An AI-based intelligent image processing method and system

Abstract

The present invention belongs to the technical field of image processing, and specifically relates to an AI-based intelligent image processing method and system, which includes the following steps: obtaining image data, segmenting the image through an AI chip, obtaining upper-layer elements and lower-layer elements in the image, and respectively processing the upper-layer elements and the lower-layer elements into multiple image data sets and corresponding set association trees. The image data sets of the upper-layer elements and the image data sets of the lower-layer elements form a set association tree, and the set association tree includes multiple groups; based on the AI chip, the upper-layer elements and the lower-layer elements are repaired, and the upper-layer elements and the lower-layer elements are judged and classified; based on the AI algorithm, the upper-layer elements and the lower-layer elements are respectively processed, a 3D model is established, the upper-layer element positions and the lower-layer elements are synthesized, and dynamic display is realized. The present invention can reduce the hardware cost of image processing equipment and improve the processing efficiency of image processing.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and particularly relates to an AI-based intelligent image processing method and system. Background Art

[0002] Image processing technology is a technology that uses a computer to analyze and process images to achieve the desired results. It usually refers to digital image processing, that is, using a computer to process digital image data. Common image processing technologies include image matting, image enhancement, and image restoration.

[0003] Problems existing in the prior art:

[0004] When generating a three-dimensional image from a two-dimensional image, the image is processed by an AI chip and a GPU. In the prior art, the image is processed as a whole at one time. Since the AI chip needs to identify and process the elements in the image when recognizing the image, it often causes insufficient computing power. Although the operation mode can be changed by hardware, it also requires increasing the hardware cost to achieve split processing, and the computing power also needs to be improved. Especially for small and medium-sized enterprises or advertising companies, cost control needs to be more emphasized. When generating a three-dimensional image, for the AI chip and the GPU, during the entire image processing process, the device often crashes, which affects the user experience. And for the split image, after processing, the correlation between image elements decreases, and the user experience for specific application scenarios becomes worse. Summary of the Invention

[0005] The purpose of the present invention is to provide an AI-based intelligent image processing method and system, which can reduce the hardware cost of image processing equipment and improve the processing efficiency of image processing.

[0006] The technical solution adopted by the present invention is specifically as follows:

[0007] An AI-based intelligent image processing method includes the following steps:

[0008] Obtain image data, and segment the image through an AI chip, and obtain upper-layer elements and lower-layer elements in the image, and process the upper-layer elements and the lower-layer elements into multiple image data sets and corresponding set association trees respectively. The image data set of the upper-layer elements and the image data set of the lower-layer elements form a set association tree, and the set association tree includes multiple groups;

[0009] Based on the AI chip, repair the upper-layer elements and the lower-layer elements, and judge and classify the upper-layer elements and the lower-layer elements;

[0010] Based on the AI algorithm, process the upper-layer elements and the lower-layer elements respectively, establish a 3D model, synthesize the upper-layer element positions and the lower-layer elements, and realize dynamic display.

[0011] The steps for the AI chip to segment an image include using a deep learning algorithm to identify and segment objects in the image, and thus constructing upper-layer elements and lower-layer elements based on the objects in the segmented image according to depth information and visual hierarchy. The specific steps are as follows:

[0012] S1: Hierarchically identify upper-layer elements according to the integrity of visual objects, extract multiple sub-elements, store the sub-elements in independent sub-element file memories respectively, and attach coordinate information;

[0013] S2: According to depth information, incorporate the information between the positions of each sub-element and the lower-layer elements into the coordinate information to form a coordinate information set;

[0014] S3: Sequentially identify down to the lower-layer elements from step S1 and step S2, establish a data coordinate axis with the lower-layer elements, and modify the coordinate sets of each sub-element in the corresponding upper-layer elements;

[0015] S4: Establish a relationship tree to associate data sets through the coordinate sets between the upper-layer elements and the lower-layer elements.

[0016] The set association tree establishes an image data association set in a tree structure by associating each group of relationship tree associated data sets, distributes each sub-element to each corresponding image processor, and collaborates with the AI chip to implement the processing of the image and sub-elements.

[0017] The method for the image processor to collaborate with the AI chip to implement the processing of the image and sub-elements includes the following steps:

[0018] Obtain sub-elements and lower-layer elements;

[0019] Through the image processor collaborating with the AI chip, repair the sub-elements and lower-layer elements to improve clarity and reduce noise;

[0020] Meanwhile, perform feature recognition on the sub-elements and lower-layer elements;

[0021] Supplement and restore the missing parts of the sub-elements and lower-layer elements;

[0022] Restore the 2D sub-elements and lower-layer elements to a 3D model through 3D model synthesis technology;

[0023] Attach a coordinate set to the generated 3D model to establish a relationship tree to associate data sets.

[0024] A three-dimensional motion coordinate system is established with the center of the generated 3D model, a three-dimensional motion coordinate system is established with motion nodes, and a time-axis coordinate system starting from the above-mentioned motion coordinate system. For the motion speed coordinate system corresponding to the time-axis coordinate system corresponding to the above-mentioned motion coordinate system, through the time correlation formula of the motion speed in the motion coordinate system, it is used to control the change of the motion speed during the motion process;

[0025] A motion vector is set for the above-mentioned change in motion speed to control the motion direction.

[0026] The set association tree established between the upper-layer elements and the lower-layer elements is used to establish a motion base point with the sub-elements as the origin. A restricted area is set between each sub-element of the lower-layer elements and the upper-layer elements. Through the intelligent motion method realized by the AI chip's intelligent operation in the restricted area, the motion mode of the sub-elements to avoid the restricted area is realized.

[0027] The motion method for the sub-elements to avoid the restricted area includes the following steps:

[0028] Obtain the position coordinate information of the restricted area and the sub-elements;

[0029] Identify the sub-elements, measure the volume and the range of motion states, and set the motion mode and activity mode of the sub-elements;

[0030] Judge or set whether the restricted area is dynamically changing;

[0031] By setting the motion mode between the sub-elements and the restricted area, the motion mode includes any one of: avoidance motion, circumferential motion, or acting motion;

[0032] Decide the motion type;

[0033] Plan the decision motion path;

[0034] Execute the decision motion instruction;

[0035] Simulate the execution instruction.

[0036] An AI-based intelligent image processing system includes:

[0037] Image acquisition module: responsible for directly acquiring, synthesizing, or downloading images, as well as the data transmission process of images;

[0038] Image preprocessing module: correct the noise existing in the images and perform color correction on the images;

[0039] AI chip and deep learning module: process and preprocess the images through the AI chip, and record and learn the recognition methods and processing methods through the deep learning module, and realize deep learning and optimization through the dynamic motion display method;

[0040] Image processor: namely GPU, which is used to generate 3D images from 2D images, render 2D and 3D images, and process the motion data operations and real-time rendering of the motion states of sub-elements, upper-layer elements and lower-layer elements;

[0041] CPU, which is used to adjust computing power and allocate work among the image processor, AI chip and deep learning module;

[0042] Motion control and dynamic display module: Based on the computing power of the image processor, CPU, AI chip and deep learning module, it is responsible for processing the dynamic control of sub-elements in 2D or 3D space and storing the dynamic motion formulas and algorithms to be called;

[0043] Storage module;

[0044] Interaction module: including input and output modules.

[0045] It also includes a dynamic simulation display module, a 3D printing module and an output module.

[0046] An AI-based intelligent image processing terminal, including;

[0047] At least one terminal for running an intelligent image processing system;

[0048] One or more terminals running the intelligent image processing system cooperate to execute the intelligent image processing method.

[0049] The technical effects achieved by the present invention are:

[0050] In the present invention, by realizing the splitting of images and independently processing sub-elements and upper-layer elements, the requirements for computing power of the AI chip and GPU can be reduced, the hardware configuration can be reduced, and thus the cost can be reduced, and the image processing efficiency can be improved.

[0051] In the present invention, by establishing a coordinate system for the 3D model and realizing the dynamic control of the 3D model through the dynamic motion of the coordinate system, when it is necessary to change the dynamic display effect, by changing the dynamic display operation data, non-technical personnel can quickly learn and get started, and the operation method is simple. Brief Description of the Drawings

[0052] Figure 1 It is a schematic flowchart of the method of the present invention;

[0053] Figure 2 It is a schematic diagram of image processing of the method of the present invention;

[0054] Figure 3 It is a schematic structural diagram of the intelligent image processing system in the present invention. Detailed implementation manners

[0055] In order to make the objectives and advantages of the present invention clearer and more understandable, the present invention will be specifically described below in conjunction with embodiments. It should be understood that the following text is only used to describe one or several specific implementation manners of the present invention, and does not strictly limit the scope of protection specifically claimed by the present invention.

[0056] For the coordinate information, coordinate system, coordinate set, and relationship tree associated data set in the embodiments;

[0057] In the embodiments, the coordinate system is an independent coordinate system centered on the center of the sub-element or the motion node;

[0058] The coordinate information includes the independent coordinate information corresponding to the upper-layer element and the sub-element located on the lower-layer element;

[0059] The coordinate set includes an independent coordinate set based on the coordinate system or coordinate information, or a coordinate set based on multiple coordinate systems and coordinate information;

[0060] The relationship tree associated data set is an associated data set between any of the sub-elements and the upper-layer elements and the lower-layer elements.

[0061] It should be noted that the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0062] According to an embodiment of the present invention, an embodiment of an AI-based intelligent image processing method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system including at least one set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0063] This method embodiment can also be executed in an electronic device including a memory and a processor, a similar control device, or the cloud. Taking the electronic device as an example, the electronic device may include one or more processors and a memory for storing data. Optionally, the above-mentioned electronic device may further include a communication device for communication functions and a display device. Those of ordinary skill in the art can understand that the above structural description is only illustrative and does not limit the structure of the above-mentioned electronic device. For example, the electronic device may further include more or fewer components than the above structural description, or have a configuration different from the above structural description.

[0064] The processor may include one or more processing units. For example, the processor may include a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processing (DSP) chip, a microcontroller unit (MCU), a field-programmable gate array (FPGA), a neural-network processing unit (NPU), a tensor processing unit (TPU), an artificial intelligent (AI) type processor, and other processing devices. Among them, different processing units may be independent components or integrated in one or more processors. In some instances, the electronic device may also include one or more processors.

[0065] The memory can be used to store computer programs. For example, it stores the computer program corresponding to the power equipment fire recognition method in the embodiments of the present invention. The processor realizes the above-mentioned power equipment fire recognition method by running the computer program stored in the memory. The memory may include a high-speed random access memory and may also include a non-volatile memory, such as one or more magnetic storage devices, a flash memory, or other non-volatile solid-state memories. In some instances, the memory may further include a memory remotely set relative to the processor, and these remote memories can be connected to the electronic device through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and their combinations.

[0066] The display device can be a touch-screen liquid crystal display (LCD) and a touch display (also known as a "touch screen" or "touch display screen"). The liquid crystal display enables the user to interact with the user interface of the electronic device. In some embodiments, the above electronic device has a graphical user interface (GUI), and the user can perform human-computer interaction with the GUI by touching and / or gestures of fingers on the touch-sensitive surface. The executable instructions for performing the above human-computer interaction functions are configured / stored in a computer program product or a readable storage medium executable by one or more processors.

[0067] As Figure 1 and Figure 2 shown, an AI-based intelligent image processing method includes the following steps:

[0068] Obtain image data, segment the image through an AI chip, obtain the upper-layer elements and lower-layer elements in the image, and process the upper-layer elements and lower-layer elements into multiple image data sets and corresponding set association trees respectively. The image data sets of the upper-layer elements and the image data sets of the lower-layer elements form set association trees, and there are multiple groups of set association trees;

[0069] Based on the AI chip, complete the repair of the upper-layer elements and lower-layer elements, and judge and classify the upper-layer elements and lower-layer elements;

[0070] Based on the AI algorithm, process the upper-layer elements and lower-layer elements separately, build a 3D model, synthesize the upper-layer element position and the lower-layer element, and achieve dynamic display.

[0071] Example 1, the above-mentioned dynamic display is as follows:

[0072] For example, in the image, there are combined characters and background objects. By independently splitting multiple characters, three-dimensional character images are generated respectively and dynamically displayed in the background. The display methods include dynamic rotation display of characters, establishing a three-dimensional space coordinate system according to the limbs of the characters, and forming a partial relationship tree association data set with the coordinate information established in the background. Through the association display between the characters and the coordinate information, the association display between the characters and the lower-layer element background can be realized.

[0073] When the set association tree established between the upper-layer elements and the lower-layer elements is used to establish a motion base point with the sub-elements as the origin, set a restricted area between the lower-layer elements and each sub-element in the upper-layer elements, and an intelligent motion method realized by the intelligent operation of the AI chip in the restricted area to realize the motion mode of the sub-elements avoiding the restricted area.

[0074] Preferably, the motion method for the sub-elements to avoid the restricted area includes the following steps:

[0075] Obtain the position coordinate information of the restricted area and the sub-elements;

[0076] Identify the sub-elements, measure the volume and the activity range of the motion state, and set the motion mode and activity mode of the sub-elements;

[0077] Judge or set whether the restricted area is dynamically changing;

[0078] By setting the motion mode between the sub-elements and the restricted area, the motion mode includes any one of evasive motion, circular motion or acting motion;

[0079] Decide the motion type, and decide the motion type of the character according to the set motion mode;

[0080] Plan the decision-making movement path, determine the movement type by the position of the restricted area or the route that can pass through the restricted area, and can be applied to dynamic advertising or dynamic poster display, the movement development of dynamic image 3D games, the display of scientific and technological effects in scientific and technological achievements, etc.;

[0081] Execute the decision-making movement instruction. After determining the movement type, execute the movement decision-making instruction;

[0082] Simulate the operation instruction, display the effect, fix the problems in the corresponding movement, and achieve a better display method by modifying the parameters of the coordinate system and coordinate information.

[0083] Among them, the method for planning the decision-making movement path includes the following steps:

[0084] Path calculation: First, calculate the path available for the sub-elements to move according to the position and shape of the restricted area, which is achieved through algorithms such as the A* algorithm, Dijkstra algorithm or any other applicable path planning algorithm. These algorithms can calculate the best path from the starting point to the end point considering all dynamic and static obstacles, or set the movement path manually;

[0085] Dynamic adjustment: If the restricted area is dynamically changing (such as the area size, shape or position changing over time), then the path planning needs to be updated in real time. The system can receive real-time data such as the coordinate information of the restricted area on the lower-level elements, the area size, the movement speed and the movement path, and quickly recalculate the path to adapt to the changes in the environment;

[0086] Multiple alternative plans: When planning the path, multiple alternative paths need to be prepared. When the main path becomes infeasible due to some unforeseen events (such as suddenly emerging new obstacles), the system can quickly switch to the alternative path, such as setting the movement path manually;

[0087] Smoothing processing: In order to make the movement of the animation or game character more natural, the path needs to be smoothed, which includes removing sharp corners, optimizing the path curve, ensuring that the changes in movement speed and direction are continuous and smooth. Specifically, it is necessary to adjust the coordinate system of the sub-elements, upper-level elements or characters. The coordinate system adjustment methods include three-axis movement, as well as the tilt angle and acceleration of the three axes during the movement;

[0088] In the above steps: By adjusting the movement mode, the adjustment of movement speed, movement direction or posture can be achieved. After the specific movement mode is completed, rendering is realized through the AI chip and GPU.

[0089] The above A* algorithm, Dijkstra algorithm, or algorithms involving specific movement directions and speeds all belong to the prior art, and will not be elaborated here. It involves functions such as goodFeaturesToTrack().

[0090] In the above embodiments, the person can be correspondingly replaced with a commodity, an object, or a naked-eye 3D virtual object for scientific and technological display.

[0091] In this embodiment, by splitting the image and independently processing the sub-elements and upper-level elements, the requirements for computing power of the AI chip and GPU can be reduced, the hardware configuration can be reduced, and thus the cost can be reduced. Additionally, during the independent processing of the sub-elements and upper-level elements, an independent storage area is separated from the storage module to store the segmented sub-elements, upper-level elements, lower-level elements, and the relationship tree association dataset. The CPU distributes the data in each independent storage module to the AI chip and GPU for data processing. After separately processing the upper-level elements and lower-level elements of the running data and establishing a 3D model, the CPU synthesizes the upper-level element positions and lower-level elements according to the relationship tree association dataset corresponding to the set association tree, thereby generating a dynamically displayed effect diagram and storing the result output in the storage module.

[0092] In addition, it can also improve the image processing efficiency.

[0093] Embodiment 2:

[0094] Furthermore, the steps for the AI chip to segment the image include using a deep learning algorithm to identify and segment the objects in the image, and thus constructing upper-level elements and lower-level elements based on the depth information and visual hierarchy of the segmented objects in the image. The specific steps are as follows:

[0095] S1: Hierarchically identify the upper-level elements according to the visual object integrity, extract multiple sub-elements, store the sub-elements in independent sub-element file memories respectively, and attach coordinate information;

[0096] S2: According to the depth information, incorporate the information between the positions of each sub-element and the lower-level elements into the coordinate information to form a coordinate information set;

[0097] S3: Sequentially identify the lower-level elements from step S1 and step S2, establish a data coordinate axis with the lower-level elements, and modify the coordinate sets of each sub-element in the corresponding upper-level elements;

[0098] S4: Establish a relationship tree association dataset through the coordinate sets between the upper-level elements and the lower-level elements.

[0099] In the above steps, the objects in the segmented image are used to establish an image data association set in a tree structure by associating the set association tree with the relationship tree groups to the data set, and each sub-element is distributed to the corresponding image processor, and the AI chip is coordinated to process the image and the sub-elements. Therefore, in the process of the image processor coordinating with the AI chip to process the image information of a single element, the required computing power is small, and various mid- and low-end image processors and AI chips can be used to achieve the processing. Furthermore, independent data processing can be achieved. Through the establishment of the set relationship tree, after processing, the rapid combination of data and the establishment of the motion state can be realized. Through the data correlation of the computer, the establishment is more convenient and fast, and the risk of data disorder can be greatly reduced.

[0100] Furthermore, the method for the image processor to cooperate with the AI chip to process the image and the sub-elements includes the following steps:

[0101] Obtain the sub-elements and the lower-level elements;

[0102] Through the image processor cooperating with the AI chip, repair the sub-elements and the lower-level elements to improve the clarity and reduce the noise;

[0103] At the same time, perform feature recognition on the sub-elements and the lower-level elements;

[0104] Supplement and restore the missing parts of the sub-elements and the lower-level elements;

[0105] Restore the 2D sub-elements and the lower-level elements to a 3D model through 3D model synthesis technology;

[0106] Attach a coordinate set to the generated 3D model to establish a relationship tree associated data set.

[0107] Among them, for the generated 3D model, a three-dimensional motion coordinate system is established with the center, a three-dimensional motion coordinate system is established with the motion nodes, and a time-axis coordinate system starting from the above motion coordinate system. For the motion speed coordinate system corresponding to the time-axis coordinate system corresponding to the above motion coordinate system, through the time correlation formula of the motion speed in the motion coordinate system, it is used to control the change of the motion speed during the motion process. Set a motion vector for the above motion speed change to control the motion direction.

[0108] In this embodiment: By establishing a coordinate system for the 3D model and through the dynamic motion of the coordinate system, the dynamic control of the 3D model is realized. When it is necessary to change the dynamic display effect, by changing the dynamic display operation data, non-technical personnel can quickly learn and get started, and the operation method is simple.

[0109] The above ways of changing the dynamic display operation data include modifying the operation parameters or realizing it by setting the operation route.

[0110] Example 3:

[0111] Please refer to Figure 3 as shown, an AI-based intelligent image processing system includes:

[0112] Image acquisition module: responsible for directly acquiring, synthesizing or downloading images, as well as the data transmission process of images;

[0113] Image preprocessing module: corrects the noise in the image and performs color correction on the image;

[0114] AI chip and deep learning module: processes and preprocesses the image through the AI chip, and records and learns the recognition method and processing method through the deep learning module, and realizes deep learning and optimization through the dynamic motion display method;

[0115] Image processor: i.e., GPU, used to generate 3D images from 2D images, render 2D and 3D images, and process the motion data operation and real-time rendering of the motion state of sub-elements, upper-level elements and lower-level elements;

[0116] CPU, used to adjust computing power and allocate the work among the image processor, AI chip and deep learning module;

[0117] Motion control and dynamic display module: based on the computing power of the image processor, CPU, AI chip and deep learning module, responsible for processing the dynamic control of sub-elements in 2D or 3D space, and used to store the dynamic motion formulas and algorithms required for calling;

[0118] Storage module;

[0119] Interaction module: includes input and output modules.

[0120] It also includes a dynamic simulation display module, a 3D printing module and an output module. The dynamic simulation display module and the output module can be included in the interaction module.

[0121] An AI-based intelligent image processing terminal includes;

[0122] At least one terminal for running the intelligent image processing system;

[0123] One or more terminals running the intelligent image processing system cooperate to execute the intelligent image processing method.

[0124] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention. The structures, devices, and operation methods not specifically described and explained in the present invention are implemented according to the conventional means in the art without special instructions and limitations.

Claims

1. An AI-based intelligent image processing method, characterized in that, It includes the following steps: Obtain image data, segment the image through an AI chip, obtain the upper-layer elements and lower-layer elements in the image, and process the upper-layer elements and lower-layer elements into multiple image data sets and corresponding set association trees respectively. The image data sets of the upper-layer elements and the image data sets of the lower-layer elements form a set association tree, and the set association tree includes multiple groups; Based on the AI chip, complete the repair of the upper-layer elements and the lower-layer elements, and judge and classify the upper-layer elements and the lower-layer elements; Based on the AI algorithm, process the upper-layer elements and the lower-layer elements respectively, establish a 3D model, synthesize the upper-layer element positions and the lower-layer elements, and realize dynamic display; Establish a three-dimensional motion coordinate system centered on the generated 3D model, a three-dimensional motion coordinate system established with motion nodes, and a time-axis coordinate system starting from the above-mentioned motion coordinate system. For the motion speed coordinate system corresponding to the time-axis coordinate system corresponding to the above-mentioned motion coordinate system, use the time correlation formula of the motion speed in the motion coordinate system to control the change of the motion speed during the motion process; Set a motion vector for the above-mentioned change in motion speed to control the motion direction; The set association tree established between the upper-layer elements and the lower-layer elements is used to establish a motion base point with the sub-elements as the origin. Set a restricted area between each sub-element in the lower-layer elements and the upper-layer elements. Through the intelligent operation of the AI chip in the restricted area, an intelligent motion method is realized to achieve the motion mode of the sub-elements avoiding the restricted area.

2. The AI-based intelligent image processing method according to claim 1, wherein, The steps of the AI chip segmenting the image include using a deep learning algorithm to identify and segment the objects in the image, and thus constructing the upper-layer elements and the lower-layer elements based on the depth information and visual hierarchy of the objects in the segmented image. The specific steps are as follows: S1: Hierarchically identify the upper-layer elements according to the integrity of the visual objects, extract multiple sub-elements, store the sub-elements in independent sub-element file memories respectively, and attach coordinate information; S2: According to the depth information, incorporate the information between the positions of each sub-element and the lower-layer elements into the coordinate information to form a coordinate information set; S3: Sequentially identify the lower-layer elements from step S1 and step S2, establish a data coordinate axis with the lower-layer elements, and modify the coordinate sets of each sub-element in the corresponding upper-layer elements; S4: Establish a relationship tree associated data set through the coordinate sets between the upper-layer elements and the lower-layer elements.

3. An AI-based intelligent image processing method according to claim 2, characterized in that: The set association tree establishes an image data association set in a tree structure by associating each group of relationship tree associated data sets, distributes each sub-element to each corresponding image processor, and cooperates with the AI chip to realize the processing of the image and the sub-elements.

4. An AI-based intelligent image processing method according to claim 3, characterized in that, The method for the image processor to cooperate with the AI chip to realize the processing of the image and the sub-elements includes the following steps: Obtain the sub-elements and the lower-layer elements; Through the image processor, cooperate with the AI chip to repair the sub-elements and the lower-layer elements to improve the clarity and reduce the noise; At the same time, perform feature recognition on the sub-elements and the lower-layer elements; Supplement and restore the missing parts of the sub-elements and the lower-layer elements; Restore the 2D sub-elements and the lower-layer elements into a 3D model through 3D model synthesis technology; Attach a coordinate set to the generated 3D model to establish a relationship tree associated data set.

5. The method for AI-based intelligent image processing according to claim 1, wherein The method for implementing the movement of a sub-element to avoid a restricted area includes the following steps: Obtain the position coordinate information of the restricted area and the sub-element; Identify the sub-element, measure its volume and the range of motion states, and set the movement mode and activity mode of the sub-element; Judge or set whether the restricted area is dynamically changing; By setting the movement mode between the sub-element and the restricted area, the movement mode includes any one of: avoidance movement, circumferential movement, or acting movement; Decide the type of movement; Plan the decision movement path; Execute the decision movement instruction; Simulate the execution instruction.

6. An AI-based intelligent image processing system that uses the AI-based intelligent image processing method according to any one of claims 1-5, characterized in that, Including: Image acquisition module: responsible for directly acquiring, synthesizing, or downloading images, and the data transmission process of images; Image preprocessing module: correct the noise in the image and perform color correction on the image; AI chip and deep learning module: process and preprocess the image through the AI chip, and record and learn the recognition method and processing method through the deep learning module, and realize deep learning and optimization of the dynamic movement display method; Image processor: namely GPU, used to generate 3D images from 2D images, render 2D images and 3D images, and process the motion data operation and real-time rendering of the motion state of sub-elements, upper-level elements, and lower-level elements; CPU, used to adjust computing power and allocate the work among the image processor, AI chip, and deep learning module; Motion control and dynamic display module: based on the computing power of the image processor, CPU, AI chip, and deep learning module, responsible for processing the dynamic control of sub-elements in 2D or 3D space, and used to store the dynamic motion formulas and algorithms to be called; Storage module; Interaction module: including input and output modules.

7. An AI-based intelligent image processing system according to claim 6, characterized in that It also includes a dynamic simulation display module, a 3D printing module, and an output module.

8. An AI-based intelligent image processing terminal, characterized in that, Including; At least one terminal for running the intelligent image processing system according to claim 7; One or more terminals running the intelligent image processing system cooperate to execute the intelligent image processing method.

Images