Intelligent camera shooting platform and method thereof for realizing automatic tracking and position angle adjustment

Abstract

The invention discloses an intelligent camera shooting platform capable of automatically tracking and adjusting a position angle. The image quality calculation module is used for calculating the image quality index of the current shot image in real time. The parameter estimation module is used for adopting a trained parameter estimation network to obtain coarse tuning parameters capable of enabling the central target to move to a position near an optimal shooting position in a shooting picture; and fine adjustment parameters capable of improving the quality index of the shot picture are also obtained. And the control module is used for judging whether the picture quality index after fine adjustment by the adjustment module is superior to the picture quality index before fine adjustment. If yes, the parameter estimation module recalculates the fine adjustment parameter, and the optimization step is repeated. If not, the adjustment is completed. According to the method, the shooting picture is evaluated, the central target is kept to be placed at the optimal shooting position, and the position and / or angle of the shooting equipment are / is automatically adjusted to improve the shooting picture quality.

Description

Technical Field

[0001] The present application relates to a camera platform, which is used to carry a camera device (smartphone, digital camera, tablet computer, etc.) and can adjust the position and angle of the camera device. Background Art

[0002] When taking photos and videos everyday, the photographer and the subject usually need to adjust the shooting position and angle several times to take satisfactory photos and videos. Especially when taking selfies, it is more difficult to adjust the camera position and angle to take satisfactory photos. Summary of the invention

[0003] The technical problem to be solved by the present application is: how to drive the camera device to automatically track the photographed object and automatically adjust the relative position and angle with the photographed object, so as to improve the shooting quality.

[0004] In order to solve the above technical problems, the present application proposes an intelligent camera platform for automatically tracking and adjusting the position angle, including: an identification and tracking module, an image quality calculation module, a parameter estimation module, an adjustment module, and a control module. The identification and tracking module is used to automatically identify the central target in the shooting picture of the camera device, and is also used to automatically track the central target in the shooting picture of the camera device. The image quality calculation module is used to calculate the quality index of the current shooting picture in real time, referred to as the image quality index. The parameter estimation module is used to use a trained neural network, called a parameter estimation network, according to the position of the central target in the current shooting picture, to obtain a coarse adjustment parameter that can move the central target to the vicinity of the optimal shooting position in the shooting picture; it is also used to use the trained parameter estimation network, according to the current shooting picture, to obtain a fine adjustment parameter that can improve the quality index of the shooting picture; the coarse adjustment parameter and the fine adjustment parameter both include any one or more of the following: a horizontal displacement parameter of the camera platform, an azimuth rotation parameter of the camera platform, an up and down displacement parameter of the camera device, and a pitch angle rotation parameter of the camera device. The adjustment module is used to drive the camera platform and / or camera device to perform corresponding coarse adjustment or fine adjustment based on the coarse adjustment parameter or fine adjustment parameter. The control module is used to determine whether the image quality index after the fine adjustment by the adjustment module is better than the image quality index before the fine adjustment; if so, the parameter estimation module calculates the fine adjustment parameter again, and the adjustment module performs corresponding fine adjustment again according to the new fine adjustment parameter, and the control module determines again whether the image quality index before and after the new fine adjustment is optimized; if not, it means that the adjustment is completed.

[0005] Optionally, the identification and tracking module is changed to allow the user to select a central target in the shooting picture of the camera device.

[0006] Furthermore, the image quality calculation module uses a neural network technology, called an image quality calculation network, to combine objective indicators of image quality and subjective preferences of users to calculate the quality indicators of the current captured image. The objective indicators of image quality include: any one or more of image brightness, contrast, and clarity. The subjective preferences of users include: any one or more of the photographer's preference for color saturation, the size ratio of people in the landscape, and the aesthetics of composition. The image quality calculation network first obtains a basic image quality calculation network that is not affected by the subjective preferences of users by learning samples, and then further obtains a personalized image quality calculation network that meets the subjective preferences of different users by learning user feedback during the use of the basic image quality calculation network.

[0007] Furthermore, the input of the parameter estimation network includes: the current shooting picture, the position of the central target in the current shooting picture; the output of the parameter estimation network includes: the displacement direction of the next step of the camera platform, the displacement step of the camera platform, the azimuth angle of the camera platform, the lifting height of the robotic arm, and the pitch angle of the camera lens.

[0008] Furthermore, the training data set used by the parameter estimation network includes photo samples of different scenes, lighting conditions and shooting distances; it also includes adjustment parameters corresponding to each photo sample - any one or more of the optimal displacement direction, step size, azimuth angle, lifting height and pitch angle; the adjustment parameters are used to improve the position of the central target in the photo sample and improve the image quality index of the photo sample; each photo sample and the corresponding adjustment parameter constitute a data pair, and all data pairs constitute the training data set.

[0009] Furthermore, the parameter estimation network is implemented using a convolutional neural network for extracting image features; during the training process, the parameter estimation network uses a back propagation algorithm to optimize the weights of the neural network and uses batch gradient descent or Adam optimizer for training.

[0010] Furthermore, the adjustment module drives the camera platform to move in the horizontal plane according to the displacement direction and displacement step of the camera platform, drives the camera platform to rotate left and right in the horizontal plane according to the azimuth rotation parameter of the camera platform, drives the camera device to rise and fall according to the lifting height of the camera device, and drives the camera device to rotate up and down according to the pitch rotation parameter of the camera device. The adjustment module includes a translation unit and a mechanical arm; the translation unit is used to drive the entire camera platform and the camera device to move in the horizontal plane and rotate left and right; the mechanical arm is used to mount the camera device and drive the camera device to move up and down, and rotate up and down.

[0011] Furthermore, the smart camera platform also has a matching application App, and the two communicate with each other and transmit data; the matching App is installed on the user's smart electronic device, and the matching App displays the current shooting picture in real time to accept the user's confirmation or adjustment.

[0012] The present application also proposes a method for an intelligent camera platform to automatically track and adjust the position and angle of a camera device, comprising the following steps. Step S1: The camera device is fixed on the camera platform, and the camera device collects the shooting picture in real time. The recognition and tracking module automatically recognizes the central target in the shooting picture, and also automatically tracks the central target in the shooting picture. Step S2: If the recognition and tracking module finds that the position of the central target in the camera picture moves, the parameter estimation module gives a coarse adjustment parameter that can make the central target be located near the best shooting position of the shooting picture according to the position of the central target in the current shooting picture. Step S3: The adjustment module makes a corresponding coarse adjustment to the position and / or angle of the camera platform and / or the camera device according to the coarse adjustment parameter, and after the coarse adjustment, the central target is located near the best shooting position of the shooting picture. If the recognition and tracking module does not find that the position of the central target in the camera picture moves, steps S2 and S3 are omitted. Step S4: The image quality calculation module calculates the image quality index of the current shooting picture. Step S5: The parameter estimation module gives a fine adjustment parameter that can improve the quality index of the shooting picture according to the current shooting picture. Step S6: The position adjustment module performs corresponding fine-tuning on the position and / or angle of the camera platform and / or camera device according to the fine-tuning parameters. Step S7: The image quality calculation module calculates the image quality index of the captured image after fine-tuning. Step S8: The control module determines the image quality index before and after fine-tuning; if the image quality index after fine-tuning is better than the image quality index before fine-tuning, return to step S5 to calculate the fine-tuning parameters again, and repeat steps S6 to S8; if the image quality index after fine-tuning is not better than the image quality index before fine-tuning, the adjustment of the camera platform and / or camera device is completed.

[0013] Furthermore, a second threshold is set in the step S8; if the image quality index after fine-tuning is not better than the image quality index before fine-tuning, the value of the failure adjustment calculator is increased by one; if the value of the failure adjustment calculator is less than the second threshold, return to step S5 to calculate the adjustment parameters again, and repeat steps S6 to S8; if the value of the failure adjustment calculator is ≥ the second threshold, it indicates that the adjustment of the camera platform and / or the camera device is completed.

[0014] The technical effect achieved by the present application is: by evaluating the captured image, the central target is kept in the optimal shooting position, and the position and / or angle of the shooting device is automatically adjusted to improve the captured image quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1It is a structural schematic diagram of an intelligent camera platform proposed in this application for realizing automatic tracking and adjusting position angle.

[0016] Figure 2 It is a flow chart of the method for realizing automatic tracking and adjusting position angle by the intelligent camera platform proposed in this application.

[0017] Explanation of the reference numerals in the figures: identification and tracking module 1, image quality calculation module 2, parameter estimation module 3, adjustment module 4, control module 5. DETAILED DESCRIPTION

[0018] See also Figure 1 The intelligent camera platform (hereinafter referred to as the camera platform or intelligent camera platform) proposed in this application for realizing automatic tracking and adjusting position angle includes: an identification and tracking module 1, an image quality calculation module 2, a parameter estimation module 3, an adjustment module 4, and a control module 5.

[0019] The identification and tracking module 1 is used to automatically identify the central target in the shooting picture of the camera device (i.e., the picture in the viewfinder of the camera device), or the user can select the central target in the shooting picture of the camera device instead. The central target identification and tracking module 1 is also used to automatically track the central target in the shooting picture of the camera device.

[0020] The image quality calculation module 2 is used to identify the elements (people, objects and landscapes) in the current shooting picture, and at the same time calculate the quality index of the current shooting picture (referred to as image quality index) in real time according to factors such as the integrity, proportion, aesthetics, lighting and clarity of each element.

[0021] As a preferred example, the image quality calculation module 2 uses neural network technology when calculating the image quality index, for example, a multi-layer convolutional neural network (CNN) is used as the image quality calculation network, combining the objective indicators of image quality and the user's subjective preferences to comprehensively evaluate the quality of the current captured picture.

[0022] Objective indicators of image quality include: image brightness, contrast, and clarity.

[0023] The subjective preferences of users (e.g., photographers) include: (1) the photographer's preference for color saturation. Some users prefer bright colors, while others prefer soft tones. (2) The size ratio of people to landscape. Some users like people to occupy a large proportion of the photo, while others want people to occupy a smaller proportion of the photo and focus on the landscape. (3) Composition aesthetics. Some users prefer symmetrical compositions, while others prefer asymmetrical artistic styles.

[0024] The image quality calculation network first obtains a basic image quality calculation network that is not affected by the user's subjective preferences by learning from a large number of samples, and then further obtains a personalized image quality calculation network that meets the subjective preferences of different users by learning from user feedback during the use of the basic image quality calculation network.

[0025] The parameter estimation module 3 is used to use a trained neural network, such as a parameter estimation network, to obtain a coarse adjustment parameter that can move the central target to the vicinity of the optimal shooting position in the shooting picture according to the position of the central target in the current shooting picture. The parameter estimation module 3 is also used to use the trained parameter estimation network to obtain a fine adjustment parameter that can improve the quality index of the shooting picture according to the current shooting picture. The coarse adjustment parameter and the fine adjustment parameter both include any one or more of the following: a horizontal displacement parameter of the camera platform, an azimuth rotation parameter of the camera platform, an up-and-down displacement parameter of the camera device, and a pitch angle rotation parameter of the camera device.

[0026] As a preferred example, the input of the parameter estimation network includes: the current shooting picture (image data), the position of the central target in the current shooting picture (coordinate information, provided by the central target recognition and tracking module 1).

[0027] The output of the parameter estimation network includes: the displacement direction of the camera platform for the next step, the displacement step length L of the camera platform step , the azimuth angle α of the camera platform, the lifting height H of the robot arm, and the pitch angle β of the camera lens. The camera platform has 8 possible displacement directions for the next step, namely front, back, left, right, left front 45 degrees, right front 45 degrees, left back 45 degrees, and right back 45 degrees; it can also be changed to front, back, left, and right, a total of 4 possible directions. The next displacement direction of the camera platform and the displacement step length L of the camera platform step (indicates the distance of a single movement) These two constitute the horizontal plane movement parameters. The azimuth angle α of the camera platform indicates how many degrees to rotate the camera platform and the camera equipment fixed on it to the left or right in the horizontal plane. The lifting height H of the robot arm is the up and down movement parameter. For example, a positive value indicates that it is to be raised, and a negative value indicates that it is to be lowered. The pitch angle α of the camera lens indicates how many degrees to rotate the camera lens upward or downward in the plane (vertical plane) perpendicular to the horizontal plane where the camera lens is located.

[0028] The training data set used by the parameter estimation network includes a large number of photo samples, including photo samples of different scenes, lighting conditions and shooting distances; and also includes adjustment parameters corresponding to each photo sample - such as the optimal displacement direction, step size, azimuth angle, lifting height and pitch angle. The adjustment parameters (including coarse adjustment parameters and fine adjustment parameters) are used to optimize the shooting effect - improve the position of the central target in the photo sample and improve the image quality index of the photo sample. The adjustment parameters are generated by annotations of professional photographers or automated algorithms. Each photo sample and the corresponding adjustment parameter constitute a data pair, and all data pairs constitute a training data set.

[0029] As a preferred example, the parameter estimation network uses a convolutional neural network suitable for image input to extract image features. For example, ResNet (Residual Neural Network) is suitable for dealing with the gradient vanishing problem of deep networks and can extract more complex image features. As another example, MobileNet has a lightweight network structure, which is suitable for real-time processing on mobile devices and meets the needs of smart camera platforms. Preferably, during the training process, the parameter estimation network uses a back-propagation algorithm to optimize the weights of the neural network, uses batch gradient descent (Batch Gradient Descent, BGD) or Adam optimizer for training to increase the convergence speed, and sets an early stopping mechanism to avoid overfitting problems.

[0030] The adjustment module 4 is used to drive the camera platform and / or the camera device to perform corresponding coarse adjustment or fine adjustment based on the coarse adjustment parameter or fine adjustment parameter. Specifically, the adjustment module 4 drives the camera platform to move in the horizontal plane according to the horizontal plane displacement parameter of the camera platform, drives the camera platform to rotate left and right in the horizontal plane according to the azimuth rotation parameter of the camera platform, drives the camera device to rise and fall according to the up and down displacement parameter of the camera device, and drives the camera device to rotate up and down according to the pitch angle rotation parameter of the camera device.

[0031] As an example, the adjustment module 4 includes a translation unit and a mechanical arm. The translation unit is used to drive the entire camera platform and the camera device to move in the horizontal plane and rotate left and right. The mechanical arm is used to mount the camera device and drive the camera device to move up and down and rotate up and down.

[0032] The control module 5 is used to determine whether the image quality index after fine-tuning by the adjustment module 4 is better than the image quality index before fine-tuning. If yes, the parameter estimation module 3 calculates the fine-tuning parameters again, and the adjustment module 4 performs corresponding fine-tuning again according to the new fine-tuning parameters, and the control module 5 determines again whether the image quality index before and after the new fine-tuning is optimized. If no, it means that the adjustment of the camera platform and / or the camera device is completely completed.

[0033] Optionally, the smart camera platform also has a matching App (application program), and the two communicate with each other and transmit data. The matching App is installed on the user's smart electronic device, which is convenient for the user to display the current shooting picture in real time in the App of the smart electronic device carried by the user, and the user can also confirm or manually adjust before the final shooting.

[0034] See also Figure 2 , which is the method proposed in this application for the intelligent camera platform to realize automatic tracking and adjustment of the position and angle of the camera device, including the following steps.

[0035] Step S1: The camera device is fixed on the camera platform, and the camera device collects the shooting picture in real time. The recognition and tracking module 1 automatically recognizes the central target in the shooting picture, or the user can select the central target from the shooting picture instead. The recognition and tracking module 1 also automatically tracks the central target in the shooting picture.

[0036] Step S2: If the recognition and tracking module 1 finds that the position of the central target in the camera image moves, the parameter estimation module 3 gives coarse adjustment parameters that can make the central target be located near the optimal shooting position of the shooting image based on the position of the central target in the current shooting image.

[0037] Step S3: the adjustment module 4 performs a corresponding coarse adjustment on the position and / or angle of the camera platform and / or the camera device according to the coarse adjustment parameters, and after the coarse adjustment, the central target is located near the optimal shooting position of the shooting picture.

[0038] If the recognition and tracking module 1 does not find that the position of the central target in the camera image moves, step S2 and step S3 are omitted.

[0039] Step S4: the image quality calculation module 2 calculates the image quality index of the current (after coarse adjustment, or initially when no coarse adjustment is required) captured image.

[0040] Step S5: the parameter estimation module 3 provides fine-tuning parameters that can improve the quality index of the captured image according to the current captured image.

[0041] Step S6: the position adjustment module 4 performs corresponding fine adjustment on the position and / or angle of the imaging platform and / or the imaging device according to the fine adjustment parameters.

[0042] Step S7: the image quality calculation module 2 calculates the image quality index of the fine-tuned captured image.

[0043] Step S8: The control module 5 determines the image quality indexes before and after fine-tuning, which are obtained from steps S4 and S7 respectively.

[0044] If the image quality index after fine-tuning is better than the image quality index before fine-tuning, for example, the image quality index after fine-tuning - the image quality index before fine-tuning ≥ the first threshold Q TN , return to step S5 to calculate the fine-tuning parameters again, and repeat steps S6 to S8.

[0045] If the image quality index after fine-tuning is not better than the image quality index before fine-tuning, for example, the image quality index after fine-tuning - the image quality index before fine-tuning < the first threshold Q TN , it means that the adjustment of the camera platform and / or the camera equipment is completely completed.

[0046] Alternatively, the step S8 further sets a second threshold C Th , which is preferably a positive integer within 5. If the image quality index after fine-tuning is not better than the image quality index before fine-tuning, the value of the adjustment calculator C is failed. F Add 1. If it fails, adjust the value of the calculator C F <Second threshold C Th , return to step S5 to calculate the adjustment parameter again, and repeat steps S6 to S8. If the value C of the adjustment calculator fails F ≥ the second threshold C Th , indicating that the adjustment of the camera platform and / or the camera equipment is completely completed.

[0047] Optionally, in the steps S1 to S8, the supporting App of the smart camera platform synchronously displays the current shooting picture of the camera device in the supporting App of the user's smart electronic device. The supporting App prompts the user through text or voice to decide whether the current shooting picture is satisfactory and whether manual adjustment is required before the interactive adjustment countdown ends. If no feedback is received from the user when the interactive adjustment countdown is completed, it is assumed that the user is satisfied with the current shooting picture and is ready to actually shoot.

[0048] Optionally, after step S8 is completed, the intelligent camera platform issues a voice prompt that the shooting is about to begin, prompting the subject to pay attention to the lens or change posture to cooperate with the shooting, and then controls the camera device to trigger the shooting to obtain high-quality photos and videos.

[0049] The intelligent camera platform proposed in this application can locate and track the central target, and displace and / or rotate to a suitable position, so as to move the central target of interest to the best position for shooting the picture. At the same time, the intelligent camera platform proposed in this application can improve the image quality in terms of light, clarity, etc. by evaluating the image quality, realizing the composition of the central target at the best position of the shooting picture. The intelligent camera platform can be used with camera devices such as smart phones, digital cameras, and tablet computers, and has a supporting App to enrich the user experience.

[0050] The above are only preferred embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should all be included in the protection scope of the present application.

Claims

1. An intelligent camera platform that automatically tracks and adjusts position angles, characterized in that: include: Identification and tracking module, image quality calculation module, parameter estimation module, adjustment module, control module; The recognition and tracking module is used to automatically recognize the central target in the shooting picture of the camera device, and is also used to automatically track the central target in the shooting picture of the camera device; The image quality calculation module is used to calculate the quality index of the current captured image in real time, referred to as the image quality index; The parameter estimation module is used to use a trained neural network, called a parameter estimation network, to obtain a coarse adjustment parameter that can move the central target to a vicinity of an optimal shooting position in the shooting picture according to the position of the central target in the current shooting picture; and is also used to use the trained parameter estimation network to obtain a fine adjustment parameter that can improve the quality index of the shooting picture according to the current shooting picture; the coarse adjustment parameter and the fine adjustment parameter both include any one or more of the following: a horizontal plane displacement parameter of the camera platform, an azimuth rotation parameter of the camera platform, an up-and-down displacement parameter of the camera device, and a pitch angle rotation parameter of the camera device; The adjustment module is used to drive the imaging platform and / or the imaging device to perform corresponding coarse adjustment or fine adjustment based on the coarse adjustment parameter or the fine adjustment parameter; The control module is used to determine whether the image quality index after fine-tuning by the adjustment module is better than the image quality index before fine-tuning; if so, the parameter estimation module calculates the fine-tuning parameters again, and the adjustment module performs corresponding fine-tuning according to the new fine-tuning parameters again, and the control module determines again whether the image quality index before and after the new fine-tuning is optimized; if not, it indicates that the adjustment is completed.

2. The intelligent camera platform for automatic tracking and adjusting position angle according to claim 1 is characterized in that: The identification and tracking module is changed to allow the user to select a central target in the shooting picture of the camera device.

3. The intelligent camera platform for automatic tracking and adjusting position angle according to claim 1 is characterized in that: The image quality calculation module uses a neural network technology, called an image quality calculation network, to calculate the quality index of the current captured image by combining the objective index of image quality and the subjective preference of the user; The objective indicators of image quality include: any one or more of image brightness, contrast, and clarity; The user's subjective preference includes: any one or more of the photographer's preference for color saturation, the size ratio of people in the landscape, and the aesthetics of composition; The image quality calculation network first obtains a basic image quality calculation network that is not affected by the user's subjective preferences by learning samples, and then further obtains a personalized image quality calculation network that meets the subjective preferences of different users by learning user feedback during the use of the basic image quality calculation network.

4. The intelligent camera platform for automatic tracking and adjusting position angle according to claim 1 is characterized in that: The input of the parameter estimation network includes: the current shooting picture and the position of the central target in the current shooting picture; the output of the parameter estimation network includes: the displacement direction of the next step of the camera platform, the displacement step of the camera platform, the azimuth angle of the camera platform, the lifting height of the robotic arm, and the pitch angle of the camera lens.

5. The intelligent camera platform for automatic tracking and adjusting position angle according to claim 4 is characterized in that: The training data set used by the parameter estimation network includes photo samples of different scenes, lighting conditions and shooting distances; it also includes adjustment parameters corresponding to each photo sample - any one or more of the optimal displacement direction, step size, azimuth angle, lifting height and pitch angle; the adjustment parameters are used to improve the position of the central target in the photo sample and improve the image quality index of the photo sample; each photo sample and the corresponding adjustment parameter constitute a data pair, and all data pairs constitute the training data set.

6. The intelligent camera platform for automatic tracking and adjusting position angle according to claim 4 is characterized in that: The parameter estimation network is implemented by a convolutional neural network for extracting image features; during the training process, the parameter estimation network uses a back propagation algorithm to optimize the weights of the neural network and adopts batch gradient descent or Adam optimizer for training.

7. The intelligent camera platform for automatic tracking and adjusting position angle according to claim 4 is characterized in that: The adjustment module drives the camera platform to move in the horizontal plane according to the displacement direction and displacement step of the camera platform, drives the camera platform to rotate left and right in the horizontal plane according to the azimuth rotation parameter of the camera platform, drives the camera device to rise and fall according to the lifting height of the camera device, and drives the camera device to rotate up and down according to the pitch angle rotation parameter of the camera device; The adjustment module includes a translation unit and a mechanical arm; the translation unit is used to drive the entire camera platform together with the camera equipment to move in the horizontal plane and rotate left and right; the mechanical arm is used to mount the camera equipment and drive the camera equipment to move up and down and rotate up and down.

8. The intelligent camera platform for automatic tracking and adjusting position angle according to claim 1 is characterized in that: The smart camera platform also has a matching application App, and the two communicate with each other and transmit data; the matching App is installed on the user's smart electronic device, and the matching App displays the current shooting picture in real time to accept the user's confirmation or adjustment.

9. A method for realizing automatic tracking and adjustment of the position and angle of a camera device by an intelligent camera platform, characterized in that: The method comprises the following steps: Step S1: The camera device is fixed on the camera platform, and the camera device collects the shooting picture in real time. The recognition and tracking module automatically recognizes the central target in the shooting picture, and also automatically tracks the central target in the shooting picture; Step S2: If the recognition and tracking module finds that the position of the central target in the camera image moves, the parameter estimation module gives a coarse adjustment parameter that can make the central target be located near the optimal shooting position of the shooting image according to the position of the central target in the current shooting image; Step S3: the adjustment module performs a corresponding coarse adjustment on the position and / or angle of the camera platform and / or the camera device according to the coarse adjustment parameters, and after the coarse adjustment, the central target is located near the optimal shooting position of the shooting picture; If the recognition and tracking module does not find that the position of the central target in the camera image has moved, step S2 and step S3 are omitted; Step S4: the image quality calculation module calculates the image quality index of the current captured image; Step S5: the parameter estimation module provides fine-tuning parameters that can improve the quality index of the captured image according to the current captured image; Step S6: the position adjustment module performs corresponding fine adjustment on the position and / or angle of the imaging platform and / or the imaging device according to the fine adjustment parameters; Step S7: the image quality calculation module calculates the image quality index of the fine-tuned captured image; Step S8: The control module determines the image quality index before and after fine-tuning; if the image quality index after fine-tuning is better than the image quality index before fine-tuning, return to step S5 to calculate the fine-tuning parameters again, and repeat steps S6 to S8; if the image quality index after fine-tuning is not better than the image quality index before fine-tuning, the adjustment of the camera platform and / or camera equipment is completed.

10. The method for realizing automatic tracking and adjusting the position and angle of a camera device by an intelligent camera platform according to claim 9, characterized in that: A second threshold is set in step S8; if the image quality index after fine-tuning is not better than the image quality index before fine-tuning, the value of the failure adjustment calculator is increased by one; if the value of the failure adjustment calculator is less than the second threshold, return to step S5 to calculate the adjustment parameters again, and repeat steps S6 to S8; if the value of the failure adjustment calculator is greater than or equal to the second threshold, it indicates that the adjustment of the camera platform and / or the camera device is completed.

Images