Image processing method and system based on artificial intelligence

Abstract

The invention discloses an image processing method and system based on artificial intelligence, and relates to the technical field of image processing, and the method comprises the following steps: carrying out the preliminary extraction of image feature information, optimizing the image quality, obtaining optimized image data, carrying out the training of a model, and carrying out the target detection. Identifying a target area in the image and providing an identifier, carrying out local feature alignment on the target area, carrying out classification and labeling on the target area, carrying out enhancement and repair processing, and customizing an output format according to an application scene; and a function module is constructed. By setting a self-adaptive image enhancement algorithm, target detection, a feature extraction technology and a generative adversarial network (GAN) for image restoration, efficient positioning, optimization and restoration of an image target area are realized. The method not only has significant advantages in noise removal and target detection precision, but also avoids distortion caused by excessive processing through dynamic adjustment of an enhancement strategy, and ensures prominent performance and natural fusion of a target area in an image.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to an image processing method and system based on artificial intelligence. Background Art

[0002] With the rapid development of artificial intelligence and deep learning technology, image processing technology has been widely used in many fields, such as medical image analysis, autonomous driving, security monitoring, industrial detection, etc. However, existing image processing methods still have some shortcomings when facing tasks such as target detection, target area enhancement and restoration in complex scenes.

[0003] Traditional image processing methods often rely on manually designed feature extraction algorithms. It is difficult to balance accuracy and efficiency in the processing process, and it is easily affected by factors such as image noise, blur, and uneven lighting. In addition, the existing target detection technology has weak recognition capabilities in high noise, low resolution, or complex backgrounds, resulting in inaccurate target positioning and affecting the quality of subsequent processing. Although deep learning methods have made significant progress in image recognition, existing models still have limitations in processing detail optimization of target areas, image restoration, and adaptive enhancement of target areas, and cannot fully meet the needs of real-time, high efficiency, and high quality.

[0004] In the existing technology, although deep learning models have achieved good performance in the field of target detection, in the process of image enhancement and restoration, they usually rely on fixed enhancement strategies and over-reliance on generative adversarial network (GAN) restoration models, which may cause some details to be over-processed or distorted, affecting the naturalness and authenticity of the image. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides an image processing method and system based on artificial intelligence to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions: In a first aspect, an embodiment of the present invention provides an image processing method based on artificial intelligence, comprising the following steps: S1. Preliminary extraction of image feature information based on image input signal; S2. Based on the extracted preliminary feature information, optimize the image quality, remove noise and improve the image clarity to obtain optimized image data; S3, using the optimized image data, training the model and performing target detection, identifying the target area in the image and providing identification; S4. After completing the target detection, the local features of the target area are aligned, and the target area is classified and labeled according to the semantic information; S5, performing enhancement and restoration processing on the marked target area, performing image enhancement and image restoration, and obtaining processed image data; S6. Customize the output format of the processed image data according to the application scenario.

[0007] Further optimizing the technical solution, in the step S1, a deep learning model of a convolutional neural network (CNN) is used to perform preliminary feature extraction on the image, identify low-level features in the image, including edges, textures, and color distribution, and form a preliminary image description; In the initial extraction process, convolution kernels are used to extract local features of the image, and pooling operations are combined to reduce the dimension and extract the spatial hierarchical structure information of the image.

[0008] To further optimize the technical solution, in step S2, an adaptive filter is used to suppress noise. The adaptive filter dynamically adjusts the parameters of the filter according to the local characteristics of the image, and suppresses different types of noise to different degrees according to the noise characteristics of different areas in the image.

[0009] To further optimize the technical solution, in step S3, the model trained using the optimized image data is as follows: ; in, : The total loss function of the training model; : The number of training samples; : The coordinate set of the real box and the predicted box; : Intersection-over-union ratio, which measures the overlap between the predicted box and the true box; : Loss weight coefficient, used to balance position loss and category loss; : total number of categories; :sample In category The real label on :sample In category The predicted probability of .

[0010] To further optimize the technical solution, in step S3, when performing target detection, a multi-scale target detection model is constructed, and the multi-scale target detection model is as follows: ; in, : The set of detected targets; : The overall inference function of the model; : A set of feature maps of the input image; : The weight parameter set of the model; : A set of multi-scale layers, representing the resolution levels of different feature maps; :For the Object detection function of layer-scale feature map; : No. Layer feature maps and corresponding weights.

[0011] To further optimize the technical solution, in step S4, a local feature matching algorithm is used to align local features of the target area, including: Key point detection; Feature descriptor extraction; Feature matching; Geometric transformation estimation; Target alignment.

[0012] To further optimize the technical solution, in step S4, a deep semantic network is used to classify and label the target area according to the semantic information, including: Image description generation; Semantic information embedding; Refine the classification; Object annotation and relationship reasoning; Output the annotations.

[0013] To further optimize the technical solution, in step S5, the image enhancement adopts an adaptive enhancement algorithm, constructs an adaptive enhancement loss function, and adaptively enhances the target area. The adaptive enhancement loss function is as follows: ; in, : Adaptive enhancement total loss; : Number of target areas; : Target area The original contrast value of : Target area The enhanced contrast value of : Target area The original brightness value of : Target area The enhanced brightness value of : For the target area Sharpness measurement function of ; : Weight coefficient, used to balance the optimization of contrast, brightness and sharpness; : Features of the target area after enhancement.

[0014] Further optimizing the technical solution, in step S5, the image restoration is performed using a generative adversarial network (GAN), which generates and optimizes the image through a discriminator and a generator to ensure the naturalness and authenticity of the target area; The loss function between the discriminator and the generator of the generative adversarial network is designed as follows: ; in, : GAN total loss; : The discriminator is trained on real images judgment; : The image generated by the generator, based on the noise vector ; : Real data distribution; : Noisy data distribution.

[0015] An artificial intelligence-based image processing system is constructed based on the above-mentioned artificial intelligence-based image processing method. The functional modules of the system include a preliminary feature extraction module, a data preprocessing module, a target area detection module, a target classification and labeling module, an image enhancement and restoration module, and an image format conversion module.

[0016] In a second aspect, an embodiment of the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of an artificial intelligence-based image processing method and system as described in the first aspect of the present invention are implemented.

[0017] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of an artificial intelligence-based image processing method and system as described in the first aspect of the present invention are implemented.

[0018] Compared with the prior art, the present invention provides an image processing method and system based on artificial intelligence, which has the following beneficial effects: This AI-based image processing method and system achieves efficient positioning, optimization, and restoration of the target area of ​​the image by setting an adaptive image enhancement algorithm, target detection, feature extraction technology, and generative adversarial network GAN for image restoration. It not only has significant advantages in noise removal and target detection accuracy, but also can dynamically adjust the enhancement strategy to avoid distortion caused by over-processing, ensuring the prominent performance and natural integration of the target area in the image. Compared with traditional image processing technology, it effectively improves the image quality and detail restoration capabilities, has stronger robustness and adaptability, can handle more complex image scenes, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 A schematic diagram of a flow chart of an image processing method based on artificial intelligence proposed by the present invention; Figure 2 A schematic diagram of the flow of a local feature matching algorithm in an artificial intelligence-based image processing method proposed by the present invention; Figure 3 A schematic diagram of the process of a deep semantic network in an artificial intelligence-based image processing method proposed by the present invention; Figure 4 This is a schematic diagram of the functional module composition of an artificial intelligence-based image processing system proposed in the present invention. DETAILED DESCRIPTION

[0021] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0022] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0023] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0024] Embodiment 1: Reference Figure 1 to Figure 3 , which is the first embodiment of the present invention, provides an image processing method based on artificial intelligence, comprising the following steps: S1. Based on the image input signal, preliminary extraction of image feature information is performed.

[0025] In this embodiment, a convolutional neural network (CNN) deep learning model is used to perform preliminary feature extraction on the image, identify low-level features in the image, including edges, textures, and color distribution, and form a preliminary image description; these low-level features will provide a basis for subsequent image enhancement, target detection, and classification tasks.

[0026] In the initial extraction process, convolution kernels are used to extract local features of the image, and pooling operations are combined to reduce the dimension and extract the spatial hierarchical structure information of the image. The dataset after feature extraction not only provides rough outline information for image understanding, but also provides effective input for subsequent reasoning and recognition.

[0027] S2. Based on the extracted preliminary feature information, optimize the image quality, remove noise and improve image clarity to obtain optimized image data.

[0028] In this embodiment, an adaptive filter is used to suppress noise. The adaptive filter dynamically adjusts the parameters of the filter according to the local characteristics of the image, and suppresses different types of noise to different degrees according to the noise characteristics of different areas in the image.

[0029] Traditional fixed filters often have poor effects when processing complex scenes, while adaptive filtering can autonomously optimize the filtering process based on the noise characteristics of different areas in the image. This method can not only reduce the impact of noise on image quality, but also retain important details in the image, avoid excessive blurring, and thus improve the accuracy of subsequent image analysis tasks.

[0030] S3. Use the optimized image data to train the model and perform target detection, identify the target area in the image and provide identification.

[0031] In this embodiment, the model trained using the optimized image data is as follows: ; in, : The total loss function of the training model; : The number of training samples; : The coordinate set of the real box and the predicted box; : Intersection-over-union ratio, which measures the overlap between the predicted box and the true box; : Loss weight coefficient, used to balance position loss and category loss; : total number of categories; :sample In category The real label on :sample In category The predicted probability of .

[0032] The model aims to jointly optimize object location detection and category classification. Calculate the error of the target position to ensure that the model can accurately predict the bounding box of the object; the second item It is a classification loss based on cross entropy and is used to evaluate the accuracy of category prediction. During training, by optimizing Minimize and gradually adjust the weights of the deep network so that it can learn the salient features of the target in the optimized data (step S2) Furthermore, when performing target detection, a multi-scale target detection model is constructed, and the multi-scale target detection model is as follows: ; in, : The set of detected targets; : The overall inference function of the model; : A set of feature maps of the input image; : The weight parameter set of the model; : A set of multi-scale layers, representing the resolution levels of different feature maps; :For the Object detection function of layer-scale feature map; : No. Layer feature maps and corresponding weights.

[0033] The core of this model is multi-scale target detection. The function performs target detection on feature maps at different resolutions, and the detection results are combined between different scales. Fusion is performed to ensure accurate detection of large and small objects. Each layer of detection function The regional features are extracted through convolution operation, and then combined with non-maximum suppression (NMS) to filter repeated detection boxes to generate the final target set. .

[0034] The multi-scale design can effectively solve the problem of target scale differences. For example, small targets rely more on high-resolution feature maps for detection, while large targets are more captured by low-resolution feature maps. In actual use, the detection process is carried out in steps: 1. Input optimized image features . 2. Process each scale feature map in turn . 3. Combine the detection results of all scales to generate a target set .

[0035] The model is first trained using the training model. The image data optimized in step S2 is input into the training framework, and the model weights are adjusted through the loss function so that the model can learn the high-level features in the image.

[0036] The feature map optimized in step S2 is input into the multi-scale target detection model to generate feature levels with multiple resolutions (such as high resolution to capture small targets and low resolution to capture large targets). The feature maps of different scales are processed layer by layer to generate candidate targets.

[0037] S4. After completing the target detection, the local features of the target area are aligned, and the target area is classified and labeled according to the semantic information.

[0038] In this embodiment, a local feature matching algorithm (such as SIFT and SURF) is used to align local features of the target area, including: Key point detection; First, the key points in the image are detected by SIFT or SURF algorithm. Key points are usually areas with significant local features in the image, such as edges, corners, and places where texture changes significantly.

[0039] Feature descriptor extraction; Local feature descriptors are extracted for the area around each key point. SIFT and SURF generate descriptors (such as 128-dimensional vectors generated by SIFT and 64-dimensional vectors generated by SURF), which can effectively capture the texture, color and shape information of the local area and are rotation and scale invariant.

[0040] Feature matching; Use a nearest neighbor search algorithm (such as brute force matching or kd tree) to match the key point descriptors in different images to find the most similar feature point pairs. These matching point pairs will help to align the images in the subsequent stage.

[0041] Geometric transformation estimation; The geometric transformation (such as the homography matrix) between the images is estimated by matching key point pairs, and one image is transformed into the coordinate system of the other image. Commonly used methods include RANSAC (random sampling consensus algorithm) to remove mismatched points.

[0042] Goal alignment; After geometric transformation, the target regions in the image are precisely aligned. These aligned image regions can provide accurate input data for subsequent target recognition, classification and labeling.

[0043] Use deep semantic networks (such as BERT) to classify and label target areas based on semantic information, including: Image description generation; After key point matching and object alignment in the image, the system generates a preliminary image description based on the contextual information of the image, such as the background around the object, the spatial relationship between objects, etc. This description can be a short sentence in natural language, summarizing the main objects in the image and their relationships.

[0044] Semantic information embedding; Input the image description into a pre-trained semantic network (such as BERT). The deep semantic network converts the input text into a higher-level semantic representation through multiple levels of semantic understanding modules. Through the relationship of context, the network can understand the semantic connection between different objects.

[0045] Refine the classification; Based on the semantic representation of the deep semantic network, the system will refine the classification of the objects in the image. That is, the model not only recognizes the general category of the object (such as "car" or "person"), but also further describes the attributes of the object (such as "red car", "person sitting on a chair") to improve the accuracy and details of the classification.

[0046] Object annotation and relationship reasoning; After completing the refined classification, the deep semantic network will infer the relationship between objects. For example, the model may infer that the relationship between a "person" and a "car" is "standing next to" or "driving", further improving the semantic understanding of the image.

[0047] Output annotation; Finally, the network outputs the identified objects and their attributes (such as category, location, and relationship) in the form of text annotations. This information can be used as the basis for further processing, display, or storage.

[0048] S5. Perform enhancement and restoration processing on the marked target area, perform image enhancement and image restoration, and obtain processed image data.

[0049] In this embodiment, the image enhancement adopts an adaptive enhancement algorithm, constructs an adaptive enhancement loss function, and adaptively enhances the target area. The adaptive enhancement loss function (AELF for short) is as follows: ; in, : Adaptive enhancement total loss; : Number of target areas; : Target area The original contrast value of : Target area The enhanced contrast value of : Target area The original brightness value of : Target area The enhanced brightness value of : For the target area Sharpness measurement function of ; : Weight coefficient, used to balance the optimization of contrast, brightness and sharpness; : Features of the target area after enhancement.

[0050] The AELF loss function mainly optimizes the enhancement effect through three parts: Contrast optimization: First The contrast difference of the target area is calculated to ensure that the contrast of the enhanced target area is improved without excessive distortion. Brightness Optimization: Item 2 Aims to adjust the brightness of the target area so that the target can maintain clarity while adapting to the complex background. Sharpness optimization: the third item Used to enhance the sharpness of the target edge. Through sharpening operations (such as the Laplacian operator or the sharpening convolution kernel in the deep learning model), the details of the target area are enhanced to make the target more prominent. During the enhancement process, the weight coefficient It will be dynamically adjusted according to the specific content of the target area to avoid loss of details or distortion of information caused by over-processing.

[0051] Image restoration is performed using a generative adversarial network (GAN). GAN generates and optimizes images through a discriminator and a generator to ensure the naturalness and authenticity of the target area. The loss function between the discriminator and the generator of the generative adversarial network is designed as follows: ; in, : GAN total loss; : The discriminator is trained on real images judgment; : The image generated by the generator, based on the noise vector ; : Real data distribution; : Noisy data distribution.

[0052] Discriminator Judge the input image through training Is it from the real data distribution or fake data generated by the generator. By accepting the random noise vector Generate images that are as realistic as possible.

[0053] pass The generator is able to fill in missing details in the target area and perform detail restoration and image quality improvement during the enhancement process.

[0054] In this step, based on the detection and calibration in step S4, the features of the target area in the image are extracted, including the boundary, brightness, contrast and sharpness of the target, etc. These features will be used as input for the enhancement process.

[0055] Using the adaptive enhancement loss function, the contrast, brightness and sharpness of the target area are optimized through the AELF loss function. The enhancement strategy is adaptively adjusted according to the local features of the target area to ensure that the target details are enhanced while the background area is not affected too much.

[0056] After enhancement, a GAN generative adversarial network is used to further repair the missing parts or details in the target area. The generator generates the missing details based on the characteristics of the target area, and the discriminator verifies the authenticity of the image to ensure the naturalness and consistency of the supplementary content in the target area.

[0057] Finally, the enhanced target area is optimized through a deep learning model to ensure that the target is more prominent against a complex background, enhancing the image visualization and usability of practical applications.

[0058] S6. Customize the output format of the processed image data according to the application scenario.

[0059] In this embodiment, the image is intelligently optimized and output according to the target application scenario. Combined with the specific requirements of the application scenario, such as the need for images to be used in medical diagnosis, autonomous driving, industrial detection and other fields, images of different formats and resolutions are output. The image saving format (such as JPG, PNG, TIFF, etc.) and the compatibility and processing efficiency of subsequent images are considered. According to the specific needs of the application, the optimization algorithm can automatically select the best output parameters and perform the final compression, format conversion and data storage on the image. Deeply combine artificial intelligence with specific application scenarios to ensure that the processed images can maximize their effects in practical applications.

[0060] Embodiment 2: Reference Figure 4 , which is the second embodiment of the present invention, and this embodiment provides an artificial intelligence-based image processing system, which is constructed based on the artificial intelligence-based image processing method described in Example 1. The functional modules of the system include a preliminary feature extraction module, a data preprocessing module, a target area detection module, a target classification and labeling module, an image enhancement and restoration module, and an image format conversion module.

[0061] The above modules are all constructed accordingly based on the artificial intelligence-based image processing method described in Example 1 to implement the process of the method.

[0062] Embodiment three: This embodiment also provides a computer device, which is applicable to an image processing method and system based on artificial intelligence, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement an image processing method and system based on artificial intelligence as proposed in the above embodiment.

[0063] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, an image processing method and system based on artificial intelligence as proposed in the above embodiment is implemented.

[0064] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, 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 and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0065] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0066] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0067] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk case (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0068] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit with a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit with a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0069] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An image processing method based on artificial intelligence, characterized in that: The following steps are involved: S1. Preliminary extraction of image feature information based on image input signal; S2. Based on the extracted preliminary feature information, optimize the image quality, remove noise and improve the image clarity to obtain optimized image data; S3, using the optimized image data, training the model and performing target detection, identifying the target area in the image and providing identification; S4. After completing the target detection, the local features of the target area are aligned, and the target area is classified and labeled according to the semantic information; S5, performing enhancement and restoration processing on the marked target area, performing image enhancement and image restoration, and obtaining processed image data; S6. Customize the output format of the processed image data according to the application scenario.

2. The image processing method based on artificial intelligence according to claim 1, characterized in that: In the step S1, a convolutional neural network (CNN) deep learning model is used to perform preliminary feature extraction on the image, identify low-level features in the image, including edges, textures, and color distribution, and form a preliminary image description; In the initial extraction process, convolution kernels are used to extract local features of the image, and pooling operations are combined to reduce the dimension and extract the spatial hierarchical structure information of the image.

3. The image processing method based on artificial intelligence according to claim 1, characterized in that: In step S2, an adaptive filter is used to suppress noise. The adaptive filter dynamically adjusts the parameters of the filter according to the local characteristics of the image, and suppresses different types of noise to different degrees according to the noise characteristics of different areas in the image.

4. The image processing method based on artificial intelligence according to claim 1, characterized in that: In step S3, the model trained using the optimized image data is as follows: ； in, : The total loss function of the training model; : The number of training samples; : The coordinate set of the real box and the predicted box; : Intersection-over-union ratio, which measures the overlap between the predicted box and the true box; : Loss weight coefficient, used to balance position loss and category loss; : total number of categories; :sample In category The real label on :sample In category The predicted probability of .

5. The image processing method based on artificial intelligence according to claim 1, characterized in that: In step S3, when performing target detection, a multi-scale target detection model is constructed, and the multi-scale target detection model is as follows: ； in, : The set of detected targets; : The overall inference function of the model; : A set of feature maps of the input image; : The weight parameter set of the model; : A set of multi-scale layers, representing the resolution levels of different feature maps; :For the Object detection function of layer-scale feature map; : No. Layer feature maps and corresponding weights.

6. The image processing method based on artificial intelligence according to claim 1, characterized in that: In step S4, local feature matching algorithm is used to align local features of the target area, including: Key point detection; Feature descriptor extraction; Feature matching; Geometric transformation estimation; Target alignment.

7. The image processing method based on artificial intelligence according to claim 1, characterized in that: In step S4, a deep semantic network is used to classify and label the target area according to the semantic information, including: Image description generation; Semantic information embedding; Refine the classification; Object annotation and relationship reasoning; Output labels.

8. The image processing method based on artificial intelligence according to claim 1, characterized in that: In step S5, the image enhancement adopts an adaptive enhancement algorithm, constructs an adaptive enhancement loss function, and adaptively enhances the target area. The adaptive enhancement loss function is as follows: ； in, : Adaptive enhancement total loss; : Number of target areas; : Target area The original contrast value of : Target area The enhanced contrast value of : Target area The original brightness value of : Target area The enhanced brightness value of : For the target area Sharpness measurement function of ; : Weight coefficient, used to balance the optimization of contrast, brightness and sharpness; : Features of the target area after enhancement.

9. The image processing method based on artificial intelligence according to claim 1, characterized in that: In step S5, the image restoration is performed using a generative adversarial network (GAN), which generates and optimizes the image through a discriminator and a generator to ensure the naturalness and authenticity of the target area; The loss function between the discriminator and the generator of the generative adversarial network is designed as follows: ； in, : GAN total loss; :The discriminator is trained on real images judgment; : The image generated by the generator, based on the noise vector ; : Real data distribution; : Noisy data distribution.

10. An artificial intelligence-based image processing system, constructed based on the artificial intelligence-based image processing method according to any one of claims 1 to 9, characterized in that: The functional modules of the system include preliminary feature extraction module, data preprocessing module, target area detection module, target classification and labeling module, image enhancement and restoration module, and image format conversion module.

Images