CN116385308B - Combined image processing optimization strategy selection system - Google Patents

Abstract

The invention relates to a joint image processing optimization strategy selection system, which comprises: the model building device is used for respectively performing learning operation on the convolutional neural network by adopting various reference images so as to obtain an artificial intelligent predictor; and the traversing processing mechanism is used for taking the image to be analyzed as an input image of the artificial intelligent predictor, and traversing various orders of the content filtering processing type, the content sharpening processing type, the content enhancement processing type and the content filtering processing, the content sharpening processing and the content enhancement processing to obtain an image optimization mode with minimum maximum noise amplitude and highest image content definition. The combined image processing optimization strategy selection system is compact in design and wide in application. Because different optimization effects of different image enhancement processing combinations can be calculated by adopting an artificial intelligent model aiming at each frame of picture to be optimized, a best-matching optimization algorithm is obtained for each frame of picture to be optimized.

Description

Combined image processing optimization strategy selection system

Technical Field

The invention relates to the field of image processing, in particular to a joint image processing optimization strategy selection system.

Background

Image enhancement is an important image processing mode for improving the picture quality of an image, making it clear or converting it into a form more suitable for human or machine analysis.

Unlike image restoration, image enhancement does not require faithful reflection of the original image. In contrast, an image that contains some distortion (e.g., a prominent contour) may be more sharp than the original image without distortion. The common image enhancement methods are: (1) gray level histogram processing: the patent technology CN115761638A of Guangzhou polar science and technology Co., ltd.s.A. is an on-line real-time intelligent analysis method and terminal equipment based on image data, the method comprises the following steps: s1, establishing a database; s2, collecting image data; s3, marking the acquired image sample according to a standard; s4, preprocessing an image; s5, performing secondary image processing; s6, performing face recognition on the acquired image data, wherein the image preprocessing comprises gray level histogram processing, interference suppression, edge sharpening and pseudo-color processing, and secondary processing, the brightness of the image is adjusted by adopting an optical module and an image processing module, harmful light is eliminated, the quality of the acquired image is improved, and whether a person in a monitoring area wears a mask or not is detected in real time; real-time video monitoring and early warning on-duty personnel wear safety helmets or not, and if abnormality is detected, immediately giving an alarm to a platform; (2) interference suppression: the random interference superimposed on the image is restrained by low-pass filtering, multi-image averaging, performing certain types of spatial domain operators and other processing; (3) edge sharpening: the patent technology CN115311242a of the photo-electric technology company of the fei-eicose family is an industrial camera image sharpening method, device, equipment and storage medium, which can enhance the contour line of the graph through high-pass filtering, differential operation or some transformation, and comprises the following steps: sharpening the edge weight of the enhanced image of the input original image to obtain a preliminary sharpening layer for highlighting the edge of the image and an edge layer for reflecting the edge of the image; the method comprises the steps of giving a fixed value of a central pixel of an area to an edge layer for noise removal according to the gray level change degree of the pixel of the area to obtain a denoising edge layer; and the FPGA performs weighted fusion on the noise-reduced sharpening weight processing on the noise-reduced edge image layer and the primary sharpening image layer to form a sharpened image which is sharpened only for the edge. According to the method, the edge judgment is utilized to remove isolated noise points, so that poisson noise and Gaussian noise are removed; the influence of noise on the sharpening result is further reduced through weighted fusion, and the image sharpening effect is improved; (4) pseudo-color processing: the patent technology CN113436110a of Xiamen university is a method for pseudo-color processing of synthetic aperture radar gray-scale images, which comprises the following steps: 1) Filtering preprocessing is carried out on the input SAR image so as to inhibit speckle noise; 2) Carrying out sectional coding on the gray value of the SAR image; 3) RGB band combining is performed to form a pseudo-color enhanced image. The image after pseudo-color enhancement contains rich color information, and has outstanding details, clear textures and good visual effects on land and sea. The invention is similar to the time used for rainbow code pseudo-color processing, and is approximately 4 times faster than the pixel self-conversion method. The calculation efficiency is high, and the method is simple and practical. In addition, the SAR image can be enhanced in pseudo color, is suitable for other gray scale images with narrow gray scale range and large noise, and has good universality. The gray level image is processed by pseudo color to enrich and highlight useful and detailed information of the ground object, so that visual interpretation is facilitated. .

However, since each frame of picture content to be subjected to image enhancement is different, the noise distribution state and other performances are different, when a plurality of different optimization modes are adopted, such as image sharpening, image filtering and image enhancement, various image optimization strategies with various types being selectable and various image processing modes being sequentially variable exist, which image optimization strategy is most suitable for the current picture content cannot be determined, so that the finally adopted image optimization strategy cannot achieve the requirement of the optimal optimization effect, and if the processing of each frame of picture content is performed with various image optimization strategies, the comparison of the processed picture quality is obviously impractical.

Disclosure of Invention

In order to solve the technical problems in the related art, the invention provides a joint image processing optimization strategy selection system, which comprises:

a content storage device for storing various reference images, which have the same resolution and different contents;

The network establishing device is used for establishing a convolutional neural network for executing image optimization effect analysis after multi-layer image enhancement, the convolutional neural network takes the maximum noise amplitude and the image content definition of an output image obtained after three-layer optimization processing of the input image in sequence based on the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number as two output information, wherein the hue component value, the brightness component value and the saturation component value of the input image are respectively corresponding to each pixel point in an HSB color space, the optimization algorithm identification based on the content filtering processing, the optimization algorithm identification based on the content sharpening processing, the optimization algorithm identification based on the content enhancement processing sequence number, the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number are all input information, and the resolution of the input image is the same as that of each reference image;

The model building device is connected with the network building device and is used for respectively executing learning operation on the convolutional neural network by adopting various reference images so as to obtain the convolutional neural network after the completion of the learning operation and output the convolutional neural network as an artificial intelligent predictor;

a traversal processing mechanism connected with the model construction device, for taking the image to be analyzed, which has the same resolution as each reference image, as an input image of the artificial intelligent predictor, and traversing the content filtering processing type, the content sharpening processing type, the content enhancement processing type and various sequences in which the content filtering processing, the content sharpening processing and the content enhancement processing are executed so as to obtain an image optimization mode with minimum maximum noise amplitude and highest image content definition;

Wherein for an image to be analyzed having the same resolution as each reference image, taking it as an input image of the artificial intelligence predictor, simultaneously traversing various orders in which a content filtering process type, a content sharpening process type, a content enhancement process type, and a content filtering process, a content sharpening process, and a content enhancement process are performed to obtain an image optimization mode in which a maximum noise amplitude is minimum and an image content definition is highest, comprises: and outputting the content filtering processing type, the content sharpening processing type, the content enhancement processing type and the sequence in which the content filtering processing, the content sharpening processing and the content enhancement processing are executed corresponding to the output image as an image optimization mode when the maximum noise amplitude of the output image corresponding to the image to be analyzed is minimum and the definition of the image content is highest.

The combined image processing optimization strategy selection system can calculate different optimization effects of different image enhancement processing combinations by adopting an artificial intelligent model aiming at each frame of picture to be optimized, and the different orders of the image enhancement processing in the image enhancement processing combinations also cause the difference of the optimization effects, so that the fine analysis of the optimal optimization strategy is realized with smaller operation cost.

Because different optimization effects of different image enhancement processing combinations can be calculated by adopting an artificial intelligent model aiming at each frame of picture to be optimized, a best-matching optimization algorithm is obtained for each frame of picture to be optimized.

Drawings

Embodiments of the present invention will be described below with reference to the accompanying drawings, in which:

Fig. 1 is a block diagram showing the structure of a joint image processing optimization strategy selection system according to an embodiment of the present invention.

Fig. 2 is a block diagram showing the structure of a joint image processing optimization strategy selection system according to the B embodiment of the present invention.

Fig. 3 is a block diagram showing the structure of a joint image processing optimization strategy selection system according to the C embodiment of the present invention.

Detailed Description

Embodiments of the joint image processing optimization strategy selection system of the present invention will be described in detail below with reference to the accompanying drawings.

Embodiment A

Fig. 1 is a block diagram showing a structure of a joint image processing optimization strategy selection system according to an embodiment of the present invention, the system including:

a content storage device for storing various reference images, which have the same resolution and different contents;

The network establishing device is used for establishing a convolutional neural network for executing image optimization effect analysis after multi-layer image enhancement, the convolutional neural network takes the maximum noise amplitude and the image content definition of an output image obtained after three-layer optimization processing of the input image in sequence based on the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number as two output information, wherein the hue component value, the brightness component value and the saturation component value of the input image are respectively corresponding to each pixel point in an HSB color space, the optimization algorithm identification based on the content filtering processing, the optimization algorithm identification based on the content sharpening processing, the optimization algorithm identification based on the content enhancement processing sequence number, the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number are all input information, and the resolution of the input image is the same as that of each reference image;

The model building device is connected with the network building device and is used for respectively executing learning operation on the convolutional neural network by adopting various reference images so as to obtain the convolutional neural network after the completion of the learning operation and output the convolutional neural network as an artificial intelligent predictor;

a traversal processing mechanism connected with the model construction device, for taking the image to be analyzed, which has the same resolution as each reference image, as an input image of the artificial intelligent predictor, and traversing the content filtering processing type, the content sharpening processing type, the content enhancement processing type and various sequences in which the content filtering processing, the content sharpening processing and the content enhancement processing are executed so as to obtain an image optimization mode with minimum maximum noise amplitude and highest image content definition;

Wherein for an image to be analyzed having the same resolution as each reference image, taking it as an input image of the artificial intelligence predictor, simultaneously traversing various orders in which a content filtering process type, a content sharpening process type, a content enhancement process type, and a content filtering process, a content sharpening process, and a content enhancement process are performed to obtain an image optimization mode in which a maximum noise amplitude is minimum and an image content definition is highest, comprises: when the maximum noise amplitude of the output image corresponding to the image to be analyzed is minimum and the definition of the image content is highest, outputting the content filtering processing type, the content sharpening processing type, the content enhancement processing type and the sequence in which the content filtering processing, the content sharpening processing and the content enhancement processing are executed corresponding to the output image as an image optimization mode;

Wherein for an image to be analyzed having the same resolution as each reference image, taking it as an input image of the artificial intelligence predictor, simultaneously traversing various orders in which a content filtering process type, a content sharpening process type, a content enhancement process type, and a content filtering process, a content sharpening process, and a content enhancement process are performed to obtain an image optimization mode in which a maximum noise amplitude is minimum and an image content definition is highest, comprises: when the image optimization mode is acquired, the priority with the smallest maximum noise amplitude is smaller than the priority with the highest definition of the image content;

The method for obtaining the convolutional neural network after the learning operation is completed by adopting various reference images to respectively execute the learning operation on the convolutional neural network, and the method for obtaining the convolutional neural network after the learning operation is completed and outputting the convolutional neural network as an artificial intelligent predictor comprises the following steps: the number of times that learning has been performed when the learning operation is completed is positively correlated with the resolution of each reference image.

B embodiment

Fig. 2 is a block diagram showing the structure of a joint image processing optimization strategy selection system according to an embodiment B of the present invention, including:

a content storage device for storing various reference images, which have the same resolution and different contents;

The network establishing device is used for establishing a convolutional neural network for executing image optimization effect analysis after multi-layer image enhancement, the convolutional neural network takes the maximum noise amplitude and the image content definition of an output image obtained after three-layer optimization processing of the input image in sequence based on the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number as two output information, wherein the hue component value, the brightness component value and the saturation component value of the input image are respectively corresponding to each pixel point in an HSB color space, the optimization algorithm identification based on the content filtering processing, the optimization algorithm identification based on the content sharpening processing, the optimization algorithm identification based on the content enhancement processing sequence number, the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number are all input information, and the resolution of the input image is the same as that of each reference image;

The model building device is connected with the network building device and is used for respectively executing learning operation on the convolutional neural network by adopting various reference images so as to obtain the convolutional neural network after the completion of the learning operation and output the convolutional neural network as an artificial intelligent predictor;

a traversal processing mechanism connected with the model construction device, for taking the image to be analyzed, which has the same resolution as each reference image, as an input image of the artificial intelligent predictor, and traversing the content filtering processing type, the content sharpening processing type, the content enhancement processing type and various sequences in which the content filtering processing, the content sharpening processing and the content enhancement processing are executed so as to obtain an image optimization mode with minimum maximum noise amplitude and highest image content definition;

The identification storage mechanism is used for storing the optimization algorithm identification of each type of content filtering processing, the optimization algorithm identification of each type of content sharpening processing and the optimization algorithm identification of each type of content enhancement processing;

Wherein the respective types of content filtering processing include a guide filtering processing, an edge preserving smoothing filtering processing, a combination filtering processing, FRANGI filtering processing, and a direction filtering processing;

The content sharpening processing of each type comprises horizontal sharpening processing, vertical sharpening processing, kirsch operator sharpening processing and Roberts operator sharpening processing;

Wherein the respective types of content enhancement processing include histogram equalization processing, logarithmic image enhancement processing, and exponential image enhancement processing.

C embodiment

Fig. 3 is a block diagram showing the structure of a joint image processing optimization strategy selection system according to the embodiment of the present invention, including:

a content storage device for storing various reference images, which have the same resolution and different contents;

The network establishing device is used for establishing a convolutional neural network for executing image optimization effect analysis after multi-layer image enhancement, the convolutional neural network takes the maximum noise amplitude and the image content definition of an output image obtained after three-layer optimization processing of the input image in sequence based on the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number as two output information, wherein the hue component value, the brightness component value and the saturation component value of the input image are respectively corresponding to each pixel point in an HSB color space, the optimization algorithm identification based on the content filtering processing, the optimization algorithm identification based on the content sharpening processing, the optimization algorithm identification based on the content enhancement processing sequence number, the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number are all input information, and the resolution of the input image is the same as that of each reference image;

The model building device is connected with the network building device and is used for respectively executing learning operation on the convolutional neural network by adopting various reference images so as to obtain the convolutional neural network after the completion of the learning operation and output the convolutional neural network as an artificial intelligent predictor;

a traversal processing mechanism connected with the model construction device, for taking the image to be analyzed, which has the same resolution as each reference image, as an input image of the artificial intelligent predictor, and traversing the content filtering processing type, the content sharpening processing type, the content enhancement processing type and various sequences in which the content filtering processing, the content sharpening processing and the content enhancement processing are executed so as to obtain an image optimization mode with minimum maximum noise amplitude and highest image content definition;

and the model storage mechanism is connected with the model construction device and used for storing various model parameters of the artificial intelligent predictor so as to realize the storage of the artificial intelligent predictor.

Next, a further explanation of the specific structure of the joint image processing optimization policy selection system of the present invention will be continued.

In a joint image processing optimization strategy selection system according to various embodiments of the present invention:

Respectively performing learning operation on the convolutional neural network by adopting various reference images to obtain the convolutional neural network after the learning operation is completed and outputting the convolutional neural network as an artificial intelligent predictor, wherein the learning operation comprises the following steps of: performing a plurality of learning operations on the convolutional neural network using each reference image, each learning operation differing in at least one of a content filtering process-based optimization algorithm identification, a content sharpening process-based optimization algorithm identification, a content enhancement process-based optimization algorithm identification, a content filtering process sequence number, a content sharpening process sequence number, and a content enhancement process sequence number;

Wherein, a plurality of learning operations are performed on the convolutional neural network using each reference image, each learning operation differing in at least one of an optimization algorithm identification based on content filtering processing, an optimization algorithm identification based on content sharpening processing, an optimization algorithm identification based on content enhancement processing, a content filtering processing sequence number, a content sharpening processing sequence number, and a content enhancement processing sequence number, including: when the optimizing algorithm identification based on the content filtering process, the optimizing algorithm identification based on the content sharpening process and the optimizing algorithm identification based on the content enhancement process are the same, the order in which the content filtering process, the content sharpening process and the content enhancement process are executed is different, and the corresponding learning operation is different;

Wherein, a plurality of learning operations are performed on the convolutional neural network using each reference image, each learning operation differing in at least one of an optimization algorithm identification based on content filtering processing, an optimization algorithm identification based on content sharpening processing, an optimization algorithm identification based on content enhancement processing, a content filtering processing sequence number, a content sharpening processing sequence number, and a content enhancement processing sequence number, including: when the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number are the same, the content filtering processing type, the content sharpening processing type and/or the content enhancement processing type are different, and the corresponding learning operation is different.

And in a joint image processing optimization strategy selection system according to various embodiments of the present invention:

The maximum noise amplitude and the image content definition of the output image obtained after three-layer optimization processing based on the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number, which are performed sequentially on the input image, are used as two items of output information, including: the processing with the forefront sequence number in the sequence number of the content filtering processing, the sequence number of the content sharpening processing and the sequence number of the content enhancement processing is used as the first layer of optimization processing executed on the input image;

The method for processing the input image by using the content filter processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number as two output information comprises the following steps: the processing of centering the sequence number in the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number is used as second-layer optimization processing executed on the input image;

The method for processing the input image by using the content filter processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number as two output information comprises the following steps: and the processing with the last sequence number in the sequence numbers of the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number is used as the third layer optimization processing executed on the input image.

In addition, in the joint image processing optimization policy selection system, the convolutional neural network uses respective hue component values, respective brightness component values and respective saturation component values, an optimization algorithm identifier based on content filtering processing, an optimization algorithm identifier based on content sharpening processing, an optimization algorithm identifier based on content enhancement processing, a content filtering processing sequence number, a content sharpening processing sequence number and a content enhancement processing sequence number, which correspond to respective pixel points of an input image in an HSB color space, as two output information, and uses a maximum noise amplitude and an image content sharpness of an output image obtained after performing sequential three-layer optimization processing on the input image based on the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number, wherein the input image and the resolution of each reference image are the same, and the two output information comprise: the optimizing algorithm identification based on the content filtering process, the optimizing algorithm identification based on the content sharpening process, the optimizing algorithm identification based on the content enhancement process, the content filtering process sequence number, the content sharpening process sequence number and the content enhancement process sequence number are respectively represented by adopting different binary number codes.

The technical scheme of the invention has the technical effects that:

(1) Establishing an optimization effect prediction model of an input image under the same resolution under a plurality of different image optimization types and different optimization sequences, so as to traverse various optimization types and various optimization sequences aiming at any image with the same resolution to obtain an optimization strategy of the optimal optimization effect, thereby being capable of obtaining reliable optimized image quality under extremely low operation cost for various image contents;

(2) The specific model is a convolutional neural network after learning, the convolutional neural network takes the values of all hue components, brightness components and saturation components of an input image corresponding to all pixel points in an HSB color space, an optimization algorithm identifier based on content filtering processing, an optimization algorithm identifier based on content sharpening processing, an optimization algorithm identifier based on content enhancement processing, a content filtering processing sequence number, a content sharpening processing sequence number and a content enhancement processing sequence number as various input information, and takes the maximum noise amplitude and the image content definition of an output image obtained after three-layer optimization processing based on the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number which are performed on the input image as two output information, and simultaneously applies a targeted learning mode, thereby improving the stability and the effectiveness of a model prediction result;

(3) When the optimal optimization strategy is obtained, the priority with the smallest maximum noise amplitude is smaller than the priority with the highest definition of the image content.

The foregoing description of the exemplary embodiments of the invention has been presented for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form disclosed. Obviously, many modifications and variations will be apparent to practitioners skilled in the art. The exemplary embodiments were chosen and described in order to best explain the principles of the invention and its practical application, thereby enabling others skilled in the art to understand the invention for various embodiments and with various modifications as are suited to the particular use contemplated. It is intended that the scope of the invention be defined by the following claims and their equivalents.

Claims (9)

1. A joint image processing optimization strategy selection system, the system comprising:

a content storage device for storing various reference images, which have the same resolution and different contents;

The network establishing device is used for establishing a convolutional neural network for executing image optimization effect analysis after multi-layer image enhancement, the convolutional neural network takes the maximum noise amplitude and the image content definition of an output image obtained after three-layer optimization processing of the input image in sequence based on the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number as two output information, wherein the hue component value, the brightness component value and the saturation component value of the input image are respectively corresponding to each pixel point in an HSB color space, the optimization algorithm identification based on the content filtering processing, the optimization algorithm identification based on the content sharpening processing, the optimization algorithm identification based on the content enhancement processing sequence number, the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number are all input information, and the resolution of the input image is the same as that of each reference image;

The content filtering processing-based optimization algorithm identification, the content sharpening processing-based optimization algorithm identification, the content enhancement processing-based optimization algorithm identification, the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number are respectively represented by adopting different binary number value codes;

The model building device is connected with the network building device and is used for respectively executing learning operation on the convolutional neural network by adopting various reference images so as to obtain the convolutional neural network after the completion of the learning operation and output the convolutional neural network as an artificial intelligent predictor;

a traversal processing mechanism connected with the model construction device, for taking the image to be analyzed, which has the same resolution as each reference image, as an input image of the artificial intelligent predictor, and traversing the content filtering processing type, the content sharpening processing type, the content enhancement processing type and various sequences in which the content filtering processing, the content sharpening processing and the content enhancement processing are executed so as to obtain an image optimization mode with minimum maximum noise amplitude and highest image content definition;

Wherein for an image to be analyzed having the same resolution as each reference image, taking it as an input image of the artificial intelligence predictor, simultaneously traversing various orders in which a content filtering process type, a content sharpening process type, a content enhancement process type, and a content filtering process, a content sharpening process, and a content enhancement process are performed to obtain an image optimization mode in which a maximum noise amplitude is minimum and an image content definition is highest, comprises: when the maximum noise amplitude of the output image corresponding to the image to be analyzed is minimum and the definition of the image content is highest, outputting the content filtering processing type, the content sharpening processing type, the content enhancement processing type and the sequence in which the content filtering processing, the content sharpening processing and the content enhancement processing are executed corresponding to the output image as an image optimization mode;

and the identification storage mechanism is used for storing the optimization algorithm identification of each type of content filtering process, the optimization algorithm identification of each type of content sharpening process and the optimization algorithm identification of each type of content enhancement process.

2. The joint image processing optimization strategy selection system of claim 1, wherein:

For an image to be analyzed having the same resolution as each reference image, as an input image of the artificial intelligence predictor, traversing various orders in which the content filtering process type, the content sharpening process type, the content enhancing process type, and the content filtering process, the content sharpening process, and the content enhancing process are performed to obtain an image optimization mode in which the maximum noise amplitude is minimum and the image content sharpness is highest, includes: when the image optimization mode is acquired, the priority with the smallest maximum noise amplitude is smaller than the priority with the highest definition of the image content;

The method for obtaining the convolutional neural network after the learning operation is completed by adopting various reference images to respectively execute the learning operation on the convolutional neural network, and the method for obtaining the convolutional neural network after the learning operation is completed and outputting the convolutional neural network as an artificial intelligent predictor comprises the following steps: the number of times that learning has been performed when the learning operation is completed is positively correlated with the resolution of each reference image.

3. The joint image processing optimization strategy selection system of claim 1, wherein:

The various types of content filtering processing comprise guiding filtering processing, edge preserving smoothing filtering processing, combined filtering processing, FRANGI filtering processing and direction filtering processing;

The content sharpening processing of each type comprises horizontal sharpening processing, vertical sharpening processing, kirsch operator sharpening processing and Roberts operator sharpening processing;

Wherein the respective types of content enhancement processing include histogram equalization processing, logarithmic image enhancement processing, and exponential image enhancement processing.

4. The joint image processing optimization strategy selection system of claim 2, wherein said system further comprises:

and the model storage mechanism is connected with the model construction device and used for storing various model parameters of the artificial intelligent predictor so as to realize the storage of the artificial intelligent predictor.

5. A joint image processing optimization strategy selection system as claimed in any of claims 1-4 wherein:

Respectively performing learning operation on the convolutional neural network by adopting various reference images to obtain the convolutional neural network after the learning operation is completed and outputting the convolutional neural network as an artificial intelligent predictor, wherein the learning operation comprises the following steps of: and performing a plurality of learning operations on the convolutional neural network by using each reference image, wherein each learning operation is different in at least one of an optimization algorithm identification based on content filtering processing, an optimization algorithm identification based on content sharpening processing, an optimization algorithm identification based on content enhancement processing, a content filtering processing sequence number, a content sharpening processing sequence number and a content enhancement processing sequence number.

6. The joint image processing optimization strategy selection system of claim 5, wherein:

Performing a plurality of learning operations on the convolutional neural network using each reference image, each learning operation differing in at least one of a content filtering process-based optimization algorithm identification, a content sharpening process-based optimization algorithm identification, a content enhancement process-based optimization algorithm identification, a content filtering process sequence number, a content sharpening process sequence number, and a content enhancement process sequence number, including: when the content filtering process-based optimization algorithm identification, the content sharpening process-based optimization algorithm identification and the content enhancement process-based optimization algorithm identification are the same, the order in which the content filtering process, the content sharpening process and the content enhancement process are executed is different, and the corresponding learning operation is different.

7. The joint image processing optimization strategy selection system of claim 6, wherein:

Performing a plurality of learning operations on the convolutional neural network using each reference image, each learning operation differing in at least one of a content filtering process-based optimization algorithm identification, a content sharpening process-based optimization algorithm identification, a content enhancement process-based optimization algorithm identification, a content filtering process sequence number, a content sharpening process sequence number, and a content enhancement process sequence number, including: when the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number are the same, the content filtering processing type, the content sharpening processing type and/or the content enhancement processing type are different, and the corresponding learning operation is different.

8. A joint image processing optimization strategy selection system as claimed in any of claims 1-4 wherein:

The maximum noise amplitude and the image content definition of the output image obtained after three-layer optimization processing based on the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number, which are performed sequentially on the input image, are used as two items of output information, including: and the processing with the forefront sequence number in the sequence number of the content filtering processing, the sequence number of the content sharpening processing and the sequence number of the content enhancement processing is used as the first layer optimization processing executed on the input image.

9. The joint image processing optimization strategy selection system of claim 8, wherein:

The maximum noise amplitude and the image content definition of the output image obtained after three-layer optimization processing based on the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number, which are performed sequentially on the input image, are used as two items of output information, including: the processing of centering the sequence number in the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number is used as second-layer optimization processing executed on the input image;

The method for processing the input image by using the content filter processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number as two output information comprises the following steps: and the processing with the last sequence number in the sequence numbers of the content filtering processing sequence number, the content sharpening processing sequence number and the content enhancement processing sequence number is used as the third layer optimization processing executed on the input image.