An image processing system based on artificial intelligence

Abstract

The application discloses an image processing system based on artificial intelligence, relates to the technical field of image processing, and utilizes an image input module to input an original image into the system, the system receives original image data and performs image display by using a display module, adjusts brightness and contrast, suppresses noise and enhances edges, and performs histogram equalization based on the original image data, so that the image quality displayed on the display module is gradually improved, the application overcomes the defects of traditional methods, realizes more intelligent, efficient and high-quality image processing effect, and especially when processing complex and changeable images, the system can automatically analyze image content and dynamically adjust processing parameters to achieve the best processing effect, the image processing system based on artificial intelligence brings a new idea and method for the image processing field, and has wide application prospect and important research value.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to an image processing system based on artificial intelligence. Background Technology

[0002] Computer vision, as an important branch of artificial intelligence, aims to enable computer systems to simulate the human visual system, thereby understanding and processing image and video data. With the continuous development of computer technology, computer vision has gradually merged with image processing technology to form an AI-based image processing system. This system can automatically analyze, understand and process image data, providing strong support for various application scenarios.

[0003] Traditional image processing algorithms typically use fixed parameter values, failing to adapt to the specific content of the image, resulting in unsatisfactory processing effects. Furthermore, traditional methods often struggle to strike a balance between noise suppression and edge enhancement, either resulting in over-smoothing that blurs edges or insufficient noise suppression. In addition, in traditional image processing, especially when dealing with images with extreme brightness and contrast ranges, while histogram equalization can improve image contrast, it can lead to loss of image details and over-enhancement. Finally, traditional methods lack a feedback mechanism during processing, failing to dynamically adjust processing parameters based on the results to achieve optimal processing effects. Summary of the Invention

[0004] The purpose of this invention is to provide an image processing system based on artificial intelligence, which solves the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution, and the specific implementation steps are as follows:

[0006] Users input raw images into the system using the image input module. The system receives the raw image data and displays the image using the display module.

[0007] Based on the original image data, and using the image adjustment processing module, brightness and contrast are adjusted, noise is suppressed and edge is enhanced, and histogram equalization is performed to gradually improve the image quality displayed on the display module.

[0008] Among them, the local processing for noise suppression and edge enhancement selects a 3×3 pixel window;

[0009] The image adjustment and processing module includes an image visual effect improvement unit, a noise and edge enhancement unit, and an adaptive histogram equalization and feedback adjustment unit.

[0010] The device used by the image input module includes a computing device;

[0011] The device used by the display module includes a display device;

[0012] The image adjustment processing module uses an input device;

[0013] The computing device includes a CPU, which is used to perform basic mathematical operations and logical judgments; the display device includes a monitor, which is used to display the original image and the processed image; and the input device includes a keyboard, a mouse and a touch screen, which are used by the user to input parameters, select processing options and interact with the system.

[0014] Optionally, the calculation formula for the image visual effect improvement unit is as follows:

[0015] LDT = a × YST + b;

[0016] in:

[0017] LDT stands for enhanced adjusted pixel values;

[0018] 'a' is the contrast adjustment factor;

[0019] YST represents the original image data values, reflecting any channel of the grayscale matrix and the RGB color matrix.

[0020] b is the brightness adjustment value;

[0021] The specific calculations for a and b are explained below:

[0022] If a > 1, the image contrast is enhanced;

[0023] If a < 1, the image contrast will be reduced;

[0024] If b is a positive value, the image brightness will increase;

[0025] If b is a positive value, the image brightness will decrease.

[0026] Optionally, the calculation formula for the noise and edge enhancement unit is as follows:

[0027] ;

[0028] LDT c，avg = (LDT) （n1，m1） +LDT （n2，m2） +LDT （n3，m3） ......+LDT （n9，m9） ) / 9;

[0029] in:

[0030] LDT c（x，y） Increase the pixel value at the (x, y) position within the window;

[0031] CK （x，y） For a pixel window centered at (x, y), CK （x，y） A 3×3 pixel window reflecting the coordinate axes centered at (x,y);

[0032] LDT （n，m） This represents the pixel value at position (n, m) within the window;

[0033] LDT c，avg This represents the average pixel value at position (n, m) within the window.

[0034] d is the edge enhancement coefficient, which is used to adjust the intensity of edge enhancement;

[0035] LDT （x，y） LDT enhances pixel values ​​at (x, y) location. （x，y） The enhanced adjusted pixel value LDT at the (x, y) position within the overall image;

[0036] LDT of pixel average at position (n, m) within the window c，avg The specific calculation instructions are as follows;

[0037] Because the pixel window CK is located at the center (x, y) （x，y） Reflecting a 3×3 pixel window centered at (x, y), the LDT of the pixel value at position (n, m) within the 3×3 pixel window. （n，m） The total computational cost is 9, "LDT （n1，m1） +LDT （n2，m2） +LDT （n3，m3） ......+LDT （n9，m9） "LDT of pixel values ​​at position (n, m) within 9 windows" （n，m） The sum of .

[0038] Optionally, the pixel window CK at the center (x, y) position （x，y） The specific details of using a 3×3 pixel window are as follows;

[0039] A 3×3 pixel window is a moderate size, which can capture neighborhood information for analysis and processing, and avoids blurring image details due to a large window.

[0040] A 3×3 pixel window has a simple geometry and symmetry, making it easy to implement and optimize in image processing algorithms;

[0041] In image processing tasks, a 3×3 pixel window is well represented and recognized within its local neighborhood.

[0042] Optionally, the calculation formula for the adaptive histogram equalization and feedback adjustment unit is as follows:

[0043] CK（x，y） =ZTH(LDT) c（x，y） )×ZSY;

[0044] ZSY=[1+(LDT max -LDT min ) / (LDT) （x，y） -LDT min )];

[0045] in:

[0046] CK （x，y） The processed pixel value at position (x, y);

[0047] ZTH is a histogram equalization function, which is used to improve image contrast.

[0048] ZSY is an adaptive adjustment factor;

[0049] LDT min To enhance the minimum pixel value after adjustment;

[0050] LDT max To enhance the adjusted maximum pixel value.

[0051] Optionally, based on the processed (x, y) position pixel value CK （x，y） In image processing, the final visual effect is adjusted according to the dynamic range standard of different images. Taking an 8-bit image as an example, its brightness dynamic range is 0 to 255, so 0-255 is the reference range for the brightness of the image.

[0052] The formulas reflecting the brightness width based on the image visual effect improvement unit and the adaptive histogram equalization and feedback adjustment unit are as follows:

[0053] LF=LDT max -LDT min ;

[0054] LF represents the brightness width value.

[0055] Optionally, the brightness width value LF and the image brightness width reference value 255 are further analyzed and adjusted in image processing as follows:

[0056] S1. If the brightness width value LF is below 150 within the range of 0-255, it indicates that the image brightness range is narrow and the image contrast is low. The adaptive adjustment factor ZSY of the adaptive histogram equalization and feedback adjustment unit should be adjusted and applied to the adaptive histogram equalization and feedback adjustment unit for calculation. The adjustment formula for the adaptive adjustment factor ZSY is as follows:

[0057] ZSY=[1+(c×LDT max -LDT min ) / (LDT) （x，y） -LDT min )];

[0058] c is a scaling factor, used to enhance the visual effect of histogram equalization and the contrast of the image;

[0059] S2. If the brightness width value LF is between 150 and 255 within the range of 0-255, it indicates that the image brightness range is wide. While maintaining the current adaptive adjustment factor ZSY, output the pixel value CK at position (x, y). （x，y） .

[0060] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0061] I. The system of this invention uses artificial intelligence algorithms to automatically analyze image content and dynamically adjusts the parameters in the image adjustment and processing module, which includes an image visual effect improvement unit and a noise and edge enhancement unit, according to image features, in order to achieve optimal brightness and contrast adjustment, noise suppression and edge enhancement effects.

[0062] Second, this invention enables the system to intelligently distinguish noise and edge information by calculating the local mean and variance in the noise and edge enhancement unit, thereby suppressing noise while preserving or even enhancing the edge details of the image.

[0063] Third, based on histogram equalization, this invention introduces a feedback adjustment mechanism to dynamically adjust the brightness and contrast of the equalized image according to the pixel value range of the enhanced pixel value LDT, so as to avoid the problem of over-enhancement or under-enhancement.

[0064] Fourth, the system of the present invention organically combines the steps of brightness and contrast adjustment, noise suppression and edge enhancement, adaptive histogram equalization and feedback adjustment to form an efficient and complete image processing flow, thereby significantly improving the quality and efficiency of image processing. Attached Figure Description

[0065] Figure 1 This is a flowchart of the method for this artificial intelligence-based image processing system;

[0066] Figure 2 This is a schematic diagram of the image adjustment processing module of the present invention.

[0067] Figure 3 This is a schematic diagram of the brightness and contrast adjustment of the image visual improvement unit of the present invention;

[0068] Figure 4This is a schematic diagram illustrating the benefits of the 3×3 pixel window of the noise and edge enhancement unit of the present invention. Detailed Implementation

[0069] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0070] This AI-based image processing system differs from traditional image processing methods. Traditional methods often rely on fixed algorithms and parameter settings, which often fall short when dealing with complex and varied images, particularly in areas such as brightness adjustment, contrast enhancement, noise suppression, and edge enhancement. Traditional methods struggle to simultaneously meet the requirements of efficiency and high quality. Furthermore, histogram equalization, a common technique for improving image contrast, suffers from both over-enhancement and under-enhancement. This algorithm unit overcomes the shortcomings of traditional methods, achieving more intelligent, efficient, and high-quality image processing results. Especially when processing complex and varied images, the system can automatically analyze image content and dynamically adjust processing parameters to achieve optimal processing results. This AI-based image processing system brings new ideas and methods to the field of image processing, possessing broad application prospects and significant research value.

[0071] Example 1, please refer to Figures 1 to 4 This embodiment provides an image processing system based on artificial intelligence, and the specific implementation steps are as follows:

[0072] Users input raw images into the system using the image input module. The system receives the raw image data and displays the image using the display module.

[0073] Based on the original image data, and using the image adjustment processing module, brightness and contrast are adjusted, noise is suppressed and edge is enhanced, and histogram equalization is performed to gradually improve the image quality displayed on the display module.

[0074] Among them, the local processing for noise suppression and edge enhancement selects a 3×3 pixel window;

[0075] The image adjustment and processing module includes an image visual effect improvement unit, a noise and edge enhancement unit, and an adaptive histogram equalization and feedback adjustment unit;

[0076] The devices used in the image input module include computing devices;

[0077] The devices used in the display module include display devices;

[0078] The image adjustment and processing module uses input devices;

[0079] The computing device includes a CPU, which is used to perform basic mathematical operations and logical judgments; the display device includes a monitor, which is used to display raw and processed images; and the input device includes a keyboard, mouse, and touch screen, which are used by users to input parameters, select processing options, and interact with the system.

[0080] In this embodiment, the system utilizes the cooperation of three algorithm units, combining LDT and LDT... c（x，y） and CK （x，y） The three calculation results are: LDT (Laser Decomposition Transformation) and LDT (Laser Decomposition Transformation ... c（x，y） This method enhances pixel values ​​at the (x, y) position within a window, suppressing noise while preserving image edge information. Edge enhancement is achieved by calculating pixel value differences within the local window, while mean filtering is used to smooth noise. CK （x，y） To process the pixel values ​​at (x, y) positions, based on histogram equalization, the brightness and contrast of the image are further adjusted using an adaptive adjustment factor to achieve finer control. Combining the processing results from the previous two steps, histogram equalization and adaptive adjustment optimize the image's contrast and brightness, making the image more suitable for subsequent processing and observation. Furthermore, the calculation results can also influence and feed back to the LDT and LDT values. c（x，y） The calculations enable the three algorithms in this system to have a high degree of correlation and entanglement, allowing the overall algorithm system to automatically provide feedback and optimization based on actual conditions to better reflect reality.

[0081] Please see Figures 1 to 4 The calculation formula for the image visual effect improvement unit is as follows:

[0082] LDT = a × YST + b;

[0083] in:

[0084] LDT stands for enhanced adjusted pixel values;

[0085] 'a' is the contrast adjustment factor;

[0086] YST represents the original image data values, reflecting any channel of the grayscale matrix and the RGB color matrix.

[0087] b is the brightness adjustment value;

[0088] The specific calculations for a and b are explained below:

[0089] If a > 1, the image contrast is enhanced;

[0090] If a < 1, the image contrast will be reduced;

[0091] If b is a positive value, the image brightness will increase;

[0092] If b is a positive value, the image brightness will decrease.

[0093] In this embodiment: First, the algorithm unit adjusts the contrast adjustment factor a to enable the system to intelligently identify dark and bright areas in the image and enhance them respectively. When a>1, the pixel value of the dark area is further reduced and the pixel value of the bright area is further increased, thereby creating a stronger contrast effect visually. This contrast enhancement helps to highlight the details and sense of layering in the image, making the image more vivid and realistic.

[0094] Adjusting the brightness adjustment value b directly affects the overall brightness level of the image. By appropriately increasing or decreasing the brightness adjustment value b, it is possible to ensure that the image is neither overexposed nor underexposed when displayed, thereby retaining more detail information. This fine-tuning of brightness is crucial for improving the visual comfort of the image.

[0095] In this embodiment, images taken in environments with insufficient or excessive light often suffer from uneven brightness and insufficient contrast. The application of the image visual effect improvement unit can automatically compensate for the adverse effects of these lighting conditions, thereby significantly improving image quality. Furthermore, different devices have performance differences during image acquisition and transmission, resulting in inconsistent image quality. By adjusting the image visual effect improvement unit, the image quality problems caused by insufficient device performance can be compensated to a certain extent.

[0096] In this embodiment, as the first step in the image processing flow, the image visual effect improvement unit adjusts the brightness and contrast to provide more ideal input data for subsequent noise suppression and edge enhancement processing steps. This optimization helps to improve the efficiency and effectiveness of subsequent processing steps. By improving the initial quality of the image, the image visual effect improvement unit reduces the difficulty and complexity of subsequent processing steps. For example, in the noise suppression process, due to the improvement of image brightness and contrast, the system can more easily distinguish noise from useful information. In the edge enhancement process, due to the enhancement of image hierarchy, the system can more accurately identify edge information.

[0097] Please see Figures 1 to 4 The calculation formulas for noise and edge enhancement units are as follows:

[0098] ;

[0099] LDT c，avg= (LDT) （n1，m1） +LDT （n2，m2） +LDT （n3，m3） ......+LDT （n9，m9） ) / 9;

[0100] in:

[0101] LDT c（x，y） Increase the pixel value at the (x, y) position within the window;

[0102] CK （x，y） For a pixel window centered at (x, y), CK （x，y） A 3×3 pixel window reflecting the coordinate axes centered at (x,y);

[0103] LDT （n，m） This represents the pixel value at position (n, m) within the window;

[0104] LDT c，avg This represents the average pixel value at position (n, m) within the window.

[0105] d is the edge enhancement coefficient, which is used to adjust the intensity of edge enhancement;

[0106] LDT （x，y） LDT enhances pixel values ​​at (x, y) location. （x，y） The enhanced adjusted pixel value LDT at the (x, y) position within the overall image;

[0107] LDT of pixel average at position (n, m) within the window c，avg The specific calculation instructions are as follows;

[0108] Because the pixel window CK is located at the center (x, y) （x，y） Reflecting a 3×3 pixel window centered at (x, y), the LDT of the pixel value at position (n, m) within the 3×3 pixel window. （n，m） The total computational cost is 9, "LDT （n1，m1） +LDT （n2，m2） +LDT （n3，m3） ......+LDT （n9，m9） "LDT of pixel values ​​at position (n, m) within 9 windows" （n，m） The sum of .

[0109] In this embodiment, the noise and edge enhancement unit identifies noise by calculating the difference in pixel values ​​within a local window. This method utilizes the principle of spatial correlation of pixel values ​​in an image, that is, adjacent pixels usually have similar values. When the difference in pixel values ​​within a local window exceeds a certain threshold, the system regards it as noise and suppresses it. While suppressing noise, the noise and edge enhancement unit also retains the detail information in the image by adjusting the edge enhancement coefficient d. This balanced processing helps to avoid the image becoming blurry and distorted during the noise suppression process.

[0110] Edges are the boundaries between different regions in an image and are crucial for image analysis and recognition. Noise and edge enhancement units improve edge information, making edges in an image clearer and more prominent. This helps subsequent edge detection and image segmentation processes to be more accurate and efficient. Enhanced edge information not only improves image visibility but also enhances its readability. In the fields of medical image analysis and remote sensing image interpretation, clear edge information is of great significance for professionals' judgment and decision-making.

[0111] The combination of noise suppression and edge enhancement makes the processed image clearer and sharper while maintaining the original details. This combined effect helps improve the overall image quality and enhance the user experience. Clearer images help reduce visual fatigue caused by prolonged viewing. In fields such as digital entertainment and online education, this improvement in clarity is of great significance for improving the user's viewing experience and satisfaction.

[0112] Please see Figures 1 to 4 The calculation formulas for adaptive histogram equalization and feedback adjustment units are as follows:

[0113] CK （x，y） =ZTH(LDT) c（x，y） )×ZSY;

[0114] ZSY=[1+(LDT max -LDT min ) / (LDT) （x，y） -LDT min )];

[0115] in:

[0116] CK （x，y） The processed pixel value at position (x, y);

[0117] ZTH is a histogram equalization function, which is used to improve image contrast.

[0118] ZSY is an adaptive adjustment factor;

[0119] LDT minTo enhance the minimum pixel value after adjustment;

[0120] LDT max To enhance the adjusted maximum pixel value.

[0121] In this embodiment, firstly, the histogram equalization function ZTH is a function used to equalize the histogram of an image. The purpose of histogram equalization is to improve the contrast of the image and make the histogram distribution of the image more uniform. Specifically, it remaps the pixel values ​​of the image so that the frequency of each pixel value appearing in the image, i.e., the height of the histogram, is approximately equal. In this way, the pixel values ​​that were originally unevenly distributed on the histogram will be stretched to the entire brightness range, thereby enhancing the contrast of the image.

[0122] In this algorithm unit, histogram equalization is an effective method to enhance image contrast by adjusting the histogram distribution of the image. The histogram equalization function ZTH in the adaptive histogram equalization and feedback adjustment unit implements this function. It enhances the image contrast by stretching the histogram distribution range of the image. This processing makes the transition between dark and bright areas in the image smoother and more natural. While enhancing contrast, the adaptive histogram equalization and feedback adjustment unit also focuses on preserving the detail information in the image. Through reasonable histogram equalization processing parameter settings and the introduction of the adaptive adjustment factor ZSY, the system can enhance contrast while avoiding the loss and distortion of image details.

[0123] The adaptive adjustment factor ZSY in the adaptive histogram equalization and feedback adjustment unit dynamically adjusts the histogram equalization result based on the pixel value range of the enhanced pixel value LDT. This adjustment method enables the system to flexibly process different images according to their actual conditions, thereby avoiding over-enhancement or under-enhancement. The introduction of the adaptive adjustment factor ZSY enables the adaptive histogram equalization and feedback adjustment unit to achieve more stable and consistent results when processing different images. This optimization effect helps to improve the performance and reliability of the entire image processing system.

[0124] The cyclical influence of adaptive histogram equalization and feedback adjustment units on the image visual effect improvement units is mainly reflected in the feedback mechanism. When the pixel value CK at position (x, y) is processed... （x，y）When the quality is improved, it can serve as one of the input data for subsequent image processing. This feedback mechanism enables the entire image processing system to form a closed-loop control system, continuously optimizing and improving the efficiency and effect of image processing. Since the adaptive adjustment factor ZSY in the adaptive histogram equalization and feedback adjustment unit is related to the pixel value range of the enhanced adjusted pixel value LDT, it can influence the selection of a and b in the image visual effect improvement unit to a certain extent. This fine adjustment helps to achieve more accurate and personalized brightness and contrast adjustment schemes, thereby meeting the needs of different application scenarios.

[0125] Please see Figures 1 to 4 After processing, the pixel value CK at position (x, y) （x，y） In image processing, the final visual effect is adjusted according to the dynamic range standard of different images. Taking an 8-bit image as an example, its brightness dynamic range is 0 to 255, so 0-255 is the reference range for the brightness of the image.

[0126] The formulas reflecting the brightness range based on the image visual effect improvement unit and the adaptive histogram equalization and feedback adjustment unit are as follows:

[0127] LF=LDT max -LDT min ;

[0128] LF represents the brightness width value;

[0129] Based on the brightness width value LF and the image brightness width reference value of 255, the image processing is further analyzed and adjusted as follows:

[0130] S1. If the brightness width value LF is below 150 within the range of 0-255, it indicates that the image brightness range is narrow and the image contrast is low. The adaptive adjustment factor ZSY of the adaptive histogram equalization and feedback adjustment unit should be adjusted and applied to the adaptive histogram equalization and feedback adjustment unit for calculation. The adjustment formula for the adaptive adjustment factor ZSY is as follows:

[0131] ZSY=[1+(c×LDT max -LDT min ) / (LDT) （x，y） -LDT min )];

[0132] c is a scaling factor, used to enhance the visual effect of histogram equalization and the contrast of the image;

[0133] S2. If the brightness width value LF is between 150 and 255 within the range of 0-255, it indicates that the image brightness range is wide. While maintaining the current adaptive adjustment factor ZSY, output the pixel value CK at position (x, y). （x，y） .

[0134] In this embodiment, the algorithm unit further adjusts the adaptive adjustment factor ZSY based on the judgment of the brightness range to ensure the visual effect of the image;

[0135] First, the adaptive adjustment factor ZSY can dynamically adjust the intensity of histogram equalization based on the results of the image visual effect improvement unit. This means that if the brightness range of the image is narrow and the contrast is insufficient, the adaptive adjustment factor ZSY will correspondingly enhance the effect of histogram equalization to expand the brightness range of the image and enhance the contrast. This adjustment helps to make the bright and dark parts of the image more distinct, improve the detail of the image, and thus optimize the visual effect of the image.

[0136] If histogram equalization is not properly controlled, it will lead to excessive amplification of image contrast, resulting in unnatural visual effects such as color distortion and over-sharpening. The introduction of the adaptive adjustment factor ZSY effectively solves this problem. It can adjust the intensity of equalization according to the actual situation of the image, ensuring that while enhancing contrast, it avoids over-processing and maintains the naturalness and authenticity of the image.

[0137] By optimizing the contrast and brightness distribution of an image, the adaptive adjustment factor ZSY significantly improves the overall quality of the image, making it more visually appealing. It also provides better input data for subsequent image processing steps, including feature extraction, image segmentation, and target recognition. High-quality images help improve the accuracy and efficiency of these processing steps, thereby promoting the development of the entire image processing workflow.

[0138] Another important advantage of the adaptive adjustment factor ZSY is that it enhances the flexibility and adaptability of the image processing workflow. Because it can dynamically adjust the equalization parameters according to the actual situation of the image, it can process various types of images and meet different visual effect requirements. This flexibility makes the image processing workflow more universal and practical, and can be widely used in multiple fields such as medical image processing, satellite remote sensing image processing, and security monitoring video analysis.

[0139] Finally, the introduction of the adaptive adjustment factor ZSY has also promoted the innovation and development of image processing technology. It has inspired researchers to conduct in-depth research and exploration of image processing algorithms, and promoted the continuous emergence of new technologies and methods. These new technologies and methods have not only improved the speed and accuracy of image processing, but also brought more possibilities and development space to the field of image processing.

[0140] In summary, the adaptive adjustment factor ZSY brings several beneficial effects to the process of influencing adaptive histogram equalization and feedback adjustment units, and these effects together promote the development and application of image processing technology.

[0141] Example 2, please refer to Figures 1 to 4 CK, the pixel window at the center (x, y) position （x，y） The specific details of using a 3×3 pixel window are as follows;

[0142] A 3×3 pixel window is a moderate size, which can capture neighborhood information for analysis and processing, and avoids blurring image details due to a large window.

[0143] A 3×3 pixel window has a simple geometry and symmetry, making it easy to implement and optimize in image processing algorithms;

[0144] In image processing tasks, a 3×3 pixel window is well represented and recognized within its local neighborhood.

[0145] In this embodiment, by considering the neighborhood information of a 3×3 pixel window, the algorithm can better adapt to local changes in the image. For example, in noise suppression, it can smooth noise points based on the similarity of neighboring pixels while preserving edge information. Compared with global processing, local window processing usually has lower computational complexity because it only processes a small part of the image at a time, thus reducing the overall computational load. Local window processing allows the algorithm to apply different processing strategies to different regions of the image. For example, in edge detection, it can identify edges based on the gradient changes of pixels within the window.

[0146] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An image processing system based on artificial intelligence, characterized in that, The specific implementation steps are as follows: Users input raw images into the system using the image input module; The system receives raw image data and displays the image using the display module; Based on the original image data, and using the image adjustment processing module, brightness and contrast are adjusted, noise is suppressed and edge is enhanced, and histogram equalization is performed to gradually improve the image quality displayed on the display module. Among them, the local processing for noise suppression and edge enhancement selects a 3×3 pixel window; The image adjustment and processing module includes an image visual effect improvement unit, a noise and edge enhancement unit, and an adaptive histogram equalization and feedback adjustment unit. The calculation formula for the image visual effect improvement unit is as follows: LDT = a × YST + b; in: LDT stands for enhanced adjusted pixel values; 'a' is the contrast adjustment factor; YST represents the original image data values, reflecting any channel of the grayscale matrix or RGB color matrix. b is the brightness adjustment value; The specific calculations for a and b are explained below: If a > 1, the image contrast is enhanced; If a < 1, the image contrast will be reduced; If b is a positive value, the image brightness will increase; If b is a positive value, the image brightness will decrease; The calculation formula for the noise and edge enhancement unit is as follows: ； Ltd. c，avg =(LDT （n1，m1） +LDT （n1，m2） +LDT （n1，m3） ......+LDT （n3，m3） ) / 9; in: LDT c（x，y） Increase the pixel value at the (x, y) position within the window; CK （x，y） For a pixel window centered at (x, y), CK （x，y） A 3×3 pixel window reflecting the coordinate axes centered at (x,y); LDT （ni，mj） This represents the pixel value at position (ni, mj) within the window. LDT c，avg This represents the average pixel value at position (ni, mj) within the window. d is the edge enhancement coefficient, which is used to adjust the intensity of edge enhancement; LDT （x，y） LDT enhances pixel values ​​at (x, y) location. （x，y） The enhanced adjusted pixel value LDT is reflected in the (x, y) position within the overall image.

2. The image processing system based on artificial intelligence according to claim 1, characterized in that, The device used by the image input module includes a computing device; The device used by the display module includes a display device; The image adjustment processing module uses an input device.

3. The image processing system based on artificial intelligence according to claim 2, characterized in that: The pixel window CK at the center (x, y) position （x，y） The specific details of using a 3×3 pixel window are as follows; A 3×3 pixel window is a moderate size, which can capture neighborhood information for analysis and processing, and avoids blurring image details due to a large window. A 3×3 pixel window has a simple geometry and symmetry, making it easy to implement and optimize in image processing algorithms; In image processing tasks, a 3×3 pixel window is well represented and recognized within its local neighborhood.

4. The image processing system based on artificial intelligence according to claim 3, characterized in that: The calculation formula for the adaptive histogram equalization and feedback adjustment unit is as follows: CK （x，y） =ZTH(LDT c（x，y） )×ZSY; ZSY=[1+(LDT max -LDT min ) / (LDT （x，y） -LDT min )]; in: CK （x，y） The processed pixel value at position (x, y); ZTH is a histogram equalization function, which is used to improve image contrast. ZSY is an adaptive adjustment factor; LDT min To enhance the minimum pixel value after adjustment; LDT max To enhance the adjusted maximum pixel value; The processed pixel value CK at position (x, y) （x，y） In image processing, the final visual effect is adjusted according to the dynamic range standard of different images. The dynamic range of brightness is 0 to 255, so 0-255 is the reference range for the brightness of the image. The formulas reflecting the brightness width based on the image visual effect improvement unit and the adaptive histogram equalization and feedback adjustment unit are as follows: LF=LDT max -LDT min ; LF represents the brightness width value; Based on the brightness width value LF and the image brightness width reference value 255, the image processing is further analyzed and adjusted as follows: S1. If the brightness width value LF is below 150 within the range of 0-255, it indicates that the image brightness range is narrow and the image contrast is low. The adaptive adjustment factor ZSY of the adaptive histogram equalization and feedback adjustment unit should be adjusted and applied to the adaptive histogram equalization and feedback adjustment unit for calculation. The adjustment formula for the adaptive adjustment factor ZSY is as follows: ZSY=[1+(c×LDT max -LDT min ) / (LDT （x，y） -LDT min )]; c is a scaling factor, used to enhance the visual effect of histogram equalization and the contrast of the image; S2. If the brightness width value LF is between 150 and 255 within the range of 0-255, it indicates that the image brightness range is wide. While maintaining the current adaptive adjustment factor ZSY, output the pixel value CK at position (x, y). （x，y） .

5. The image processing system based on artificial intelligence according to claim 2, characterized in that: The computing device includes a CPU, which is used to perform basic mathematical operations and logical judgments; the display device includes a monitor, which is used to display the original image and the processed image; and the input device includes a keyboard, a mouse and a touch screen, which are used by the user to input parameters, select processing options and interact with the system.

Images