Unmanned aerial vehicle regulation and control method and system based on artificial intelligence

Abstract

The invention relates to the technical field of computer control, and discloses an unmanned aerial vehicle regulation and control method and system based on artificial intelligence, and the method comprises the steps: sensing the image data and environment parameters of a flight region of an unmanned aerial vehicle through a sensor array installed on the unmanned aerial vehicle; selecting an identification model suitable for analyzing the image data of the flight area from a plurality of preset identification models according to the environment parameters of the flight area; identifying a target object in the flight area from the image data of the flight area by using an identification model; determining the position of the target object; calculating a flight path of the unmanned aerial vehicle according to the position of the target object; and controlling the unmanned aerial vehicle to fly according to the flight path of the unmanned aerial vehicle. According to the invention, not only is the autonomous navigation capability of the unmanned aerial vehicle in a complex and changeable environment enhanced, but also a higher safety standard is ensured while the working efficiency is improved, and a new direction is opened up for the development of the unmanned aerial vehicle technology.

Description

Technical Field

[0001] The present application relates to the field of computer control technology, and more specifically, to an artificial intelligence-based drone control method and system. Background Art

[0002] With the development of science and technology, UAVs (Unmanned Aerial Vehicles, UAVs) are increasingly used in agriculture, logistics, monitoring, search and rescue, and other fields. In order to improve the safety and efficiency of UAV operations, artificial intelligence-based technologies are increasingly being introduced into UAV flight control.

[0003] Traditionally, drones usually fly on pre-set routes or are manually controlled by ground operators using remote control devices. However, this flying method has many limitations and challenges. First, the preset routes lack flexibility. Once the environment changes or the target object moves, the drone may not be able to respond effectively, thus affecting the execution of the mission.

[0004] Therefore, a new technical solution for controlling the flight of drones is needed, so that drones can autonomously determine their flight paths and ensure that they will reach the target object even if the environment changes or the target object moves. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed to provide an artificial intelligence-based drone control method and system, so that the drone can autonomously formulate a flight path and ensure that the drone flies to the location of the target object even if the environment changes or the target object moves.

[0006] In a first aspect, the present application provides an artificial intelligence-based drone control method, comprising: sensing the image data and environmental parameters of the flight area of ​​the drone through a sensor array installed on the drone; selecting a recognition model suitable for analyzing the image data of the flight area from a plurality of preset recognition models according to the environmental parameters of the flight area; using the recognition model to identify a target object in the flight area from the image data of the flight area; determining the position of the target object; calculating the flight trajectory P(t) = {x(t), y(t), z(t)} of the drone according to the position of the target object, and making a preset optimization function

[0007] The value of is the smallest, where x(t), y(t), z(t) are the position components of the drone on the x, y and z axes at time t, x′(t), y′(t), z′(t) are the velocity components of P(t) on the x, y and z axes, d obs(P(t)) represents the distance between the position of the UAV at time t and the target object, k1, k2, k3, k4 are preset weighting coefficients, T is the preset end time, and the position of P(t) at time T is the position of the target object; according to the flight trajectory P(t) of the UAV, the UAV is controlled to fly.

[0008] Optionally, the aforementioned artificial intelligence-based drone control method selects a recognition model suitable for analyzing the image data of the flight area from a plurality of preset recognition models according to the environmental parameters of the flight area, including: calculating the illumination coefficient of the flight area according to the environmental parameters of the flight area Among them, α, β, γ, δ are preset weight values, vis, T, H are the visibility, temperature and humidity of the flight environment included in the environmental parameters of the flight environment; according to the illumination coefficient Lc of the flight area, a recognition model suitable for analyzing the image data of the flight area is selected from the multiple recognition models corresponding to different light intensities.

[0009] Optionally, the aforementioned artificial intelligence-based drone control method, before using the recognition model to identify the target object in the flight area from the image data of the flight area, further includes: adjusting the pixel value of the image data of the flight area according to the illumination coefficient Lc of the flight area; inputting the image data of the flight area into a preset neural network-based image enhancement model to perform image enhancement processing on the image data of the flight area.

[0010] Optionally, the aforementioned artificial intelligence-based drone control method adjusts the pixel value of the image data of the flight area according to the illumination coefficient Lc of the flight area, including: for each pixel in the image data of the flight area, calculating the pixel mean of a preset size area centered on the pixel, wherein the pixel value of the i-th pixel is I i , the pixel mean of the preset size area centered on the i-th pixel is I i,clahe ; Calculate the target pixel value of each pixel in the image data of the flight area, wherein the target pixel value of the i-th pixel

[0011] Optionally, the aforementioned artificial intelligence-based drone control method, before inputting the image data of the flight area into a preset neural network-based image enhancement model and performing image enhancement processing on the image data of the flight area, further includes: when training the image enhancement model, inputting a preset original image into the image enhancement model to obtain an enhanced image output by the image enhancement model; calculating the degree of structural similarity between the enhanced image and the expected image corresponding to the original image; calculating the value of the loss function of the image enhancement model based on the degree of structural similarity between the enhanced image and the expected image; and adjusting the image enhancement model based on the value of the loss function of the image enhancement model.

[0012] Optionally, the aforementioned artificial intelligence-based drone control method calculates the degree of structural similarity between the enhanced image and the expected image, including: calculating the mean and standard deviation of the pixels of the enhanced image, calculating the mean and standard deviation of the pixels of the expected image; calculating the covariance of the pixels of the enhanced image and the expected image; calculating the degree of structural similarity between the enhanced image and the expected image. Wherein, A and B represent the enhanced image and the expected image, μ(A) and μ(S) are the means of the pixels of the enhanced image and the expected image, σ(A) and σ(S) are the standard deviations of the pixels of the expected image, σ(A, B) is the covariance of the pixels of the enhanced image and the expected image, and C1 and C2 are preset equilibrium constants.

[0013] Optionally, the aforementioned artificial intelligence-based drone control method calculates the value of the loss function of the image enhancement model according to the degree of structural similarity between the enhanced image and the expected image, including: calculating the mean square error of pixels between the enhanced image and the expected image; calculating the value of the loss function of the image enhancement model Loss=w1M(A, B)+w2(1-S(a, B)), wherein A and B represent the enhanced image and the expected image, M(A, B) represents the mean square error of pixels between the enhanced image and the expected image, S(A, B) represents the degree of structural similarity between the enhanced image and the expected image, and w1 and w2 represent preset weight values.

[0014] Optionally, the aforementioned artificial intelligence-based drone control method determines the position of the target object, including: calculating the size of the target object in the image data of the flight area; calculating the ratio of the size of the target object in the image data of the flight area to the known actual size of the target object; calculating the distance between the target object and the drone based on the ratio of the size of the target object in the image data of the flight area to the actual size of the target object; calculating the position of the target object based on the distance between the target object and the drone and the position of the drone.

[0015] In the second aspect, the present application provides an artificial intelligence-based drone control system, including: a perception module, which perceives the image data and environmental parameters of the flight area of ​​the drone through a sensor array installed on the drone; a model selection module, which selects a recognition model suitable for analyzing the image data of the flight area from a plurality of preset recognition models according to the environmental parameters of the flight area; an object recognition module, which uses the recognition model to identify the target object in the flight area from the image data of the flight area; a position determination module, which determines the position of the target object; a trajectory calculation module, which calculates the flight trajectory P(t) = {x(t), y(t), z(t)} of the drone according to the position of the target object, and uses a preset optimization function

[0016] The value of is the smallest, where x(t), y(t), z(t) are the position components of the drone on the x, y and z axes at time t, x′(t), y′(t), z′(t) are the velocity components of P(t) on the x, y and z axes, d obs (P(t)) represents the distance between the position of the UAV at time t and the target object, k1, k2, k3, k4 are preset weighting coefficients, T is the preset end time, and the position of P(t) at time T is the position of the target object; the control module controls the UAV to fly according to the flight trajectory P(t) of the UAV.

[0017] The above one or more technical solutions of the present application have at least one or more of the following beneficial effects:

[0018] According to the technical solution of the present invention, not only the autonomous navigation capability of UAVs in complex and changing environments is enhanced, but also higher safety standards are guaranteed while improving work efficiency, thus opening up a new direction for the development of UAV technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0020] Figure 1 is a flow chart of a drone control method based on artificial intelligence according to an embodiment of the present application;

[0021] Figure 2 is another flow chart of a drone control method based on artificial intelligence according to an embodiment of the present application;

[0022] Figure 3 is another flow chart of a drone control method based on artificial intelligence according to an embodiment of the present application;

[0023] Figure 4 is a partial flow chart of an artificial intelligence-based drone control method according to an embodiment of the present application;

[0024] Figure 5 is another partial flow chart of a drone control method based on artificial intelligence according to an embodiment of the present application;

[0025] Figure 6 The present invention is a block diagram of an artificial intelligence-based drone control system according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] Some embodiments of the present application are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0027] like Figure 1 As shown, an embodiment of the present application provides a drone control method based on artificial intelligence, including:

[0028] Step S110, sensing image data and environmental parameters of the flight area of ​​the drone through a sensor array installed on the drone.

[0029] Step S120 , selecting a recognition model suitable for analyzing the image data of the flight area from a plurality of preset recognition models according to the environmental parameters of the flight area.

[0030] In this embodiment, the most appropriate recognition model can be automatically selected according to the environmental parameters of different flight areas, which enables the drone to accurately perceive and understand the surrounding conditions in a variety of environments, thereby improving the flexibility and applicability of the system.

[0031] Step S130: using the recognition model to identify the target object in the flight area from the image data of the flight area.

[0032] In this embodiment, the target object may be a person or animal that moves at any time. In this embodiment, a specially trained recognition model is used to process image data, so that the moving target object (such as a person or an animal) can be located more accurately, thereby ensuring that the drone can efficiently track and respond to dynamically changing targets.

[0033] Step S140, determining the position of the target object.

[0034] Step S150, according to the position of the target object, calculate the flight trajectory P(t) = {x(t)y(t), z(t)} of the drone, and use the preset optimization function

[0035] The value of is the smallest, where x(t), y(t), z(t) are the position components of the drone on the x, y and z axes at time t, x′(t), y′(t), z′(t) are the velocity components of P(t) on the x, y and z axes, respectively. obs (P(t)) represents the distance between the position of the drone at time t and the target object, k1, k2, k3, k4 are preset weighting coefficients, T is the preset end time, and the position of P(t) at time T is the position of the target object.

[0036] In this embodiment, a fast and stable flight trajectory is calculated by minimizing a preset optimization function. This path planning not only takes into account the influence of the speed component, but also pays special attention to the safe distance from obstacles, thus ensuring the safety and stability of the UAV flight.

[0037] Step S160, controlling the UAV to fly according to the flight trajectory P(t) of the UAV.

[0038] According to the technical solution of this embodiment, not only the autonomous navigation capability of the UAV in complex and changing environments is enhanced, but also higher safety standards are guaranteed while improving work efficiency, opening up a new direction for the development of UAV technology.

[0039] like Figure 2 As shown, one embodiment of the present application provides another method for controlling a drone based on artificial intelligence. Compared with the above-mentioned embodiment, the method for controlling a drone based on artificial intelligence in this embodiment, step S120 includes:

[0040] Step S210, calculating the illumination coefficient of the flight area according to the environmental parameters of the flight area Wherein α, β, γ, δ are preset weight values, and vis, T, H are visibility, temperature and humidity of the flight environment included in the environmental parameters of the flight environment.

[0041] Step S220 , selecting a recognition model suitable for analyzing the image data of the flight area from a plurality of recognition models corresponding to different light intensities according to the light coefficient Lc of the flight area.

[0042] According to the technical solution of this embodiment, by calculating the illumination coefficient of the flight area and selecting the most suitable model for the current environment from multiple recognition models optimized for different illumination intensities based on this coefficient, the problem of inaccurate target recognition caused by light changes can be effectively solved. This allows the drone to maintain a high recognition rate under various lighting conditions (such as early morning, noon, evening or cloudy days).

[0043] like Figure 3 As shown, one embodiment of the present application provides another method for controlling a drone based on artificial intelligence. Compared with the above-mentioned embodiment, the method for controlling a drone based on artificial intelligence in this embodiment, before step S130, further includes:

[0044] Step S310: adjusting the pixel value of the image data of the flight area according to the illumination coefficient Lc of the flight area.

[0045] This step specifically includes:

[0046] (1) For each pixel in the image data of the flight area, the pixel mean of the area of ​​preset size centered on it is calculated, where the pixel value of the i-th pixel is I i , the pixel mean of the preset size area centered on the i-th pixel is I i，clahe .

[0047] (2) Calculate the target pixel value of each pixel in the image data of the flight area, where the target pixel value of the i-th pixel is

[0048] In this embodiment, by adjusting the target pixel value of each pixel according to the illumination coefficient and combining the pixel mean value of the local area, the image quality problem caused by uneven illumination can be effectively improved. This adaptive pixel value adjustment helps to balance the brightness distribution in the image and reduce the influence of shadow and highlight areas, thereby obtaining a clearer and more consistent image.

[0049] Step S320: input the image data of the flight area into a preset neural network-based image enhancement model to perform image enhancement processing on the image data of the flight area.

[0050] According to the technical solution of this embodiment, the image is further processed using a neural network-based image enhancement model to extract more detailed information, which makes the subsequent target recognition process more accurate and reliable. Especially in the case of poor lighting conditions, image enhancement processing can greatly improve the effect of the recognition model.

[0051] like Figure 4 As shown, one embodiment of the present application provides another method for controlling a drone based on artificial intelligence. Compared with the aforementioned embodiment, the method for controlling a drone based on artificial intelligence in this embodiment, before step S320, further includes:

[0052] Step S410, when training the image enhancement model, a preset original image is input into the image enhancement model to obtain an enhanced image output by the image enhancement model.

[0053] Step S420, calculating the structural similarity between the enhanced image and the desired image corresponding to the original image.

[0054] Specifically include:

[0055] (1) Calculate the mean and standard deviation of the pixels of the enhanced image, and calculate the mean and standard deviation of the pixels of the expected image.

[0056] (2) Calculate the pixel covariance between the enhanced image and the desired image.

[0057] (3) Calculate the structural similarity between the enhanced image and the expected image Wherein, A and B represent the enhanced image and the expected image, μ(A) and μ(B) are the means of the pixels of the enhanced image and the expected image, σ(A) and σ(B) are the standard deviations of the pixels of the expected image, σ(A, B) is the covariance of the pixels of the enhanced image and the expected image, and C1 and C2 are preset equilibrium constants.

[0058] In this embodiment, the similarity between the enhanced image and the expected image is measured by the structural similarity degree, so that the structural information in the image can be better preserved, so that the processed image is not only improved in brightness and contrast, but also maintains high clarity and authenticity in subtle features such as texture and edges.

[0059] Step S430, calculating the value of the loss function of the image enhancement model according to the degree of structural similarity between the enhanced image and the expected image.

[0060] Specifically include:

[0061] (1) Calculate the mean square error between the pixels of the enhanced image and the expected image.

[0062] In this embodiment, by considering both pixel-level mean square error and structural similarity in the loss function, the model can be guided to pay more attention to global and local consistency during the training process. Such a multi-objective optimization strategy helps to accelerate the convergence process, reduce the risk of overfitting, and thus improve the overall efficiency of model training.

[0063] (2) Calculate the value of the loss function of the image enhancement model: Loss = w1M(A, B) + w2(1-S(A, B)), where A and B represent the enhanced image and the expected image, M(A, B) represents the mean square error of the pixels between the enhanced image and the expected image, S(A, B) represents the degree of structural similarity between the enhanced image and the expected image, and w1 and w2 represent preset weight values.

[0064] In this embodiment, by calculating the degree of structural similarity between the enhanced image and the desired image and combining the mean square error to define the loss function, the output quality of the image enhancement model can be more comprehensively evaluated. This comprehensive evaluation method helps to ensure that the enhanced image retains the details of the original image while also being as close to the ideal visual effect as possible.

[0065] Step S440: adjusting the image enhancement model based on the value of the loss function of the image enhancement model.

[0066] According to the technical solution of this embodiment, by introducing the degree of structural similarity and the optimized loss function, the quality and robustness of the image enhancement model are significantly improved, providing strong support for achieving a higher level of autonomous navigation of drones. This method not only improves the effect of image enhancement, but also enhances the stability and reliability of the entire system.

[0067] like Figure 5 As shown, one embodiment of the present application provides another method for controlling a drone based on artificial intelligence. Compared with the above-mentioned embodiment, the method for controlling a drone based on artificial intelligence in this embodiment, step S140 includes:

[0068] Step S510, calculating the size of the target object in the image data of the flight area.

[0069] Step S520, calculating the ratio between the size of the target object in the image data of the flight area and the known real size of the target object.

[0070] Step S530 , calculating the distance between the target object and the drone according to the ratio of the size of the target object in the image data of the flight area to the actual size of the target object.

[0071] Step S540, calculating the position of the target object according to the distance between the target object and the drone and the position of the drone.

[0072] According to the technical solution of this embodiment, the actual distance between the target object and the drone can be accurately estimated by using the proportional relationship between the size of the target object in the image data and its known true size. This distance measurement method based on visual information is more intuitive and less susceptible to environmental factors than traditional distance measurement methods. By accurately calculating the position of the target object, the drone can better plan its flight path and avoid collisions with obstacles. This is particularly important for operations in complex or changing environments, helping to reduce the risk of accidents and protect equipment safety.

[0073] like Figure 6 As shown, an embodiment of the present application provides a drone control system based on artificial intelligence, including:

[0074] The perception module 610 perceives the image data and environmental parameters of the flight area of ​​the drone through the sensor array installed on the drone.

[0075] The model selection module 620 selects a recognition model suitable for analyzing the image data of the flight area from a plurality of preset recognition models according to the environmental parameters of the flight area.

[0076] In this embodiment, the most appropriate recognition model can be automatically selected according to the environmental parameters of different flight areas, which enables the drone to accurately perceive and understand the surrounding conditions in a variety of environments, thereby improving the flexibility and applicability of the system.

[0077] The object recognition module 630 uses a recognition model to recognize a target object in the flight area from the image data of the flight area.

[0078] In this embodiment, the target object may be a person or animal that moves at any time. In this embodiment, a specially trained recognition model is used to process image data, so that the moving target object (such as a person or an animal) can be located more accurately, thereby ensuring that the drone can efficiently track and respond to dynamically changing targets.

[0079] The position determination module 640 determines the position of the target object.

[0080] The trajectory calculation module 650 calculates the flight trajectory P(t) = {x(t), y(t), z(t)} of the drone according to the position of the target object, and uses the preset optimization function

[0081] The value of is the smallest, where x(t), y(t), z(t) are the position components of the drone on the x, y and z axes at time t, x′(t), y′(t), z′(t) are the velocity components of P(t) on the x, y and z axes, respectively. obs (P(t)) represents the distance between the position of the drone at time t and the target object, k1, k2, k3, k4 are preset weighting coefficients, T is the preset end time, and the position of P(t) at time T is the position of the target object.

[0082] In this embodiment, a fast and stable flight trajectory is calculated by minimizing a preset optimization function. This path planning not only takes into account the influence of the speed component, but also pays special attention to the safe distance from obstacles, thus ensuring the safety and stability of the UAV flight.

[0083] The control module 660 controls the UAV to fly according to the flight trajectory P(t) of the UAV.

[0084] According to the technical solution of this embodiment, not only the autonomous navigation capability of the UAV in complex and changing environments is enhanced, but also higher safety standards are guaranteed while improving work efficiency, opening up a new direction for the development of UAV technology.

[0085] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0086] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0087] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0088] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0089] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. The method for controlling unmanned aerial vehicles based on artificial intelligence is characterized in that: include: The sensor array installed on the drone senses the image data and environmental parameters of the flight area of ​​the drone; Selecting a recognition model suitable for analyzing the image data of the flight area from a plurality of preset recognition models according to the environmental parameters of the flight area; using the recognition model, identifying a target object in the flight area from image data of the flight area; determining a location of the target object; According to the position of the target object, the flight trajectory of the drone P(t) = {x(t), y(t), z(t)} is calculated, and the preset optimization function The value of is the smallest, where x(t), y(t), z(t) are the position components of the drone on the x, y and z axes at time t, x′(t), y′(t), z′(t) are the velocity components of P(t) on the x, y and z axes, d obs (P(t)) represents the distance between the position of the drone at time t and the target object, k1, k2, k3, k4 are preset weighting coefficients, T is the preset end time, and the position of P(t) at time T is the position of the target object; According to the flight trajectory P(t) of the UAV, the UAV is controlled to fly.

2. The method for controlling a drone based on artificial intelligence according to claim 1, characterized in that: According to the environmental parameters of the flight area, a recognition model suitable for analyzing the image data of the flight area is selected from a plurality of preset recognition models, including: Calculate the illumination coefficient of the flight area according to the environmental parameters of the flight area Wherein, α, β, γ, δ are preset weight values, and vis, T, H are visibility, temperature and humidity of the flight environment included in the environmental parameters of the flight environment; According to the illumination coefficient Lc of the flight area, a recognition model suitable for analyzing the image data of the flight area is selected from the multiple recognition models corresponding to different illumination intensities.

3. The method for controlling a drone based on artificial intelligence according to claim 2, characterized in that: Before using the recognition model to recognize the target object in the flight area from the image data of the flight area, the method further includes: adjusting the pixel value of the image data of the flight area according to the illumination coefficient Lc of the flight area; The image data of the flight area is input into a preset neural network-based image enhancement model to perform image enhancement processing on the image data of the flight area.

4. The method for controlling a drone based on artificial intelligence according to claim 3, characterized in that: According to the illumination coefficient Lc of the flight area, adjusting the pixel value of the image data of the flight area comprises: For each pixel in the image data of the flight area, the pixel mean of the area of ​​preset size centered on the pixel is calculated, where the pixel value of the i-th pixel is I i , the pixel mean of the preset size area centered on the i-th pixel is I i，clahe ; Calculate the target pixel value of each pixel in the image data of the flight area, wherein the target pixel value of the i-th pixel is 5. The method for controlling a drone based on artificial intelligence according to claim 3, characterized in that: Before inputting the image data of the flight area into a preset neural network-based image enhancement model to perform image enhancement processing on the image data of the flight area, the method further includes: When training the image enhancement model, a preset original image is input into the image enhancement model to obtain an enhanced image output by the image enhancement model; Calculating the degree of structural similarity between the enhanced image and the expected image corresponding to the original image; Calculating the value of the loss function of the image enhancement model according to the degree of structural similarity between the enhanced image and the expected image; The image enhancement model is adjusted based on the value of the loss function of the image enhancement model.

6. The method for controlling a drone based on artificial intelligence according to claim 5, characterized in that: Calculating the structural similarity between the enhanced image and the expected image includes: Calculating the mean and standard deviation of pixels of the enhanced image, and calculating the mean and standard deviation of pixels of the expected image; Calculating the covariance of pixels of the enhanced image and the desired image; Calculate the structural similarity between the enhanced image and the expected image Wherein, A, B represent the enhanced image and the expected image, μ(A), μ(B) are the means of the pixels of the enhanced image and the expected image, σ(A), σ(B) are the standard deviations of the pixels of the expected image, σ(A, B) is the covariance of the pixels of the enhanced image and the expected image, and C1, C2 are preset equilibrium constants.

7. The method for controlling a drone based on artificial intelligence according to claim 5, characterized in that: Calculating the value of the loss function of the image enhancement model according to the structural similarity between the enhanced image and the expected image includes: Calculating the mean square error between pixels of the enhanced image and the desired image; Calculate the value of the loss function of the image enhancement model: Loss = w1M(A, B) + w2(1-S(A, B)), where A and B represent the enhanced image and the expected image, M(A, B) represents the mean square error of pixels between the enhanced image and the expected image, S(A, B) represents the degree of structural similarity between the enhanced image and the expected image, and w1 and w2 represent preset weight values.

8. The method for controlling a drone based on artificial intelligence according to claim 1, characterized in that: Determining the location of the target object includes: Calculating the size of the target object in the image data of the flight area; Calculating a ratio between a size of the target object in the image data of the flight area and a known real size of the target object; Calculating the distance between the target object and the drone according to the ratio of the size of the target object in the image data of the flight area to the actual size of the target object; The position of the target object is calculated according to the distance between the target object and the drone and the position of the drone.

9. The drone control system based on artificial intelligence is characterized by: include: A perception module, which perceives image data and environmental parameters of the flight area of ​​the drone through a sensor array installed on the drone; A model selection module, which selects a recognition model suitable for analyzing the image data of the flight area from a plurality of preset recognition models according to the environmental parameters of the flight area; an object recognition module, using the recognition model to recognize a target object in the flight area from the image data of the flight area; A position determination module, for determining the position of the target object; The trajectory calculation module calculates the flight trajectory P(t)={x(t), y(t), z(t)} of the drone according to the position of the target object, and uses a preset optimization function The value of is the smallest, where x(t), y(t), z(t) are the position components of the drone on the x, y and z axes at time t, x′(t), y′(t), z′(t) are the velocity components of P(t) on the x, y and z axes, d obs (P(t)) represents the distance between the position of the drone at time t and the target object, k1, k2, k3, k4 are preset weighting coefficients, T is the preset end time, and the position of P(t) at time T is the position of the target object; The control module controls the UAV to fly according to the flight trajectory P(t) of the UAV.

Images