Method, system and device for enhancing image details in real time

Abstract

The invention aims to provide a method, a system and a device for enhancing image details in real time. The method comprises the following steps: splitting an image into a plurality of pixels; judging whether the pixel is a noisy point or not; if the pixel is not the noisy point, judging whether the pixel is located at the edge or not; and if the pixel is located at the edge, performing USM enhancement on the pixel. According to the method, optimization is specially carried out aiming at the USM problem, pixel data can be efficiently processed, so that real-time classification operation is carried out on the pixels, the pixels which do not need to be filtered are removed, various distortion expressions are avoided, and a satisfactory filtering effect can be realized by using a smaller filter core.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly to a method, system and device for real-time enhancing image details. Background Art

[0002] Enhancing image details is a commonly used image processing means. Generally speaking, complex image algorithms can obtain relatively ideal processing results, but the price is slow processing speed and complex operations, making it impossible to be implemented in hardware; simple algorithms have problems such as poor image enhancement effect and easy image distortion. Another challenge is real-time processing. Real-time processing requires the processing speed to keep up with the image transmission speed. With the continuous popularization of high-resolution content, the current image transmission speed has exceeded 1G pixels / s calculated by pixel, and it is very difficult to implement a device that can run at such a speed, and there are few available implementation options. Currently, the relatively ideal implementation solution in terms of effect and speed is to use USM (unsharp mask) filtering to enhance image details. However, as mentioned above, unsharp mask will introduce image distortion, resulting in an obvious decline in image perception and effect in some cases, leading to limitations in the implementation of the solution. The USM algorithm (Unsharp Masking) is an edge enhancement technology for image processing, mainly used to enhance the detail contrast of images and improve image clarity. The USM algorithm generates an "unsharp mask" by creating a low-frequency version of the image and subtracting it from the original image, and then combines this mask with the original image to strengthen edges and details. The USM algorithm is widely used in image editing software such as Adobe Photoshop and is a standard step in many digital photography post-processing processes. By adjusting the sharpening intensity and radius parameters, the degree of sharpening and the affected range can be controlled. The USM algorithm can enhance the edges and details of images: by highlighting the high-frequency components of the image, making the image look clearer. Reduce noise: reduce the impact of noise while sharpening. Adjust the sharpening coefficient: by adjusting the coefficient λ, the degree of sharpening can be controlled to avoid the "overshoot" phenomenon caused by oversharpening.

[0003] The effect of USM depends on the size of the filter kernel (kernel) it uses. Increasing the size of the kernel can improve the filtering effect, but the computational complexity and system complexity will both increase rapidly, seriously affecting real-time performance. In addition, USM has its inherent drawbacks. It will show false colors in the texture area and color inversion in the edge area, which are all caused by the global enhancement of the image by USM rather than selective enhancement. Summary of the Invention

[0004] The object of the present invention is to provide a method, system and device for real-time enhancing image details. This method is specifically optimized for the problems of USM, can efficiently process pixel data, thereby realizing real-time classification operations on pixels, eliminating pixels that do not need to enter the filtering, avoiding various distortion manifestations, and thus can use a smaller filter core to achieve a satisfactory filtering effect.

[0005] A method for real-time enhancing image details, comprising: Splitting an image into multiple pixels; Judging whether the pixel is a noise point; If the pixel is not a noise point, judging whether the pixel is at an edge; If the pixel is at an edge, performing USM enhancement on the pixel.

[0006] Preferably, the splitting the image into multiple pixels includes: Storing three rows of input pixels to obtain a three-row pixel cache; Using the three-row pixel cache to obtain a 3x3 pixel matrix; Calculating the correlation matrix of the 3x3 pixel matrix.

[0007] Preferably, the judging whether the pixel is a noise point includes: Judging whether the relationship between the 3x3 pixel matrix and the correlation matrix of the 3x3 pixel matrix exceeds a threshold set by the system; If the judgment of noise is established, the current pixel is directly output.

[0008] Preferably, the if the pixel is not a noise point, judging whether the pixel is at an edge includes: Multiplying by unit matrices in two directions. If the obtained result is greater than the threshold set by the system, then there is an edge in the 3x3 pixel matrix, otherwise there is no edge.

[0009] Preferably, the if the pixel is not a noise point, judging whether the pixel is at an edge further includes: Judging whether the current pixel has passed through an edge; Multiplying a unit matrix by the 3x3 pixel matrix, and comparing the maximum value of the obtained result with the result of multiplying by unit matrices in two directions; If it is greater than a set ratio, it is considered that the current pixel is on the edge.

[0010] Preferably, the if the pixel is at an edge, performing USM enhancement on the pixel includes: Introducing a bilateral filter and a Gaussian filter in USM enhancement to process the image.

[0011] Preferably, if the pixel is at the edge, the USM enhancement of the pixel includes: Calculating the mean square deviation of a 3x3 pixel matrix; Normalizing the mean square deviation; Selecting the gain intensity of the pixel according to the normalized variance of the pixel.

[0012] A system for real-time enhancing image details includes: An image processing module for splitting an image into multiple pixels; A first judgment module for judging whether the pixel is a noise point; A second judgment module for judging whether the pixel is at the edge if the pixel is not a noise point; An image enhancement module for performing USM enhancement on the pixel if the pixel is at the edge.

[0013] A method device for real-time enhancing image details includes: An image processing device for splitting an image into multiple pixels; A first judgment device for judging whether the pixel is a noise point; A second judgment device for judging whether the pixel is at the edge if the pixel is not a noise point; An image enhancement device for performing USM enhancement on the pixel if the pixel is at the edge.

[0014] An electronic device includes: a chip, a processor, and a memory. The memory is used to store computer program code, and the computer program code includes computer instructions. When the chip executes the computer instructions, the electronic device executes a method for real-time enhancing image details.

[0015] The beneficial effects of the present invention are as follows: 1. The present invention uses surrounding pixels to complete the construction of two reference parameter matrices through simple addition, multiplication, and shifting, and constrains all subsequent processing based on these two matrices. 2. The present invention makes noise judgment, edge judgment, and judgment on the edge according to the two reference matrices in accordance with the steps, and outputs the judgment result of the current pixel. 3. If the judgment is true, both the judgment result and the reference matrix are output to the subsequent filtering process. If it is false, the original pixel is directly output without performing filtering enhancement operations. 4. The present invention improves the USM algorithm and combines the USM algorithm with other image processing algorithms to obtain better processing results. 5. The present invention improves the adaptive function of USM, can adopt different enhancement coefficients according to different regions of image pixels, and has a better image enhancement effect. Description of the Drawings

[0016] The accompanying drawings here are incorporated into the specification and form a part of this specification, showing the embodiments that conform to the present invention, and are used together with the specification to explain the principles of the present invention.

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0018] Figure 1 It is a flowchart of a method for real-time enhancing image details of the present invention; Figure 2 It is a schematic diagram of the USM enhancement effect of the present invention; Figure 3 It is a schematic diagram of four unit matrices of the present invention; Figure 4 It is a schematic diagram of the hardware structure of an electronic device of the present invention. Detailed implementation manners

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0020] It should be noted that all the directional indications (such as up, down, left, right, front, back...) in the embodiments of the present invention are only used to explain the relative position relationship and movement conditions between components in a specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indications will also change accordingly.

[0021] In addition, the descriptions involving "first", "second", etc. in the present invention are only for descriptive purposes, and cannot be understood as indicating or implying their relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the fact that those of ordinary skill in the art can implement it. When the combination of technical solutions results in contradictions or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0022] The effect of USM depends on the size of the filter kernel it uses. Increasing the size of the kernel can improve the filtering effect, but both the computational complexity and system complexity will increase rapidly, seriously affecting real-time performance. Additionally, USM has its inherent drawbacks. In texture regions, it will exhibit false colors, and in edge regions, there will be color inversion phenomena. This is because USM performs global enhancement on the image rather than selective enhancement.

[0023] The present invention constructs two reference parameter matrices by using surrounding pixels through simple addition, multiplication, and shifting, and constrains all subsequent processing to be based on these two matrices. According to the two reference matrices, the present invention performs operations such as noise judgment, edge judgment, and judgment on the edge in sequence, and outputs the judgment result of the current pixel. If the judgment is true, both the judgment result and the reference matrix are output to the subsequent filtering process. If it is false, the original pixel is directly output without performing filtering enhancement operations. The present invention improves the USM algorithm and combines the USM algorithm with other image processing algorithms to obtain better processing results. The present invention improves the adaptive function of USM and can adopt different enhancement coefficients according to different regions of image pixels, having a better image enhancement effect.

[0024] Embodiment 1 A method for real-time enhancing image details, referring to Figure 1 , including: S100, splitting the image into multiple pixels; ‌The main function of splitting the image into multiple pixels is to simplify or change the representation form of the image, making it easier to understand and analyze.‌ The present invention subdivides the digital image into multiple image sub-regions (sets of pixels), making the image easier to understand and analyze, for locating objects and boundaries (lines, curves, etc.) in the image. And a label is added to each pixel in the image, so that pixels with the same label have certain common visual characteristics, such as color, brightness, texture, etc.‌

[0025] S200, judging whether the pixel is a noise point; Noise appears as granular or speckled noisy areas in an image, which can cause certain damage to the details and clarity of the image. In digital photography, noise is usually caused by signal interference generated by the image sensor at high ISO sensitivities. To reduce the impact of noise, the present invention uses a noise reduction algorithm to post-process the image to improve the quality of the image. Noise is an unnecessary or interfering signal in the image, manifested as a miscellaneous signal incompatible with the background. In photography, the generation of noise is related to factors such as the ISO (sensitivity) setting, exposure time, and temperature. High ISO settings, long exposure times, and high temperatures will all increase the generation of noise. Noise will make the details of the image become blurred and reduce the quality of the image. In an image, the noise in the shadow area will be more obvious than that in the bright area because the light signal in the shadow area is weaker, and the noise will be more obvious when reading the signal.

[0026] S300, if the pixel is not noise, determine whether the pixel is at the edge; An edge is a place where the pixel values in an image change drastically, usually the boundary between an object and the background in the image or the boundary between different regions within an object. Edge detection is a basic operation in image processing, and its main purpose is to detect the edge information of an object in the image, that is, the contour of the object. The basic principle of an edge detection algorithm is to find the positions where the pixel values change significantly in the image, and these positions are usually the contours of the object.

[0027] S400, if the pixel is at the edge, then perform USM enhancement on the pixel.

[0028] The role of USM enhancement is to highlight the details and edges in the image, and enhance the clarity and visual effect of the image. USM sharpens the image by blurring the image (such as Gaussian blur) and then enhancing the difference (high-frequency details) between the original image and the blurred image. USM can highlight the details and edges in the image and enhance the clarity and visual effect of the image. It does not affect the brightness and contrast of the smooth areas, so as to improve the detail performance of the image while keeping the overall brightness and contrast of the image unchanged.

[0029] The purpose of the present invention is to design a hardware device that can process pixel data input in real time. The hardware device implemented by this patent can process the input pixel data in real time and determine whether each pixel should be sent to the USM algorithm unit for enhancement.

[0030] Preferably, S100, splitting the image into multiple pixels includes: S110, storing three rows of input pixels to obtain a three-row pixel cache; First, store three rows of input pixels to form a three-row pixel cache.

[0031] S120, obtain a 3x3 pixel matrix using a three-row pixel cache; Continuously input a 3x3 pixel matrix using a three-row cache, denoted as , and calculate another matrix through .

[0032] S130, calculate the correlation matrix of the 3x3 pixel matrix.

[0033] The calculation process is as follows: ; ; Preferably, in S200, determining whether a pixel is a noise point includes: S210, determine whether the relationship between the 3x3 pixel matrix and the correlation matrix of the 3x3 pixel matrix exceeds the threshold set by the system; S220, if it is determined that the noise point is established, the current pixel is directly output.

[0034] ; That is, determine whether the relationship between and exceeds the threshold set by the system. If it is determined that the noise point is established, then this pixel will be directly bypassed to the output, will not enter the subsequent operations, and will not enter the USM enhancement. If the noise point determination is not established, then the following operation process will be continued.

[0035] In the embodiment of the present invention, the function of determining the noise point is to improve the image quality and enhance the visual effect. Noise points refer to unwanted interference or distortion in the image, manifested as random granular spots, inconsistencies in brightness or color, etc. These noise points may be caused by various factors, including the natural noise of the camera sensor, high ISO settings, image capture under low light conditions, errors in the signal processing process, etc. The existence of noise points will reduce the details and clarity of the image. Therefore, in photo processing, measures such as denoising are usually taken to reduce or eliminate noise points.

[0036] Preferably, in S300, if the pixel is not a noise point, determine whether the pixel is on the edge, including: Multiply the unit matrices in two directions. If the obtained result is greater than the threshold set by the system, then there is an edge in the 3x3 pixel matrix; otherwise, there is no edge.

[0037] To determine whether there is an edge in this 3x3 matrix, multiply the unit matrices and in two directions. If the obtained result and If it is greater than the threshold set by the system, then there is an edge in this 3x3 matrix; otherwise, there is no edge.

[0038] ; ; If it is not established that there is an edge, the pixel will be directly output, will not enter the subsequent operations, and will not enter the USM enhancement.

[0039] The main purpose of edge detection is to detect the edge information of objects in an image, that is, the contour of the object. An edge is a place where the pixel values in the image change drastically, usually the boundary between the object and the background in the image or the boundary between different regions inside the object. By detecting the edges, the amount of calculation in subsequent processing can be reduced. Edge detection can provide the structural information of the objects in the image. The detected edges can reflect the shape and structure of the objects.

[0040] Preferably, referring to Figure 3 , if the pixel is not a noise point, determining whether the pixel is on the edge further includes: Determining whether the current pixel has passed through an edge; Multiplying the identity matrix by the 3x3 pixel matrix, and comparing the maximum value of the result with the result of multiplying by the identity matrix in two directions; If it is greater than the set ratio, it is considered that the current pixel is on the edge.

[0041] Nine pixels that may have edges have been filtered out through the previous operations, but for the current pixel Whether it needs to be sent to the subsequent enhancement for processing still requires a judgment on whether the edge passes through this pixel. The judgment process is to multiply the original matrix by four types of identity matrices, and take the maximum value of the results and that in the fourth step for comparison. If it is greater than the set ratio, it is considered that the pixel is on the edge.

[0042] ; ; If it is still judged to be true in this step, the pixel will be sent to the subsequent USM enhancement. Before the USM enhancement, it is necessary to first judge whether the pixel is on the edge. Only the pixels that are not on the edge can be used for the subsequent USM enhancement. Therefore, edge detection is used to judge the pixels and filter out the pixels that are not on the edge to enter the USM enhancement.

[0043] Preferably, before performing USM enhancement on the image, it is also necessary to adjust the different weight parameters of the pixels in the image according to different situations. Specifically: Adjust the Laplacian weight, expressed as: ; Under the transformation of , the data in the low-dimensional space can be mapped to the high-dimensional space. The Laplace weight, in image processing, refers to measuring the contrast of an image by calculating the difference between the color channels and the luminance channel (L channel) of each pixel after converting an RGB image into an Lab image.

[0044] Adjust the significance weight, expressed as: ; where is the average pixel value of the input image in the step of adjusting the Laplace weight; is the data value of the pixel after Gaussian blur processing. The image significance weight refers to, in image processing, assigning different weight values according to the importance or significance of each part in the image. These weight values reflect the contribution degree of different regions in the image to the overall visual effect, and are usually used to guide subsequent image processing tasks, such as object detection, image segmentation, image compression, etc. The significance weight can help the algorithm quickly locate the key objects in the image, guide the image segmentation algorithm, improve the accuracy and efficiency of segmentation, preferentially retain the regions with high significance during the compression process, improve the compression ratio while maintaining the image quality, and track the targets with high significance in video processing.

[0045] Adjust the exposure weight, expressed as: ; where represents the value of the pixel position (x, y) of the input image , k is the number of pixels in the image, and the standard deviation is 0.25. The image exposure weight refers to, in image processing, the weight vector corresponding to the exposure index and the color temperature zone. The exposure index is the flag bit of the exposure time and the gain size when the camera takes a picture, starting from 0. Generally, the higher the ambient brightness, the shorter the required exposure time, the smaller the gain, and the smaller the Index.

[0046] Normalize the above three weights, expressed as: ; where is a regularization term to ensure that each input contributes to the output, takes the value of 0.15.

[0047] In the embodiments of the present invention, the function of adjusting the image pixel weights is to optimize the display quality of the image, enhance specific features of the image, and perform image processing and restoration. In image processing, weight calculation is used for image enhancement and filtering. By adjusting the weights of pixels, a clearer and higher-contrast image can be generated. For example, the weighted average method can adjust the gray intensity of pixels by assigning different weights to pixels at different positions, thereby enhancing specific regions of the image and making these regions clearer and more contrasting. Weight calculation can also be used to remove noise in the image or weaken the influence of certain regions.

[0048] Preferably, in S400, if the pixel is at the edge, performing USM enhancement on the pixel includes: Introduce a bilateral filter and a Gaussian filter in the USM enhancement to process the image.

[0049] In the embodiments of the present invention, the bilateral filter and the Gaussian filter are used to improve the USM enhancement algorithm, and the specific process is expressed as: ; Wherein, is the processed image, is the process of bilateral filtering of the original pixel, is the process of Gaussian filtering of the original pixel, is the process of performing CLAHE processing on the base region of the image after bilateral filtering, is the adaptive function of the detail enhancement part.

[0050] The bilateral filter can preserve the edge information of the image while filtering out noise. This is because the bilateral filter combines the spatial proximity of the image and the similarity of pixel values for filtering, so that while smoothing the image and filtering out noise, it can effectively preserve the edge information of the image.

[0051] Subtract the image after bilateral filtering from the image after Gaussian filtering to obtain a mask, that is, the image detail layer. At the same time, perform contrast enhancement on the image after bilateral filtering through the CLHAE algorithm to obtain the result . Then multiply the image detail layer by the adaptive function weight , and superimpose it on the image after bilateral filtering and contrast enhancement to obtain the enhanced image

[0052] The process of bilateral filtering is expressed as: ; Wherein, is the normalization parameter; is the original image pixel point, j ∈ N represents The pixel region centered on is the spatial weight, and is the grayscale weight. The spatial weight decreases as the distance from the central pixel increases, and the grayscale weight decreases as the difference in grayscale values increases.

[0053] The filtering template of Gaussian filtering is expressed as: ; In the embodiments of the present invention, convolving the Gaussian function with the image can obtain the low-frequency information of the image. The main function of Gaussian filtering is to eliminate Gaussian noise in the image and improve the image quality through smoothing processing. Gaussian filtering is a linear smoothing filter, suitable for eliminating noise that follows a normal distribution, and is applied in the noise reduction process of image processing. Gaussian filtering smooths the image by applying the Gaussian function to each pixel in the image. The Gaussian function is a bell-shaped curve that reaches its maximum value at the center point and then gradually decreases towards both sides. By adjusting the parameters of the Gaussian function, the degree of smoothing can be controlled.

[0054] Preferably, referring to Figure 2 , S400, if the pixel is at the edge, performing USM enhancement on the pixel includes: S410, calculating the mean square error of the 3x3 pixel matrix; Take any pixel point x(m,n), and define a 3 × 3 pixel matrix A centered on x(m,n). Calculate the mean square error d(m,n) of the pixel matrix A, expressed as: ; where i and j take values from -1 to 1.

[0055] S420, performing normalization processing on the mean square error; Normalization processing can eliminate the influence of extreme values, enabling better generalization during the image processing process and obtaining better processing effects.

[0056] Normalize , expressed as: ; S430, selecting the gain intensity of the pixel according to the normalized variance of the pixel.

[0057] The adaptability of the gain function is reflected in using different gain intensities for different regions of the image. s(m,n) divides each pixel point into the interval of 0-255, and then determines the current gain coefficient according to the value of each pixel.

[0058] ; where is the gain coefficient, which can be adjusted according to different application scenarios.

[0059] The gain coefficient of the USM algorithm refers to the coefficient used to adjust the weighted average of the original image and the Gaussian blurred image when performing USM sharpening. The USM sharpening algorithm enhances image details and reduces noise by subtracting the Gaussian blur from the original image.

[0060] In the embodiments of the present invention, USM enhances image details and reduces noise by subtracting the Gaussian blur from the original image, providing a more realistic sharpening effect. This method can highlight high-frequency information in the image, such as edges and textures, thereby enhancing the clarity and visual effect of the image. And while sharpening the image, the USM algorithm can remove some small interfering details and noise, making the sharpening result more realistic and reliable. According to the present invention, the USM algorithm is improved so that it can be applied to various scenarios, including photography, printing, medical images, and remote sensing images, etc., which can significantly enhance the clarity and details of the image. Moreover, the implementation of the USM algorithm of the present invention is relatively simple, and it is easy to adjust and maintain in practical applications, suitable for various image processing requirements, and can be implemented without too high hardware requirements, which can save a lot of production costs.

[0061] Embodiment 2 A system for real-time enhancing image details, comprising: An image processing module, configured to split an image into multiple pixels; A first judgment module, configured to judge whether a pixel is a noise point; A second judgment module, configured to judge whether a pixel is at an edge if the pixel is not a noise point; An image enhancement module, configured to perform USM enhancement on a pixel if the pixel is at an edge.

[0062] Embodiment 3 A method and apparatus for real-time enhancing image details, comprising: An image processing apparatus, configured to split an image into multiple pixels; A first judgment apparatus, configured to judge whether a pixel is a noise point; A second judgment apparatus, configured to judge whether a pixel is at an edge if the pixel is not a noise point; An image enhancement apparatus, configured to perform USM enhancement on a pixel if the pixel is at an edge.

[0063] The prior art is not suitable for hardware implementation. The method of this application is actually a hardware computing device carried by very large scale integration (VLSI). The computing intensity and computing conditions of other solutions are limited and cannot be implemented on a circuit. Since the method of this application is implemented using hardware, the processing speed is very fast, and real-time processing of high-resolution video can be achieved, while a large number of similar solutions can only process static pictures. Moreover, the solution of this application has very small resource overhead and has a great advantage in implementation cost.

[0064] Embodiment 4 An electronic device includes: a chip, a processor, and a memory. The memory is used to store computer program code, and the computer program code includes computer instructions. When the chip executes the computer instructions, the electronic device executes a method for real-time enhancing image details.

[0065] Reference Figure 4 , the electronic device 2 includes a processor 21, a memory 22, an input device 23, and an output device 24. The processor 21, the memory 22, the input device 23, and the output device 24 are coupled through a connector, and the connector includes various interfaces, transmission lines, or buses, etc., which are not limited in the embodiments of the present invention. It should be understood that in various embodiments of the present invention, coupling means being interconnected in a specific manner, including being directly connected or indirectly connected through other devices. For example, they can be connected through various interfaces, transmission lines, buses, etc.

[0066] The processor 21 can be one or more graphics processing units (GPUs). When the processor 21 is a single GPU, the GPU can be a single-core GPU or a multi-core GPU. Optionally, the processor 21 can be a processor group composed of multiple GPUs, and multiple processors are coupled to each other through one or more buses. Optionally, the processor can also be other types of processors, etc., which are not limited in the embodiments of the present invention.

[0067] The memory 22 can be used to store computer program instructions and various computer program codes including the program code for implementing the solution of the present invention. Optionally, the memory includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or compact disc read-only memory (CD-ROM), and the memory is used for relevant instructions and data.

[0068] An input device 23 is used to input data and / or signals, and an output device 24 is used to output data and / or signals. The output device 24 and the input device 23 can be independent devices or an integrated device.

[0069] The present invention utilizes surrounding pixels to complete the construction of two reference parameter matrices through simple addition, multiplication, and shifting, and constrains all subsequent processing to be based on these two matrices. According to the two reference matrices, the present invention performs noise judgment, edge judgment, and judgment on the edge in sequence and outputs the judgment result of the current pixel. If the judgment is true, both the judgment result and the reference matrix are output to the subsequent filtering process; if it is false, the original pixel is directly output without performing filtering enhancement operations. The present invention improves the USM algorithm and combines the USM algorithm with other image processing algorithms to obtain better processing results. The present invention improves the adaptive function of USM and can adopt different enhancement coefficients according to different regions of image pixels, having a better image enhancement effect.

[0070] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather will conform to the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for real-time enhancement of image details, characterized in that: include: Split the image into pixels; Determining whether the pixel is a noise point; If the pixel is not a noise point, determining whether the pixel is at an edge; If the pixel is at the edge, USM enhancement is performed on the pixel.

2. The method for real-time image detail enhancement according to claim 1, characterized in that: The splitting of the image into a plurality of pixels comprises: Store three rows of input pixels to obtain a three-row pixel buffer; Using the three rows of pixel cache, a 3x3 pixel matrix is ​​obtained; Calculate the incidence matrix of the 3x3 pixel matrix.

3. The method for real-time image detail enhancement according to claim 2, characterized in that: The determining whether the pixel is a noise point comprises: Determine whether the relationship between the 3x3 pixel matrix and the correlation matrix of the 3x3 pixel matrix exceeds a threshold set by the system; If the noise is determined to be true, the current pixel is directly output.

4. The method for real-time image detail enhancement according to claim 2, characterized in that: If the pixel is not a noise point, determining whether the pixel is at an edge includes: The unit matrices in two directions are multiplied. If the result is greater than a threshold set by the system, then there is an edge in the 3x3 pixel matrix; otherwise, there is no edge.

5. The method for real-time image detail enhancement according to claim 4, characterized in that: If the pixel is not a noise point, determining whether the pixel is at an edge further includes: Determine whether the current pixel passes the edge; Multiply the unit matrix by the 3x3 pixel matrix, take the maximum value of the result and compare it with the result of multiplying the unit matrix in two directions; If it is greater than the set ratio, the current pixel is considered to be on the edge.

6. The method for real-time image detail enhancement according to claim 1, characterized in that: If the pixel is at the edge, performing USM enhancement on the pixel comprises: In USM enhancement, bilateral filter and Gaussian filter are introduced to process the image.

7. The method for real-time image detail enhancement according to claim 1, characterized in that: If the pixel is at the edge, performing USM enhancement on the pixel comprises: Calculate the mean square error of the 3x3 pixel matrix; Performing standardization on the mean square error; The gain intensity of a pixel is chosen based on its normalized variance.

8. A system for enhancing image details in real time, characterized in that: include: An image processing module for splitting an image into multiple pixels; A first judging module, used to judge whether the pixel is a noise point; A second judgment module, configured to judge whether the pixel is at an edge if the pixel is not a noise point; The image enhancement module is used for performing USM enhancement on the pixel if the pixel is at the edge.

9. A method and device for real-time enhancement of image details, characterized in that: include: An image processing device for splitting an image into a plurality of pixels; A first judging device, used to judge whether the pixel is a noise point; A second judging device, for judging whether the pixel is at an edge if the pixel is not a noise point; The image enhancement device is used for performing USM enhancement on the pixel if the pixel is at the edge.

10. An electronic device, characterized in that: include: A chip, a processor and a memory, wherein the memory is used to store computer program codes, wherein the computer program codes include computer instructions, and when the chip executes the computer instructions, the electronic device executes a method for real-time enhancement of image details as described in any one of claims 1 to 7.

Images