the static scene type.

[0083] Optionally, the television image generation module includes: A noise reduction image submodule, used to preliminarily remove noise points in the processed television image through an image noise reduction algorithm to obtain a noise reduction image; A detail enhanced image submodule is used to perform texture and edge feature analysis on the denoised image to obtain an image analysis result, and enhance the details of the denoised image based on the image analysis result to generate a detail enhanced image; The final optimized television image submodule is used to perform edge-preserving filtering on the detail-enhanced image to obtain a sharpened image, and render the sharpened image to generate the final optimized television image.

[0084] Optionally, the television image generation module then includes: An ambient light detection module is used to detect the brightness and color temperature changes of the ambient light in real time to obtain ambient light information; The television image optimization module is used to adjust the final optimized television image based on the ambient light information to compensate for the color cast caused by the ambient light change.

[0085] Optionally, the television image generation module then includes: A model building module, used to build a user preference model according to the operation information of the user adjusting the television screen image; The parameter adjustment module is used to automatically adjust the image processing parameters by analyzing the user's preferences for different scenes based on the user preference model.

[0086] For the specific definition of the TV television image processing device based on scene recognition, please refer to the definition of the TV television image processing method based on scene recognition above, which will not be repeated here. The various modules in the above-mentioned TV television image processing device based on scene recognition can be implemented in whole or in part through software, hardware and combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0087] In one embodiment, a computer device is provided. The computer device may be a server, and the internal structure diagram may be as follows: Fig. 9As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. In, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used for a scene information database. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a TV image processing method based on scene recognition is implemented.

[0088] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: Acquire the scene type in the TV screen, where the scene type includes a dynamic scene type and a static scene type; According to the scene type in the TV picture, based on the preset optimization parameters, the brightness, contrast and sharpness of the TV image are dynamically adjusted to obtain a processed TV image; Noise reduction and detail enhancement are performed on the processed television image to produce a final optimized television image.

[0089] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: Acquire the scene type in the TV screen, where the scene type includes a dynamic scene type and a static scene type; According to the scene type in the TV picture, based on the preset optimization parameters, the brightness, contrast and sharpness of the TV image are dynamically adjusted to obtain a processed TV image; Noise reduction and detail enhancement are performed on the processed television image to produce a final optimized television image.

[0090] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. In the embodiments provided in this application, any reference to memory, storage, database or other media used may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0091] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0092] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features can be replaced by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A TV image processing method based on scene recognition, characterized in that: The TV image processing method based on scene recognition comprises: Acquire a scene type in a television picture, wherein the scene type includes a dynamic scene type and a static scene type; According to the scene type in the TV picture, based on preset optimization parameters, dynamically adjust the brightness, contrast and sharpness of the TV image to obtain a processed TV image; The processed television image is subjected to noise reduction and detail enhancement to generate a final optimized television image.

2. The TV image processing method based on scene recognition according to claim 1, characterized in that: The acquiring of the scene type in the television picture, wherein the scene type includes a dynamic scene type and a static scene type, comprises: By performing real-time feature analysis on the color, brightness and texture of the TV screen image, preliminary scene features are obtained; Based on the preliminary scene features, the television screen image is classified to identify the dynamically changing area and the static area in the television screen image; A motion feature analysis is performed on the dynamically changing area to identify the dynamic scene type, and a static feature analysis is performed on the static area to identify the static scene type.

3. The TV image processing method based on scene recognition according to claim 2, characterized in that: The performing motion feature analysis on the dynamically changing area to identify the dynamic scene type, and performing static feature analysis on the static area to identify the static scene type, comprises: By performing the motion feature analysis on the dynamically changing area, the motion feature of the dynamically changing area is obtained, and the motion feature of the dynamically changing area is compared with known dynamic scene types in a preset scene detection library to identify the dynamic scene type; The static feature analysis is performed on the static area to obtain the features of the static scene, and the features of the static scene are compared with the known static scene types in the preset scene detection library to identify the static scene type.

4. The TV image processing method based on scene recognition according to claim 1, characterized in that: The step of dynamically adjusting the brightness, contrast and sharpness of the television image according to the scene type in the television screen and based on preset optimization parameters to obtain a processed television image comprises: According to the dynamic scene type, based on the preset optimization parameters, the brightness of the area where the moving object is located is increased; according to the static scene type, based on the preset optimization parameters, the contrast of the area where the moving object is located is maintained; according to the static scene type, based on the preset optimization parameters, the area where the moving object is located is sharpened; According to the static scene type, based on the preset optimization parameters, the global brightness of the TV image is increased; according to the static scene type, based on the preset optimization parameters, the contrast between the dark and bright parts of the TV image is enhanced; according to the static scene type, based on the preset optimization parameters, the global sharpness of the TV image is increased.

5. The TV image processing method based on scene recognition according to claim 1, characterized in that: The step of reducing noise and enhancing details of the processed television image to generate a final optimized television image comprises: Preliminarily removing noise points in the processed television image by an image noise reduction algorithm to obtain a noise-reduced image; Performing texture and edge feature analysis on the denoised image to obtain an image analysis result, and enhancing details of the denoised image based on the image analysis result to generate a detail-enhanced image; The detail-enhanced image is subjected to edge-preserving filtering to obtain a sharpened image, and the sharpened image is subjected to rendering to generate the final optimized television image.

6. The TV image processing method based on scene recognition according to claim 1, characterized in that: The TV image processing method based on scene recognition further includes: Real-time detection of changes in brightness and color temperature of the ambient light to obtain ambient light information; Based on the ambient light information, the final optimized television image is adjusted to compensate for the color cast caused by the ambient light change.

7. The TV image processing method based on scene recognition according to claim 1, characterized in that: The TV image processing method based on scene recognition further includes: constructing a user preference model according to the operation information of the user adjusting the television screen image; Based on the user preference model, image processing parameters are automatically adjusted by analyzing the user's preferences for different scenes.

8. A TV image processing device based on scene recognition, characterized in that: The TV image processing device based on scene recognition comprises: A scene type acquisition module, used to acquire the scene type in the TV picture, wherein the scene type includes a dynamic scene type and a static scene type; A television image processing module, used to dynamically adjust the brightness, contrast and sharpness of the television image according to the scene type in the television picture based on preset optimization parameters to obtain a processed television image; The television image generation module is used to perform noise reduction and detail enhancement on the processed television image to generate a final optimized television image.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the TV image processing method based on scene recognition as claimed in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the TV image processing method based on scene recognition as claimed in any one of claims 1 to 7 are implemented.

Images