An adaptive camera posture intelligent adjustment system

Abstract

The present invention relates to the field of intelligent control technology, and specifically to an adaptive camera posture intelligent adjustment system, which includes a dynamic recognition module, a posture adjustment module, a mechanism control module, an information filtering module, and a feedback association module. In the present invention, by extracting the brightness gradient and boundary continuity of foreground pixels in an image frame, the target displacement path can be identified and tracking information can be constructed to ensure consistent perception of the target under rapid scene changes. By combining posture parameters with the target movement direction and speed, offset correction instructions are generated to improve the response stability to dynamic targets. Execution structure feedback is used to correct control outputs and reduce the risk of accumulated driving errors. By improving image boundary clarity and motion consistency, non-target disturbances are eliminated, data purity is improved, and by combining perspective execution status and path information, synchronous control of tracking and perspective adjustment is achieved, thereby enhancing the response efficiency and following robustness of the system in a dynamic environment.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control technology, and in particular to an adaptive camera posture intelligent adjustment system. Background Art

[0002] The field of intelligent control technology encompasses automated management of equipment and processes, state identification, and response regulation. The core of this technology lies in acquiring information about the external environment or the internal state of a system through sensing elements, analyzing and judging this data based on pre-set logic or models, and thereby driving actuators to complete corresponding operations. This field systematically involves multiple steps, including sensor signal acquisition, control logic calculation, and drive execution and response. It is widely used in a variety of application scenarios, including industrial automation, smart homes, traffic control, and robotic navigation. Its primary goal is to improve the stability, efficiency, and intelligent responsiveness of system operations.

[0003] The adaptive camera posture intelligent adjustment system refers to a control system that can automatically adjust the camera's posture based on the environment or target state. To address issues such as shooting angle deviation, blind spots, or target tracking offsets in dynamic scenes, the system adjusts the camera's posture through image information analysis, spatial geometric parameter calculation, and coordinated control of actuators. The system determines the rationality of the current camera's perspective based on image edge recognition and target contour determination, then corrects the three-dimensional posture parameters based on the angle change pattern and position change trajectory. The system automatically adjusts the camera's pitch angle, horizontal rotation angle, and zoom distance through a multi-axis drive actuator, achieving continuous perspective tracking and posture optimization of the target.

[0004] Existing technologies for automatic camera perspective adjustment primarily rely on static analysis of edges and target outlines in single-frame images, lacking in-depth analysis of the target's continuous motion path within an image sequence. This makes it difficult to maintain target lock stability in complex backgrounds or scenes with multiple interfering targets. Regarding posture parameter adjustment, existing systems use a pre-set angle change model for correction, failing to fully incorporate changes in the target's motion direction and speed. This leads to response lag and even loss of perspective during rapid scene switching or sudden target movement. The feedback execution phase fails to capture the real-time response state of the execution structure in a closed loop, causing drive deviations to accumulate during continuous adjustments, manifesting as slight drift or even distortion in camera posture adjustments. The dynamic separation of foreground and background areas in image information is weak, often leading to misidentification due to motion interference from non-target objects in the background, reducing target tracking accuracy. For example, in complex street scene surveillance, where multiple targets, such as pedestrians and vehicles, are present simultaneously, the system is unable to eliminate disturbances from non-critical targets, resulting in frequent jumps in posture adjustment, affecting video continuity and analysis stability. This limits the real-time performance, robustness, and control accuracy of existing technologies in dynamic environments. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an adaptive camera posture intelligent adjustment system.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solution: an adaptive camera posture intelligent adjustment system includes:

[0007] The dynamic recognition module obtains camera pose data, extracts video frames captured by the image acquisition component, extracts pixel brightness gradients and boundary contour continuity in the foreground area of the frame, identifies target position transformation data in the previous and next frames within the image stabilization structure, determines the motion trend of the area block and identifies the continuous motion path, and generates target tracking path information;

[0008] The posture adjustment module extracts the current pitch and rotation angle values of the camera based on the target tracking path information, analyzes the posture deviation degree and the required angle adjustment amplitude in combination with the target movement direction and speed range, and generates a posture correction instruction set;

[0009] The mechanism control module collects the current response state of the execution structure according to the posture correction instruction set, compares the target angle and displacement requirements in the instruction, identifies the drive error range, adjusts the controller output amplitude and direction, and obtains the mechanism execution adjustment value;

[0010] The information filtering module calls the mechanism to execute the adjustment value, extracts the boundary clarity and motion consistency features of the foreground and background areas in the image, identifies the proportion of non-target dynamic areas, and removes the boundary noise areas to generate background disturbance elimination results.

[0011] As a further solution of the present invention, the target tracking path information includes the position change amplitude, trajectory curvature characteristics, and displacement direction angle; the posture correction instruction set includes the pitch angle adjustment amount, the rotation angle correction value, and the angle offset threshold; the mechanism execution adjustment value includes the controller output gain, the execution structure rotation angle, and the real-time response displacement; the background disturbance elimination result includes the foreground boundary clarity score, the background area consistency coefficient, and the dynamic interference ratio.

[0012] As a further solution of the present invention, the dynamic recognition module includes:

[0013] The pose estimation submodule obtains the camera pose data, extracts the camera's 3D pose data and center pixel coordinates for each frame, analyzes the pose rotation angle difference and translation vector change rate in consecutive frames, determines whether the change exceeds the pose stability threshold, and generates a pose adjustment trigger signal;

[0014] The foreground perception submodule extracts the brightness gradient and boundary contour continuity of the foreground in the current frame based on the posture adjustment trigger signal, and combines the spatial offset and direction change of the target edge points in the previous and next frames to determine the perceptual integrity trend of the foreground area under the posture change and obtain the image perception integrity parameter;

[0015] The path linkage submodule dynamically corrects the current posture according to the image perception completeness parameter, adjusts the camera focus direction and rotation angle, calls the offset difference between the center point position of the corrected image and the original image, and obtains the posture adjustment execution path.

[0016] As a further solution of the present invention, the posture adjustment module includes:

[0017] The angle extraction submodule extracts the pitch angle and rotation angle data in the image frame based on the target tracking path information, identifies the pitch angle and rotation angle by combining the pixel matrix change and the image edge position sequence, and obtains the angle change value set;

[0018] The direction matching submodule calls the angle change value set, detects the direction consistency and speed change trend according to the target moving direction vector and the speed range, calculates the posture direction offset strength value, and obtains the direction offset adjustment parameter value by combining the direction matching result and the speed jump range;

[0019] The instruction generation submodule calls the direction offset adjustment parameter value, divides the gear according to the offset amplitude and speed range, determines whether it is in the adjustment response range, maps the correction action parameters, and generates a posture correction instruction set.

[0020] As a further solution of the present invention, the mechanism control module includes:

[0021] The state recognition submodule collects the current response state of the execution structure according to the posture correction instruction set, extracts the posture angle vector and displacement component, identifies the dimension deviation distance, and generates the posture displacement difference;

[0022] The error evaluation submodule calls the posture displacement difference, determines the matching offset, extracts the dynamic change amplitude and response trend in the response state, compares the change rate with the response rate threshold, analyzes the driving accuracy difference interval, and obtains the driving error range value;

[0023] The control adjustment submodule calls the driving error range value, combines the angle response sensitivity, displacement response hysteresis amplitude, and adjustment feedback cycle amplitude, corrects the adjustment amplitude and direction of the controller output, calculates the control output correction amplitude, corrects the controller signal, and obtains the mechanism execution adjustment value.

[0024] As a further solution of the present invention, the information filtering module includes:

[0025] The clarity extraction submodule calls the mechanism to execute the adjustment value, extracts the grayscale gradient of the image boundary, screens the pixels that meet the clarity threshold, and obtains the boundary clarity change rate;

[0026] The dynamic ratio judgment submodule calls the boundary clarity change rate, detects the consistency of the regional motion vector, and calculates the area ratio of the non-target area to obtain the non-target area ratio value;

[0027] The disturbance elimination submodule collects the boundary noise pixel distribution according to the non-target area ratio, performs comparison based on the foreground and background grayscale difference and error amplitude, calculates the disturbance elimination coefficient, filters out the noise area, extracts the mask layer, and generates the background disturbance elimination result.

[0028] As a further solution of the present invention, the system further includes a feedback association module:

[0029] The feedback association module evaluates the response synchronization between the camera view movement and the target path based on the background disturbance elimination result, combined with the updated target tracking path and the current posture execution state, and generates a camera posture response matching degree;

[0030] The camera posture response matching degree includes target path tracking deviation, posture adjustment delay, and synchronization coordination index.

[0031] As a further solution of the present invention, the feedback association module includes:

[0032] The view offset extraction submodule extracts the updated camera pose sequence and the starting pose data after background disturbance elimination based on the background disturbance elimination result, analyzes the pose Euler angle difference and displacement change between consecutive frames, and performs normalization processing based on the timestamp to obtain the camera view offset trend;

[0033] The path linkage comparison submodule analyzes the cosine value of the direction vector angle and the velocity amplitude change rate at the corresponding time point based on the camera view offset trend and the spatial direction change sequence of adjacent posture segments in the target tracking path, and determines the response lag in the synchronization interval to obtain the path response synchronization characteristics;

[0034] The adjustment strategy generation submodule calls the path response synchronization characteristics and the camera speed change rate at the previous moment, sets the proportion parameters according to the response lag time difference and the angle offset degree, performs weighted processing on each posture adjustment variable, and generates the camera posture response matching degree.

[0035] Compared with the prior art, the advantages and positive effects of the present invention are:

[0036] In the present invention, by dynamically extracting the pixel brightness gradient and boundary contour continuity of the foreground area in the image frame, it is possible to accurately identify the target displacement path in a stable structure, and construct real-time tracking path information based on the target's continuous motion trend, thereby maintaining a consistent perception of the target position under rapid scene changes. The current camera posture parameters are extracted, and combined with the target movement direction and speed range, offset correction instructions for pitch and rotation angles are generated, allowing the camera to quickly adapt to changes in target behavior and improve the tracking stability of dynamic targets. In response to the feedback data of the execution structure, the expected and actual angle deviations are compared, and the control output amplitude and direction are further adjusted to effectively avoid control drift caused by accumulated drive errors. Based on the boundary clarity and motion consistency characteristics of the foreground and background areas of the image, the disturbance influence of non-target areas is eliminated, the specificity of target recognition is enhanced, and the data purity of posture adjustment decisions is guaranteed. Combined with the actual execution status of the camera and the updated information of the target path, the synchronization of the perspective response is evaluated, and real-time closed-loop control between the tracking path and perspective adjustment is achieved. The above operations work together in a multi-level link of target capture, attitude correction, response adjustment and interference elimination, greatly improving the system's ability to quickly recognize dynamic scenes, the accuracy of real-time attitude adjustment, and the robustness and visual consistency during target following. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a system flow chart of the present invention;

[0038] Figure 2 This is a flow chart of the dynamic identification module in the present invention;

[0039] Figure 3 This is a flow chart of the posture adjustment module in the present invention;

[0040] Figure 4 This is a flow chart of the mechanism control module in the present invention;

[0041] Figure 5 This is a flow chart of the information filtering module in the present invention;

[0042] Figure 6 This is a flow chart of the feedback association module in the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0044] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0045] See also Figure 1 , an adaptive camera posture intelligent adjustment system includes:

[0046] The dynamic recognition module obtains camera pose data, extracts video frames captured by the image acquisition component, extracts pixel brightness gradients and boundary contour continuity in the foreground area of the frame, identifies target position transformation data in the previous and next frames within the image stabilization structure, determines the motion trend of the area block and identifies the continuous motion path, and generates target tracking path information;

[0047] The attitude adjustment module extracts the current pitch and rotation angle values of the camera based on the target tracking path information, analyzes the attitude deviation degree and the required angle adjustment range based on the target movement direction and speed range, and generates a posture correction instruction set;

[0048] The mechanism control module collects the current response status of the execution structure according to the posture correction instruction set, compares the target angle and displacement requirements in the instruction, identifies the drive error range, adjusts the controller output amplitude and direction, and obtains the mechanism execution adjustment value;

[0049] The information filtering module calls the mechanism to execute the adjustment value, extracts the boundary clarity and motion consistency features of the foreground and background areas in the image, identifies the proportion of non-target dynamic areas, and removes the boundary noise areas to generate the background disturbance elimination result;

[0050] The feedback association module evaluates the response synchronization between the camera view movement and the target path based on the background disturbance elimination results, combined with the updated target tracking path and the current posture execution status, and generates the camera posture response matching degree.

[0051] The target tracking path information includes the position change amplitude, trajectory curvature characteristics, and displacement direction angle. The posture correction instruction set includes the pitch angle adjustment amount, rotation angle correction value, and angle offset threshold. The mechanism execution adjustment value includes the controller output gain, the execution structure rotation angle, and the real-time response displacement. The background disturbance elimination results include the foreground boundary clarity score, the background area consistency coefficient, and the dynamic interference ratio. The camera posture response matching degree includes the target path tracking deviation, the posture adjustment delay, and the synchronization coordination index.

[0052] See also Figure 2 , the dynamic recognition module includes:

[0053] The pose estimation submodule obtains the camera pose data, extracts the camera's 3D pose data and center pixel coordinates for each frame, analyzes the pose rotation angle difference and translation vector change rate in consecutive frames, determines whether the change exceeds the pose stability threshold, and generates a pose adjustment trigger signal;

[0054] To acquire 3D camera pose data, the camera position and rotation angle must first be determined for each frame. This data is acquired using sensors (such as an IMU) and computer vision methods. For example, visual SLAM technology is used to calculate the pose by matching image feature points between consecutive frames. Specifically, for each frame, feature points are extracted and a least-squares optimization algorithm is applied to calculate the camera pose for the current frame, which includes a 3D translation vector and rotation matrix. After extracting the pose data, the rotation angle and translation vector differences between adjacent frames must be calculated. Rotation angles can be represented using quaternions or Euler angles, while translation vectors are directly calculated as the difference in 3D coordinates between adjacent frames. A preset stability threshold is used, such as a rotation angle difference greater than 0.5 degrees and a translation vector difference greater than 0.1 meter, to deem a significant change. If a pose change exceeding the threshold is detected, a pose adjustment signal is triggered, at which point the camera position and orientation can be reassessed to ensure stable operation. For example, if the rotation angle difference between two consecutive frames is 0.6 degrees and the translation difference is 0.12 meters, a pose adjustment signal is triggered.

[0055] The foreground perception submodule extracts the brightness gradient and boundary contour continuity of the foreground in the current frame based on the posture adjustment trigger signal. It then combines the spatial offset and directional changes of the target edge points in the previous and next frames to determine the perceptual integrity trend of the foreground area under posture changes and obtain the image perception integrity parameters.

[0056] Analyze the foreground brightness gradient and boundary contour continuity. Extract the foreground area of the current frame based on the trigger signal of the posture change, and then calculate the brightness gradient of each pixel in the foreground area. This can be achieved through image convolution operation. Use the Sobel operator or Canny edge detection to obtain the gradient value of the brightness change. Analyze the boundary contour of the foreground area and use the image contour detection algorithm, such as Canny edge detection, to extract the contour edge and calculate the continuity of the contour. If the connectivity between the boundary points is weak, it means that the perceived continuity of the foreground area is poor. On this basis, combine the spatial offset and direction change of the target edge points in the previous and next frames, and perform multi-frame analysis to determine the perceptual integrity trend of the foreground area under posture changes. For example, if the spatial displacement of the edge points in the previous and next frames exceeds 0.15 meters and the direction of the edge changes significantly, it means that the integrity of the foreground perception is reduced. The image perception integrity parameter will be updated to reflect the perception loss of the foreground area.

[0057] The path linkage submodule dynamically corrects the current posture according to the image perception completeness parameters, adjusts the camera focus direction and rotation angle, and calls the offset difference between the center point position of the corrected image and the original image to obtain the posture adjustment execution path;

[0058] The current camera pose is dynamically corrected based on the image perception completeness parameter. First, the camera's orientation is determined to have changed significantly based on the aforementioned image perception completeness analysis. If the image perception completeness is low, the current pose needs to be adjusted. The adjustment mechanism is triggered based on the perception completeness value (for example, when the completeness falls below 50%). The camera's orientation can be adjusted by controlling the rotation angle, which varies based on the magnitude of the change in perception completeness. If the completeness drops by more than a certain threshold (for example, a 20% drop), the rotation angle is adjusted using a dynamic algorithm. Correction is performed based on the image center point and offset difference. A new correction path is calculated using the offset difference between the adjusted image center point and the original image. For example, if the original image offset is 0.05 meters and the adjusted image center point offset is 0.07 meters, the path adjustment difference is 0.02 meters, indicating that the camera needs to adjust its position along this path. This determines the pose adjustment execution path, ensuring improved image quality and perception completeness after the pose adjustment.

[0059] See also Figure 3 , the attitude adjustment module includes:

[0060] The angle extraction submodule extracts the pitch and rotation angle data from the image frame based on the target tracking path information, combines the pixel matrix changes with the image edge position sequence, identifies the pitch and rotation angles, and obtains the angle change value set;

[0061] The data is calculated by analyzing the spatial position changes of the target in consecutive video frames. The data is used to understand how the target moves in three-dimensional space. For example, in security monitoring, the camera needs to adjust its angle in real time to keep focusing on the intruder. First, the edge and pixel matrix change data are extracted from each frame of the image. This involves edge detection and optical flow estimation in image processing technology. Through this technology, the moving object in the image as well as its speed and direction can be identified. By analyzing the change trend and position data in the image sequence, the pitch and rotation angle change values of the target are calculated to adjust the pitch and rotation angle of the camera in real time to ensure that the camera's field of view remains aligned with the moving target. For example, in a security scenario, if a person passes through the monitoring area from left to right, the camera's rotation angle will be adjusted according to the person's movement speed and trajectory to keep the target in the center of the field of view, generating an angle change value set, which is a series of pitch and rotation angle values that reflects how the camera should adjust its angle from one point in time to another to track the target.

[0062] The direction matching submodule calls the angle change value set and detects the direction consistency and speed change trend based on the target moving direction vector and speed range. The formula is:

[0063] ;

[0064] Calculate the attitude direction offset strength value, combine the direction matching result with the speed jump interval, and obtain the direction offset adjustment parameter value;

[0065] in, Represents the attitude direction offset strength value, represents the expected value of the path velocity, Represents the current speed value, is the desired direction angle, is the current rotation angle change value, is the desired pitch angle, is the current pitch angle change value, Represents path frame The position value of Represents path frame The position value of is the number of path frames;

[0066] After calling the angle change value set, it is necessary to numerically express the changes in the camera's pitch angle and rotation angle in the current frame, and analyze them in conjunction with the target's actual moving direction, speed difference and other data to form the attitude direction offset strength value and the target speed. The path node position change between consecutive frames can be obtained by dividing the time interval. Let the path node position point , , sampling interval , then the speed is calculated as , for example, when ,but ;

[0067] Current speed Derived from the real-time calculation of the instantaneous change of the path, collecting the current time point The spatial position difference between the previous and next frames, let The front and back frames are and , the sampling interval is still ,but ;

[0068] Expected pitch angle The target trajectory simulation path is calculated and set here. , expected value of rotation angle , the current change value of the camera is , , then the angle difference is calculated as ;

[0069] The changes between the position values of the nodes in consecutive frames of the path are calculated by accumulating the absolute values of the inter-frame differences and then averaging them;

[0070] Assume there is A sequence of nodes, ;

[0071] but , the average path change is ;

[0072] Substituting the above data into the formula: ;

[0073] The result shows that the attitude direction offset strength value is 7.12, which means that there is a significant offset trend in the current frame path direction, and it is necessary to enter the adjustment process to match the corresponding offset correction instruction set. In this calculation, the units of each parameter are unified into SI units: speed is , the angle is uniformly converted to dimensionless when it is input in degrees for geometric difference operation, and the spatial position is is the basic unit, the final result is the comprehensive value of the offset strength, and the dimension is a scalar unitless;

[0074] By introducing the coupled expression of three physical variables: speed difference, directional angle error, and path continuity change, we avoid dependence on single angle or speed data and enhance the response adaptability in the complex background of dynamic trajectories.

[0075] The instruction generation submodule calls the direction offset adjustment parameter value, divides the gear according to the offset amplitude and speed range, determines whether it is in the adjustment response range, maps the correction action parameters, and generates the posture correction instruction set;

[0076] The direction offset adjustment parameter value is converted into a specific camera adjustment instruction. The parameter is calculated and provided by the previous sub-module. The camera posture needs to be adjusted based on the offset parameter and the current speed range. This involves determining whether the current offset exceeds the preset response threshold. For example, in urban traffic monitoring, if a pedestrian or vehicle is detected to suddenly change direction, the camera needs to quickly adjust its angle to track the target and select the appropriate adjustment gear and parameter mapping. The gear is preset according to different speeds and offsets to optimize the camera's response speed and accuracy. For example, in a shopping mall monitoring, when a child is monitored moving in a crowd, the camera will select a more sensitive adjustment gear to track a fast-moving small target, generate a posture correction instruction set, and guide the camera to adjust its physical position to ensure that the camera can effectively track and record dynamic changes in the scene.

[0077] See also Figure 4 , the mechanism control module includes:

[0078] The state recognition submodule collects the current response state of the execution structure according to the posture correction instruction set, extracts the posture angle vector and displacement component, identifies the dimension deviation distance, and generates the posture displacement difference;

[0079] The starting position is based on the target angle and displacement requirements. By calculating the deviation distance in each dimension, this process is aimed at the position adjustment scenario of the robotic arm in the automated assembly line. During implementation, the robotic arm first collects the current posture data, such as angle and position coordinates, against the preset target state, and then feeds back real-time data through the sensor to calculate the difference between the current state and the target state. The calculation includes the angle difference and displacement difference of each axis. The specific value of each difference is determined by the algorithm. For example, the angle error is calculated as the angle Δθ, and the displacement error is Δx, Δy and Δz. The data is then used to adjust the control instructions to ensure that the robotic arm can accurately reach the preset position. Through continuous monitoring and real-time adjustment, the error rate can be significantly reduced, the assembly accuracy can be improved, and the posture displacement difference is generated. This is a specific numerical data representing the specific angle and displacement that the robotic arm needs to adjust to ensure consistency with the predetermined target.

[0080] The error assessment submodule calls the attitude displacement difference, determines the matching offset, extracts the dynamic change amplitude and response trend in the response state, compares the change rate with the response rate threshold, analyzes the driving accuracy difference interval, and obtains the driving error range value;

[0081] The starting position calls the attitude displacement difference. For the error correction scenario in automated navigation, the obtained attitude displacement difference is first used for preliminary analysis to determine how to adjust the algorithm parameters according to the size of the quantity. For example, if the displacement difference exceeds the preset response rate threshold, the proportion and direction that need to be adjusted will be calculated, and then the path planning will be optimized through dynamic adjustment of the algorithm. This adjustment is not only based on real-time data, but also considers the original data pattern to predict future trends and errors, thereby achieving more accurate navigation. The calculation in this process involves multiple parameters, such as thresholds, change rates, and quantitative standards for continuous trends. Actual data-driven simulations are used to ensure real-time updating and application of parameters. The final generated driving error range value is an accurate data that indicates the optimal error range that can be achieved under current technical conditions.

[0082] The control adjustment submodule calls the drive error range value, combines the angle response sensitivity, displacement response hysteresis amplitude, and adjustment feedback cycle amplitude, and corrects the adjustment amplitude and direction of the controller output using the formula:

[0083] ;

[0084] Calculate the control output correction amplitude, correct the controller signal, and obtain the mechanism execution adjustment value;

[0085] in, Indicates the control output correction amplitude, Represents the driving error range value, represents the angular response sensitivity, represents the displacement response hysteresis amplitude, represents the direction correction weight vector, Representative feedback cycle The amplitude of the adjustment feedback cycle, Representative feedback cycle The table response lag compensation ratio, is the dimension of the number of feedback cycles;

[0086] The driving error range value is called at the starting position. The control factor is constructed by combining the angle response sensitivity, displacement response hysteresis, and adjustment feedback cycle amplitude. The adjustment amplitude and direction of the controller output are corrected. The control output correction amplitude E is calculated using the formula. This process is applied to scenarios such as dynamic adjustment of the camera's real-time posture control in adaptive camera operation. To ensure consistency in the calculations between the parameters, all parameters are first dimensionally unified. The angle unit is unified into radians, the displacement unit is unified into millimeters, and the time-related parameters are unified into seconds. The following describes how to obtain each parameter and the example value, and then substitutes them into the formula for a complete calculation:

[0087] A: Drive error range value, which represents a comprehensive measure of the deviation of the camera's current posture from the target posture. It is determined by the angle difference and displacement difference. It is obtained by calculating the difference between the target and actual response on each axis and averaging the squares. Assuming the actual angle deviation is 0.07 rad and the displacement deviation is 1.5 mm, A=1.85 is set according to the comprehensive weighting rule (uniformly in numerical dimensions).

[0088] R: Angle response sensitivity, obtained by the change in angle response caused by unit input excitation. After adjusting the control input by 0.1 unit, the angle changes by 0.015rad, and R=0.015;

[0089] B: Displacement response hysteresis amplitude, which indicates the hysteresis of displacement response after the control output changes. It can be measured by the average response delay per unit control output. If the average response hysteresis is 2.1mm, set B=2.1;

[0090] F: Direction correction weight vector, used to adjust the error deflection direction. It is quantified by the original response deviation feature clustering result, and the main deviation direction can be quantified to 1.2;

[0091] : Adjust the feedback cycle amplitude, which represents the amount of change in the control signal within each feedback cycle. Set the system for five consecutive feedback cycles, and adjust the control amount per cycle to 0.8, 0.9, 1.0, 1.1, and 1.2 respectively. The unit is unified as unit signal strength.

[0092] : Response hysteresis compensation ratio, indicating the hysteresis error compensation amount per cycle, in the order of 0.3, 0.4, 0.35, 0.45, and 0.5, and the unit is the error compensation unit;

[0093] Here are the steps:

[0094] Calculate the square root: ;

[0095] Weighted item: (A + square root item) ;

[0096] Multiply by F: ;

[0097] Calculate the denominator:

[0098] ;

[0099] Finally, the control output correction amplitude is obtained ;

[0100] The results show that under the current response state, the controller output should be dynamically adjusted with an amplitude of 0.677 units, and the posture correction should be performed in the direction indicated by the direction correction vector F to obtain the mechanism execution adjustment value;

[0101] The benefit of the formula is that by integrating the angle response sensitivity R and the displacement hysteresis amplitude B into the square root operation, the system's sensitivity to the two sources of deviation is enhanced, and the actual control input is used to calculate the system's error. and compensation coefficient It is equalized based on the reference, and can dynamically adapt the output signal strength in a multi-cycle feedback structure to form a stable and controllable adjustment mechanism.

[0102] See also Figure 5 , the information filtering module includes:

[0103] The clarity extraction submodule calls the mechanism to execute the adjustment value, extracts the grayscale gradient of the image boundary, filters the pixels that meet the clarity threshold, and obtains the boundary clarity change rate;

[0104] The calling mechanism performs the adjustment. This adjustment is based on edge detection technology within image processing algorithms. By extracting the grayscale gradient change of the boundary in the image, the technique is used to locate the clear and blurred boundaries in the image. In surveillance cameras, such as those used for traffic monitoring, this technology can identify the clear boundaries of vehicles and road signs, ensuring that traffic management obtains accurate visual information. The sharpness change between adjacent boundary pixels is calculated and compared against a preset sharpness change threshold. If the sharpness change between pixels exceeds the threshold, the pixel is identified as a clear boundary. The grayscale gradient of the boundary pixel is compared with the threshold to determine whether the clear boundary condition is met. For example, in a specific surveillance scenario, the threshold can be set to 20% of the grayscale change. When the grayscale gradient change of the boundary pixel exceeds this threshold, the pixel is marked and used for further analysis. In this way, pixels that meet the sharpness threshold can be screened out, effectively improving image processing quality and monitoring accuracy. Through refined image processing, the boundary sharpness change rate is obtained. This data can be used to further optimize the image processing process and improve the accuracy and efficiency of image recognition.

[0105] The dynamic ratio judgment submodule calls the boundary clarity change rate, detects the consistency of regional motion vectors, and calculates the area ratio of non-target areas to obtain the non-target area ratio value;

[0106] The boundary sharpness change rate is used to monitor the motion vectors of corresponding pixels between image frames. This step is primarily used to identify dynamically changing areas within an image, such as those used to capture fast-moving objects in surveillance video. By analyzing the pixel changes between consecutive frames in an image sequence, the motion vector of each pixel is calculated and statistically analyzed to determine whether the pixel motion is consistent. If the motion vectors of most pixels within a region are consistent, it indicates that the movement in that region is uniform and is caused by camera shake or wind, rather than movement of the target object. By setting a threshold for the proportion of non-target areas within the consistent region, such as setting the non-target motion area to no more than 30% of the total image area, non-target dynamic areas can be effectively screened out. For example, non-target areas include dynamic backgrounds caused by wind-blown leaves. This technique can prevent natural motion from being mistakenly identified as significant events. The non-target area ratio calculated in this way provides important reference data for subsequent image analysis and event determination.

[0107] The disturbance elimination submodule collects the distribution of boundary noise pixels based on the proportion of non-target areas, and compares the grayscale difference between the foreground and background with the error amplitude using the formula:

[0108] ;

[0109] Calculate the disturbance rejection coefficient, filter out the noise area, extract the mask layer, and generate the background disturbance elimination result;

[0110] in, represents the disturbance rejection coefficient, Represents the maximum value of the foreground grayscale gradient, represents the background gray gradient mean, Represents the noise area offset error amplitude, Represents the error variation of the noise area, Represents the proportion of non-target areas, Representative The noise density value of the segment boundary area, Represents the number of segments of the noise area;

[0111] The operation is performed based on the non-target area ratio. The non-target area ratio ω is calculated by the ratio of the area of the consistent motion vector region in the image to the area of the entire image. For example, in a 1920×1080 pixel image, if the consistent non-target area occupies 345600 pixels, then ω = 345600 / (1920×1080) ≈ 0.166. The noise pixel distribution feature values of the image boundary area are collected. First, the image boundary is divided into z = 8 segments with 16×16 pixels as the unit;

[0112] Sample the noise density within each segment , noise is defined as a pixel whose intensity fluctuation exceeds ±15 gray levels within three frames. For example, in the first segment, if the number of fluctuating pixels is 204 and the total number of pixels in the area is 256, then , and the rest of the segments are , , , , , , , and then get ;

[0113] Then collect the boundary grayscale features and the maximum value of the foreground grayscale gradient The maximum value calculated by taking the first-order derivative of the boundary foreground pixel is set to Grayscale / pixel, background grayscale gradient mean χ is the average value of background pixel grayscale gradient in the region, Grayscale / pixel;

[0114] Analyze the offset error in the noise area, and the maximum offset error It is defined as the maximum distance that the center of the noise point deviates from its geometric center, and the dimension is pixel. Pixel, error variation is the standard deviation of noise pixel error, let Pixels;

[0115] Substitute into the calculation: ;

[0116] The disturbance rejection coefficient Indicates the significance of noise interference in the boundary area. By setting the disturbance rejection threshold to 5.5, judging that the current Q is higher than the threshold, the rejection operation is performed to extract and shield the noise mask in the boundary area to form a clean mask layer that can be used for subsequent image fusion processing. Finally, the background disturbance elimination result is established by and offset error 、 Combine the operations and introduce the non-target area ratio into the denominator , which enhances the ability to distinguish scenes with complex background disturbances. This result shows that there is considerable disturbance in the boundary area of the current image, and masking processing is required to ensure the accuracy of boundary recognition.

[0117] See also Figure 6 , the feedback association module includes:

[0118] The view offset extraction submodule extracts the updated camera pose sequence and the starting pose data after background disturbance elimination based on the background disturbance elimination results. It analyzes the Euler angle difference and displacement changes between consecutive frames, and normalizes them with the timestamp to obtain the camera view offset trend.

[0119] The camera pose sequence is extracted using a background disturbance elimination algorithm. Background disturbance elimination is achieved by comparing environmental changes between frames and using filtering or adaptive algorithms to remove irrelevant dynamic changes to obtain stable pose data. This data includes pose and position, which are composed of rotation and displacement in the coordinate system. Based on the extracted pose data, the pose Euler angle difference and displacement change between consecutive frames are calculated. The Euler angle difference is obtained by calculating the rotation angle change between each frame, while the displacement change is obtained by directly calculating the Euclidean distance between the camera positions. The difference is normalized using the timestamp. Normalization is to eliminate the influence of the time length between different frames and map all displacement and angle differences to a unified time scale. Assuming that the camera displacement change between two frames is 0.5 meters, the rotation angle difference is 15 degrees, and the time interval is 2 seconds, it can be normalized to a displacement of 0.25 meters per second and a rotation of 7.5 degrees per second, thereby obtaining the camera's perspective offset trend.

[0120] The path linkage comparison submodule analyzes the cosine value of the direction vector angle and the rate of change of the velocity amplitude at the corresponding time point based on the camera view offset trend and the spatial direction change sequence of adjacent posture segments in the target tracking path, and determines the response lag in the synchronization interval to obtain the path response synchronization characteristics;

[0121] Viewpoint data from adjacent time points is extracted from the camera viewpoint deviation trend. This data represents the spatial orientation of the camera relative to the target tracking path. The sequence of directional changes between adjacent pose segments in the target tracking path is then extracted. For each pair of adjacent pose segments, the cosine of the angle between their direction vectors is calculated. The direction vector is obtained by transforming each pose using the rotation matrix or Euler angles. The vector indicates the spatial orientation of the path. The cosine of the angle is obtained by taking the dot product of two vectors. If the two direction vectors are identical, the cosine is 1; if they are completely opposite, the cosine is -1. For example, if the direction vectors of two adjacent poses in the target path are (1, 0, 0) and (0, 1, 0), the cosine of the angle is 0 because they are perpendicular. The rate of change of the velocity amplitude is calculated, that is, the velocity increment at each time point is obtained by varying the velocity and time. The rate of change of the velocity data can be calculated by differencing the velocity data. The cosine of the angle and the rate of change of the velocity amplitude are used to analyze the response lag within the synchronization interval. This lag is manifested as the change in the cosine of the angle lags behind the change in velocity, thereby characterizing the synchronization of the path response.

[0122] The adjustment strategy generation submodule uses the path response synchronization characteristics and the camera speed change rate at the previous moment, sets the proportion parameters according to the response lag time difference and the angle offset, and performs weighted processing on each pose adjustment variable to generate the camera pose response matching degree;

[0123] The path response synchronization characteristics and the camera's velocity change rate at the preceding moment are obtained. This data is used to determine the camera's response to path changes. The response lag time difference indicates the lag time of camera adjustment, while the angle offset refers to the degree of deviation between the camera's orientation and the path's orientation. Based on this information, weighting parameters are set. For longer lag times, a higher weight can be assigned to the lag time difference, while for larger angle offsets, a higher weight can be assigned to the angle offset. Combining these two weights yields a weighted adjustment for each pose. For example, if the lag time difference is 0.5 seconds and the angle offset is 10 degrees, these two factors can be combined according to a predetermined weighting formula to determine the final adjustment. Assuming a weight of 0.6 for the lag time difference and 0.4 for the angle offset, the final adjustment for each pose is calculated by weighting these two weights, thus generating the camera pose response matching degree.

[0124] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An adaptive camera posture intelligent adjustment system, characterized in that: The system comprises: The dynamic recognition module obtains camera pose data, extracts video frames captured by the image acquisition component, extracts pixel brightness gradients and boundary contour continuity in the foreground area of the frame, identifies target position transformation data in the previous and next frames within the image stabilization structure, determines the motion trend of the area block and identifies the continuous motion path, and generates target tracking path information; The posture adjustment module extracts the current pitch and rotation angle values of the camera based on the target tracking path information, analyzes the posture deviation degree and the required angle adjustment amplitude in combination with the target movement direction and speed range, and generates a posture correction instruction set; The mechanism control module collects the current response state of the execution structure according to the posture correction instruction set, compares the target angle and displacement requirements in the instruction, identifies the drive error range, adjusts the controller output amplitude and direction, and obtains the mechanism execution adjustment value; The information filtering module calls the mechanism to execute the adjustment value, extracts the boundary clarity and motion consistency features of the foreground and background areas in the image, identifies the proportion of non-target dynamic areas, and removes the boundary noise areas to generate background disturbance elimination results.

2. The adaptive camera posture intelligent adjustment system according to claim 1, characterized in that: The target tracking path information includes the position change amplitude, trajectory curvature characteristics, and displacement direction angle; the posture correction instruction set includes the pitch angle adjustment amount, the rotation angle correction value, and the angle offset threshold; the mechanism execution adjustment value includes the controller output gain, the execution structure rotation angle, and the real-time response displacement; the background disturbance elimination result includes the foreground boundary clarity score, the background area consistency coefficient, and the dynamic interference ratio.

3. The adaptive camera posture intelligent adjustment system according to claim 1, characterized in that: The dynamic recognition module includes: The pose estimation submodule obtains the camera pose data, extracts the camera's 3D pose data and center pixel coordinates for each frame, analyzes the pose rotation angle difference and translation vector change rate in consecutive frames, determines whether the change exceeds the pose stability threshold, and generates a pose adjustment trigger signal; The foreground perception submodule extracts the brightness gradient and boundary contour continuity of the foreground in the current frame based on the posture adjustment trigger signal, and combines the spatial offset and direction change of the target edge points in the previous and next frames to determine the perceptual integrity trend of the foreground area under the posture change and obtain the image perception integrity parameter; The path linkage submodule dynamically corrects the current posture according to the image perception completeness parameter, adjusts the camera focus direction and rotation angle, calls the offset difference between the center point position of the corrected image and the original image, and obtains the posture adjustment execution path.

4. The adaptive camera posture intelligent adjustment system according to claim 3, characterized in that: The posture adjustment module includes: The angle extraction submodule extracts the pitch angle and rotation angle data in the image frame based on the target tracking path information, identifies the pitch angle and rotation angle by combining the pixel matrix change and the image edge position sequence, and obtains the angle change value set; The direction matching submodule calls the angle change value set, detects the direction consistency and speed change trend according to the target moving direction vector and the speed range, calculates the posture direction offset strength value, and obtains the direction offset adjustment parameter value by combining the direction matching result and the speed jump range; The instruction generation submodule calls the direction offset adjustment parameter value, divides the gear according to the offset amplitude and speed range, determines whether it is in the adjustment response range, maps the correction action parameters, and generates a posture correction instruction set.

5. The adaptive camera posture intelligent adjustment system according to claim 4, characterized in that: The mechanism control module includes: The state recognition submodule collects the current response state of the execution structure according to the posture correction instruction set, extracts the posture angle vector and displacement component, identifies the dimension deviation distance, and generates the posture displacement difference; The error evaluation submodule calls the posture displacement difference, determines the matching offset, extracts the dynamic change amplitude and response trend in the response state, compares the change rate with the response rate threshold, analyzes the driving accuracy difference interval, and obtains the driving error range value; The control adjustment submodule calls the driving error range value, combines the angle response sensitivity, displacement response hysteresis amplitude, and adjustment feedback cycle amplitude, corrects the adjustment amplitude and direction of the controller output, calculates the control output correction amplitude, corrects the controller signal, and obtains the mechanism execution adjustment value.

6. The adaptive camera posture intelligent adjustment system according to claim 5, characterized in that: The information filtering module includes: The clarity extraction submodule calls the mechanism to execute the adjustment value, extracts the grayscale gradient of the image boundary, screens the pixels that meet the clarity threshold, and obtains the boundary clarity change rate; The dynamic ratio judgment submodule calls the boundary clarity change rate, detects the consistency of the regional motion vector, and calculates the area ratio of the non-target area to obtain the non-target area ratio value; The disturbance elimination submodule collects the boundary noise pixel distribution according to the non-target area ratio, performs comparison based on the foreground and background grayscale difference and error amplitude, calculates the disturbance elimination coefficient, filters out the noise area, extracts the mask layer, and generates the background disturbance elimination result.

7. The adaptive camera posture intelligent adjustment system according to claim 1, characterized in that: The system also includes a feedback association module: The feedback association module evaluates the response synchronization between the camera view movement and the target path based on the background disturbance elimination result, combined with the updated target tracking path and the current posture execution state, and generates a camera posture response matching degree; The camera posture response matching degree includes target path tracking deviation, posture adjustment delay, and synchronization coordination index.

8. The adaptive camera posture intelligent adjustment system according to claim 7, characterized in that: The feedback association module includes: The view offset extraction submodule extracts the updated camera pose sequence and the starting pose data after background disturbance elimination based on the background disturbance elimination result, analyzes the pose Euler angle difference and displacement change between consecutive frames, and performs normalization processing based on the timestamp to obtain the camera view offset trend; The path linkage comparison submodule analyzes the cosine value of the direction vector angle and the velocity amplitude change rate at the corresponding time point based on the camera view offset trend and the spatial direction change sequence of adjacent posture segments in the target tracking path, and determines the response lag in the synchronization interval to obtain the path response synchronization characteristics; The adjustment strategy generation submodule calls the path response synchronization characteristics and the camera speed change rate at the previous moment, sets the proportion parameters according to the response lag time difference and the angle offset degree, performs weighted processing on each posture adjustment variable, and generates the camera posture response matching degree.

Images