Target tracking method based on fuzzy PD control

Abstract

The invention provides a target tracking method based on fuzzy PD control. The method comprises the following steps: S1, initializing a servo control system, a video acquisition system and a communication system; s2, performing real-time detection on a set target in the video according to a set control period; s3, obtaining the position of an offset center point of each frame of image according to target detection; s4, designing a fuzzy PD control system; s5, third closed-loop control; and S6, target loss control. According to the target tracking method based on fuzzy PD control, accurate and stable tracking of the moving target is achieved, the tracking precision and efficiency are improved, through application of the fuzzy PID controller, the system has good robustness and adaptability and can effectively work in the complex environment, and the target tracking efficiency is improved. The method can be widely applied to the fields of intelligent monitoring, unmanned aerial vehicle navigation, automatic driving and the like, and has relatively high commercialization and socialization values.

Description

Technical Field

[0001] The invention relates to the field of high-speed ball tracking, and in particular to a target tracking method based on fuzzy PD control. Background Art

[0002] In recent years, with the development of computer vision and control technology, video surveillance systems based on target tracking have been widely used in security, intelligent transportation and other fields. Video surveillance systems based on target tracking use image processing technology to analyze and understand video streams and, based on this, to link the servo control system that carries the video surveillance system, thereby making it intelligent. Video surveillance systems based on target tracking mainly complete the following functions:

[0003] Use computer vision technology to detect the target location;

[0004] The servo device is linked according to the detected target position to realize the tracking function.

[0005] There are many security monitoring systems now, and high-speed ball tracking system is one of them. High-speed ball is a highly integrated product, integrating pan-tilt system, communication system, and camera system. The pan-tilt system refers to the rotating part driven by the motor, the communication system refers to the control of the motor and the processing of images and signals, and the camera system refers to the all-in-one machine core used. The main control core CPU and the power supply part play a horizontal connection between the major systems. The power supply part is powered by diodes, transistors and other micro-currents in many places, and the core CPU is the basis for the normal operation of all functions.

[0006] The current high-speed ball tracking system has the following problems:

[0007] Jitter and blur: Under fast movement or vibration, some surveillance dome cameras may produce jittery or blurred images, affecting the image quality and viewing experience;

[0008] Low tracking accuracy: The tracking accuracy of some surveillance cameras may not be high enough in fast-moving or complex environments, resulting in unstable or inaccurate tracking of the target object, and causing the video screen to be stuck and incoherent during tracking control;

[0009] Lack of adaptability: Some surveillance cameras are not adaptable enough to different scenarios and environments, and cannot flexibly respond to surveillance needs in various complex situations;

[0010] The control method is not sensitive: using step control, the response curve is not fast enough.

[0011] Therefore, it is necessary to provide a target tracking method based on fuzzy PD control to solve the above technical problems. Summary of the invention

[0012] The invention provides a target tracking method based on fuzzy PD control, which solves the problems of jitter and blur, low tracking accuracy, insufficient adaptability and insensitive control mode of the current high-speed ball tracking system.

[0013] In order to solve the above technical problems, the present invention provides a target tracking method based on fuzzy PD control, comprising the following steps:

[0014] S1. Initialize the servo control system, video acquisition system, and communication system;

[0015] S2, performing real-time detection of a set target in the video according to a set control cycle;

[0016] S3, obtaining the offset center point position of each frame image according to target detection;

[0017] S4. Design fuzzy PD control system;

[0018] S5, three closed-loop control;

[0019] S6. Target lost control.

[0020] Preferably, S4 comprises the following steps:

[0021] S41. Assuming that the input signal is the error e(t) and the rate of change of the error (de(t) / dt), and the output signal is the control quantity u(t), the formula example of the fuzzy PD controller can be expressed as:

[0022] u(t)=Kpe(t)+Kd*(de(t) / dt);

[0023] S42. Assume that the error e(t) and the rate of change of the error (de(t) / dt) are fuzzyized into three membership functions: "negative", "zero", and "positive", respectively. The control quantity u(t) is also fuzzyized into three membership functions: "decrease", "unchanged", and "increase".

[0024] Preferably, in S4, Kp is a proportional gain and Kd is a differential gain. In a fuzzy PD controller, these two gains are usually fuzzified, that is, their values ​​are fuzzy rather than definite constants. The fuzzification process converts the gains into fuzzy sets according to the system requirements and performance indicators, so that the controller can process fuzzy input information.

[0025] Preferably, the following fuzzy rules are formulated in S42:

[0026] S421. If e(t) is negative and (de(t) / dt) is negative, then u(t) is increasing;

[0027] S422. If e(t) is zero and (de(t) / dt) is zero, then u(t) is unchanged;

[0028] S423. If e(t) is positive and (de(t) / dt) is positive, then u(t) decreases.

[0029] According to the fuzzy set values ​​of the current error e(t) and the error change rate (de(t) / dt), fuzzy rules are used for reasoning to obtain the fuzzy set value of the output u(t). Then, the fuzzy defuzzification method is used to convert the fuzzy output into a specific control quantity.

[0030] Preferably, S5 includes:

[0031] S51, position closed loop control;

[0032] S52, speed closed loop control;

[0033] S53, current closed loop control;

[0034] S54, parameter adjustment feedback and adjustment.

[0035] Preferably, in S51, the image coordinate value is first detected according to the target, and the offset angle A1 of the image relative to the camera center point is inferred according to the image coordinate value. Assuming that the target position is the expected position, the camera needs to be moved by the angle A1, and the speed output of the system is adjusted according to the error between the expected position and the current position so that the position of the system reaches the expected value.

[0036] Preferably, in S52, a speed control link is further added on the basis of position control. The speed control link adjusts the output of the system according to the error between the desired speed and the current speed so that the speed of the system reaches the desired value. The speed control is achieved by differentiating the position error.

[0037] Preferably, the current of the system is closed-loop controlled in S53 to ensure that the current output of the system meets the requirements. The current control link adjusts the output of the system according to the error between the expected current and the current so that the current of the system reaches the expected value. The current control uses a current sensor to measure the current signal and provide feedback.

[0038] Preferably, the PD parameters are adjusted in S54 according to the experience of the system and the intuition of the operator. According to the observation and experience of the dynamic response of the system, the proportional gain Kp is manually adjusted, and after the proportional gain is adjusted to the oscillation point, the proportional parameter is reduced to 0.8 of the current parameter, and then the differential parameter Kd is adjusted until a satisfactory control effect is achieved.

[0039] Preferably, in said S6, the offset angle of the target relative to the center point of the ball camera video is determined according to the target offset center point position. If the target is lost, the last detected target center point position relative to the camera is used as the final control holding position.

[0040] Compared with the related art, the target tracking method based on fuzzy PD control provided by the present invention has the following beneficial effects:

[0041] The present invention provides a target tracking method based on fuzzy PD control, which realizes accurate and stable tracking of a moving target and improves tracking accuracy and efficiency.

[0042] Through the application of fuzzy PID controller, the system has good robustness and adaptability and can work effectively in complex environments.

[0043] It can be widely used in intelligent monitoring, drone navigation, autonomous driving and other fields, and has high commercial and social value.

[0044] Wide range of applications: Position-speed dual closed-loop control is suitable for many systems that require precise control of position and speed, such as robot control, motor control, etc.

[0045] Position-speed dual closed-loop control is widely used in many fields and can effectively improve the control performance and stability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A schematic structural diagram of a preferred embodiment of a target tracking method based on fuzzy PD control provided by the present invention;

[0047] Figure 2 It is the flow chart of position closed-loop control;

[0048] Figure 3 It is the circuit closed-loop control flow chart;

[0049] Figure 4 It is the flow chart of speed closed loop control;

[0050] Figure 5 This is the system framework diagram. DETAILED DESCRIPTION

[0051] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0052] Please refer to Figure 1 , Figure 2 , Figure 3 , Figure 4 and Figure 5 ,in, Figure 1A schematic structural diagram of a preferred embodiment of a target tracking method based on fuzzy PD control provided by the present invention; Figure 2 It is the flow chart of position closed-loop control; Figure 3 It is the circuit closed-loop control flow chart; Figure 4 It is the flow chart of speed closed loop control; Figure 5 A target tracking method based on fuzzy PD control includes the following steps:

[0053] S1. Initialize the servo control system, video acquisition system, and communication system;

[0054] S2, performing real-time detection of a set target in the video according to a set control cycle;

[0055] S3, obtaining the offset center point position of each frame image according to target detection;

[0056] S4. Design fuzzy PD control system;

[0057] S5, three closed-loop control;

[0058] S6. Target lost control.

[0059] The S4 comprises the following steps:

[0060] S41. Assuming that the input signal is the error e(t) and the rate of change of the error (de(t) / dt), and the output signal is the control quantity u(t), the formula example of the fuzzy PD controller can be expressed as:

[0061] u(t)=Kpe(t)+Kd*(de(t) / dt);

[0062] S42. Assume that the error e(t) and the rate of change of the error (de(t) / dt) are fuzzyized into three membership functions: "negative", "zero", and "positive", respectively. The control quantity u(t) is also fuzzyized into three membership functions: "decrease", "unchanged", and "increase".

[0063] In S4, Kp is the proportional gain and Kd is the differential gain. In a fuzzy PD controller, these two gains are usually fuzzified, that is, their values ​​are fuzzy rather than definite constants. The fuzzification process converts the gains into fuzzy sets based on the system requirements and performance indicators, so that the controller can process fuzzy input information.

[0064] The following fuzzy rules are formulated in S42:

[0065] S421. If e(t) is negative and (de(t) / dt) is negative, then u(t) is increasing;

[0066] S422. If e(t) is zero and (de(t) / dt) is zero, then u(t) is unchanged;

[0067] S423. If e(t) is positive and (de(t) / dt) is positive, then u(t) decreases.

[0068] According to the fuzzy set values ​​of the current error e(t) and the error change rate (de(t) / dt), fuzzy rules are used for reasoning to obtain the fuzzy set value of the output u(t). Then, the fuzzy defuzzification method is used to convert the fuzzy output into a specific control quantity.

[0069] The S5 includes:

[0070] S51, position closed loop control;

[0071] S52, speed closed loop control;

[0072] S53, current closed loop control;

[0073] S54, parameter adjustment feedback and adjustment.

[0074] In the S51, the image coordinate value is first detected according to the target, and the offset angle A1 of the image relative to the camera center point is inferred based on the image coordinate value. Assuming that the target position is the expected position, the camera needs to be moved by the angle A1, and the speed output of the system is adjusted according to the error between the expected position and the current position so that the position of the system reaches the expected value.

[0075] In S52, a speed control link is further added on the basis of position control. The speed control link adjusts the output of the system according to the error between the desired speed and the current speed so that the speed of the system reaches the desired value. The speed control is achieved by differentiating the position error.

[0076] In S53, the current of the system is closed-loop controlled to ensure that the current output of the system meets the requirements. The current control link adjusts the output of the system according to the error between the expected current and the current so that the current of the system reaches the expected value. The current control uses a current sensor to measure the current signal and provide feedback.

[0077] In S54, the PD parameters are adjusted according to the system experience and the operator's intuition. According to the observation and experience of the system dynamic response, the proportional gain Kp is manually adjusted. After the proportional gain is adjusted to the oscillation point, the proportional parameter is reduced to 0.8 of the current parameter and the differential parameter Kd is adjusted until a satisfactory control effect is achieved.

[0078] In the step S6, the offset angle of the target relative to the center point of the ball camera video is determined according to the offset center point position of the target. If the target is lost, the last detected target relative to the center point of the camera is used as the final control holding position.

[0079] By combining the advantages of fuzzy logic and classic PI control, accurate and stable tracking of fast-moving targets in videos is achieved through three-loop servo control. The fuzzy PD (fuzzy proportional-differential) controller is a control method that combines fuzzy logic with PD control. It uses fuzzy logic to process complex and fuzzy input information, and combines the fast response and stability of the PD controller to achieve control of the system. The design of the fuzzy PD controller requires the formulation of fuzzy rules and membership functions, and the determination of the fuzzy set value of the gain through fuzzy reasoning. The design of specific fuzzy rules and membership functions will vary according to the characteristics of the system and application requirements, and usually requires a combination of expert experience and system modeling to determine.

[0080] Position, speed and current closed-loop control is a common control method used to control the movement of a motor system or a mechanical system. In this control method, the controller controls the movement of the system through closed-loop feedback based on the position, speed and current signals of the system to achieve the desired position, speed and current output.

[0081] More stable control: By controlling position and velocity simultaneously, the system can be regulated more stably, reducing position and velocity errors.

[0082] Faster response: The speed loop controller can speed up the system's response speed, allowing the system to reach the desired position and speed more quickly.

[0083] More precise control: By jointly controlling position and speed, more precise control of the system can be achieved, improving system performance and accuracy.

[0084] Compared with the related art, the target tracking method based on fuzzy PD control provided by the present invention has the following beneficial effects:

[0085] The present invention provides a target tracking method based on fuzzy PD control, which realizes accurate and stable tracking of a moving target and improves tracking accuracy and efficiency.

[0086] Through the application of fuzzy PID controller, the system has good robustness and adaptability and can work effectively in complex environments.

[0087] It can be widely used in intelligent monitoring, drone navigation, autonomous driving and other fields, and has high commercial and social value.

[0088] Wide range of applications: Position-speed dual closed-loop control is suitable for many systems that require precise control of position and speed, such as robot control, motor control, etc.

[0089] Position-speed dual closed-loop control is widely used in many fields and can effectively improve the control performance and stability of the system.

[0090] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A target tracking method based on fuzzy PD control, characterized in that: The following steps are involved: S1. Initialize the servo control system, video acquisition system, and communication system; S2, performing real-time detection of a set target in the video according to a set control cycle; S3, obtaining the offset center point position of each frame image according to target detection; S4. Design fuzzy PD control system; S5, three closed-loop control; S6. Target lost control.

2. The target tracking method based on fuzzy PD control according to claim 1 is characterized in that: The S4 comprises the following steps: S41. Assuming that the input signal is the error e(t) and the rate of change of the error (de(t) / dt), and the output signal is the control quantity u(t), the formula example of the fuzzy PD controller can be expressed as: u(t)=Kpe(t)+Kd*(de(t) / dt); S42. Assume that the error e(t) and the rate of change of the error (de(t) / dt) are fuzzyized into three membership functions: "negative", "zero", and "positive", and the control quantity u(t) is also fuzzyized into three membership functions: "decrease", "unchanged", and "increase".

3. The target tracking method based on fuzzy PD control according to claim 1 is characterized in that: In S4, Kp is the proportional gain, and Kd is the differential gain. In a fuzzy PD controller, these two gains are usually fuzzified, that is, their values ​​are fuzzy rather than definite constants. The fuzzification process converts the gains into fuzzy sets according to the system requirements and performance indicators, enabling the controller to process fuzzy input information.

4. The target tracking method based on fuzzy PD control according to claim 2 is characterized in that: The following fuzzy rules are formulated in S42: S421. If e(t) is negative and (de(t) / dt) is negative, then u(t) is increasing; S422. If e(t) is zero and (de(t) / dt) is zero, then u(t) is unchanged; S423. If e(t) is positive and (de(t) / dt) is positive, then u(t) decreases. According to the fuzzy set values ​​of the current error e(t) and the error change rate (de(t) / dt), fuzzy rules are used for reasoning to obtain the fuzzy set value of the output u(t). Then, the fuzzy defuzzification method is used to convert the fuzzy output into a specific control quantity.

5. The target tracking method based on fuzzy PD control according to claim 1, characterized in that: The S5 includes: S51, position closed loop control; S52, speed closed loop control; S53, current closed loop control; S54, parameter adjustment feedback and adjustment.

6. The target tracking method based on fuzzy PD control according to claim 5, characterized in that: In the S51, the image coordinate value is first detected according to the target, and the offset angle A1 of the image relative to the camera center point is inferred based on the image coordinate value. Assuming that the target position is the expected position, the camera needs to be moved by the angle A1, and the speed output of the system is adjusted according to the error between the expected position and the current position so that the position of the system reaches the expected value.

7. The target tracking method based on fuzzy PD control according to claim 5, characterized in that: In S52, a speed control link is further added on the basis of position control. The speed control link adjusts the output of the system according to the error between the desired speed and the current speed so that the speed of the system reaches the desired value. The speed control is achieved by differentiating the position error.

8. The target tracking method based on fuzzy PD control according to claim 5, characterized in that: In S53, the current of the system is closed-loop controlled to ensure that the current output of the system meets the requirements. The current control link adjusts the output of the system according to the error between the expected current and the current so that the current of the system reaches the expected value. The current control uses a current sensor to measure the current signal and provide feedback.

9. The target tracking method based on fuzzy PD control according to claim 5, characterized in that: In S54, the PD parameters are adjusted according to the system experience and the operator's intuition. According to the observation and experience of the dynamic response of the system, the proportional gain Kp is manually adjusted. After the proportional gain is adjusted to the oscillation point, the proportional parameter is reduced to 0.8 of the current parameter and then the differential parameter Kd is adjusted until a satisfactory control effect is achieved.

10. The target tracking method based on fuzzy PD control according to claim 1, characterized in that: In the step S6, the offset angle of the target relative to the center point of the ball camera video is determined according to the offset center point position of the target. If the target is lost, the last detected target relative to the center point of the camera is used as the final control holding position.

Images