Method and system for automatic compensation of dynamic strike delay of a laser weeding robot

Abstract

The application discloses a dynamic striking delay automatic calculation compensation method and system of a laser weeding robot, and belongs to the technical field of agricultural intelligent equipment. The method is realized through the following steps: in the static calibration stage, the laser galvanometer, the camera and the cooperative parameters of the two are corrected to eliminate the laser scanning nonlinear error, the camera imaging distortion error and the pose and mapping error between the galvanometer coordinate system and the camera coordinate system; in the dynamic test stage, the robot travels at a constant speed, the laser strikes the preset marked wood piece, the position deviation between the actual striking point and the theoretical target point is obtained through the image recognition technology; the delay time of the system is deduced based on the deviation and the speed of the robot, and the compensation parameters are iteratively updated; finally, the compensation parameters are solidified to the control system after the iterative optimization striking deviation is less than the set precision; the application is suitable for the laser weeding robot and other dynamic visual striking systems.

Description

Technical Field

[0001] This invention relates to the field of intelligent agricultural equipment, and in particular to a method and system for automatic time delay compensation in dynamic target attack by a laser weeding robot, which is used to improve the attack accuracy of the robot in mobile operation. Background Technology

[0002] Existing laser weeding robots are often affected by system latency when identifying and striking targets (such as weeds) with lasers during their movement.

[0003] The program executes laser emission commands based on the timestamp at the time of execution and the corresponding spatial position of the target. There is a time difference Δt between the laser emission control command being issued and the laser actually being executed. Therefore, even if the strike is based on the position (x, y, z) at time t, the robot is still moving during this time period. When the laser is executed, the laser strike position will be offset relative to the target: Δd = Δt * v. Theoretically, the strike point will lag behind the target point. Therefore, the factor of Δt needs to be taken into account in the execution command part and predicted in advance so that the laser strikes the target position exactly when it is executed.

[0004] This delay is caused by the superposition of multiple factors, including image acquisition delay, control signal transmission delay, laser execution delay, and vehicle motion deviation.

[0005] Currently, the system typically relies on static calibration, which cannot accurately compensate for such delays in dynamic operations, resulting in significant strike deviations. Summary of the Invention

[0006] This invention overcomes the shortcomings of existing technologies and provides a dynamic self-calibration delay compensation method and system for laser weeding robots. Through continuous dynamic identification and impact testing, the laser execution delay time compensation value is automatically calculated, enabling the robot to achieve high-precision impact while moving.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: an automatic calculation and compensation method for dynamic impact delay of laser weeding robots, implemented in the following manner.

[0008] S1. Static calibration stage;

[0009] The inherent parameters of the robot system are calibrated in a static state, including the nonlinear error correction parameters of the laser galvanometer, the distortion error correction parameters of the camera, and the pose error correction parameters of the galvanometer-camera mounting. This enables the theoretical target position in the galvanometer coordinate system to be accurately located by the target pixel position, thus achieving precise laser strike.

[0010] S2. Dynamic testing phase;

[0011] The robot moves at a constant speed and dynamically strikes the target wood chips that enter the test area. It uses a camera to detect the corners and black spots of the target, calculates the strike error, and then calculates the time delay compensation parameters to update the system parameters.

[0012] S3. Iterative optimization phase;

[0013] Repeat the above process until the hit deviation of multiple consecutive tests is less than the set threshold, at which point the delay compensation is considered to have converged.

[0014] S4. Parameter solidification stage;

[0015] The final compensation value is written into the control system parameter area for automatic application later, and the dynamic calibration process can be restarted through the one-click calibration function.

[0016] Furthermore, the static calibration method in S1 can be implemented through the following steps:

[0017] S11. Place the wooden chip with the center mark;

[0018] S12. Control the laser to strike the marked point and record the actual and theoretical laser strike positions;

[0019] S13. Calculate the nonlinear error of the galvanometer, the distortion error of the camera, and the galvanometer-camera mounting pose error by recording data;

[0020] S14. Update system parameters to achieve the correspondence between the target pixel position and the theoretical target position in the galvanometer coordinate system, thereby achieving precise laser strike.

[0021] Furthermore, the dynamic testing method in S2 is as follows:

[0022] S21. The robot moves at a constant speed v;

[0023] S22. Identify and track the marked wood chips, and initiate the strike procedure;

[0024] S23. The image recognition module detects the position of the wood chip, the corner of the wood chip, and the center of the black spot, and calculates the pixel position P of the actual impact black spot. actual Pixel position P relative to the theoretical target center theory ;

[0025] S24. Calculate the spatial deviation Δd (unit: cm) between the actual impact point and the theoretical target center.

[0026] S25. Calculate the delay Δt = Δd / v based on the robot's travel speed v;

[0027] S26. Update the system time compensation parameters, where the initial compensation parameter Told Set to 0, and then iteratively update based on the time delay Δt calculated for each dynamic strike, T new =T old +Δt.

[0028] Furthermore, a fixed-size square wooden chip is used as a dynamic calibration target carrier. Target marks with fixed positions are preset on the wooden chip. During the movement, the program detects the target and then strikes the target mark with a laser. By detecting the actual position of the black spot formed on the wooden chip after the laser strike, the strike deviation from the theoretical center of the wooden chip is calculated in the local coordinate system of the wooden chip. Combined with the robot's movement speed, the system time delay is deduced, thus realizing the self-learning and convergence of the delay compensation parameters.

[0029] Furthermore, the calculation and compensation process is as follows:

[0030] The first step is to detect the pixels at the four corners of the wood chip;

[0031] The coordinates of the four corner pixels are represented as follows:

[0032] C1(u1,v1),C2(u2,v2),C3(u3,v3),C4(u4,v4);

[0033] The second step is to detect the central pixel of the dark spot;

[0034] The coordinates of the center pixel of the black spot are represented as follows:

[0035] P hit (u h ,v h );

[0036] The third step is to map pixels to the actual coordinates of the wood chips.

[0037] Length and width of the wood chip: Let the length of the wood chip be L. x Width is L y (Unit: cm);

[0038] Normalization coefficient:

[0039] ;

[0040] The position of the black spot relative to the center of the wood chip: First, define the pixel coordinates (u) of the center of the wood chip. c ,v c );

[0041] ;

[0042] Then calculate the position of the dark spot relative to the center (X). hit ,Y hit );

[0043] ;

[0044] The fourth step is to calculate the actual spatial deviation;

[0045] The formula for calculating the actual spatial deviation Δd is as follows (unit: cm):

[0046] ;

[0047] Fifth step, obtain the time Δt that needs to be compensated;

[0048] Δt = Δd / v;

[0049] In the formula, v (unit: cm / s) is the robot's travel speed.

[0050] Furthermore, the Δd calculated using the above method is the actual positional deviation between the target point and the impact point, and is independent of the orientation angle of the wood chip placement. The magnitude of Δd is only related to the velocity and the compensation time. The more suitable the compensation time and the lower the velocity, the smaller the value. When the compensation time is accurate, only the measurement error remains. Similarly, when the impact occurs under static conditions, only the measurement error remains.

[0051] In motion, the black impact spot and the center point of the crosshair in the center of the wood chip deviate only in the direction of motion, and the displacement deviation is Δd.

[0052] The present invention discloses a system for automatically calculating and compensating for dynamic impact delay of a laser weeding robot, comprising the following cooperating modules:

[0053] The visual perception and imaging module includes an industrial camera and a supplementary lighting unit. The industrial camera is a high-frame-rate, high-pixel industrial camera used to acquire continuous image data of the work area. The supplementary lighting unit uses a high-brightness light source that is synchronously triggered with the camera to suppress ambient light changes and reflective interference, thereby improving the imaging stability of the target wood chips and their marker points. The visual perception and imaging module is used to identify, track, and strike target wood chips in dynamic scenes, and to identify and extract sub-pixel-level feature points for strike positioning and black spots left after the strike.

[0054] The motion sensing module, including an IMU (Inertial Measurement Unit) and a wheel speed encoder, is used to collect the pose information of the laser weeding robot in real time during its movement, output the robot's motion velocity parameters in the world coordinate system, and provide motion prior data for dynamic delay calculation.

[0055] The laser galvanometer strike execution module includes a laser, a dual-axis high-speed galvanometer, and a laser control board. The dual-axis high-speed galvanometer is used to achieve high-speed deflection and positioning of the laser spot in the target plane. The laser is used to output a high-energy laser beam for strike. The laser control board integrates a high-speed control unit to achieve precise timing control of the laser pulse and galvanometer scanning control, supporting microsecond-level response.

[0056] The central computing and control module, which is an industrial control computer or embedded PC platform, is used to run visual reasoning, motion prediction, and delay compensation control programs. Specifically, the central computing and control module deploys target recognition and corner / spot detection algorithms to perform real-time reasoning on image data acquired by the camera; it calculates the total system delay from image acquisition and data processing to the actual laser strike based on the detected impact error results and the velocity information output by the motion perception module; and it generates corresponding time compensation parameters based on the total delay.

[0057] The advantages of this invention compared to existing technologies are as follows: Through a dynamic self-calibration mechanism, the system can identify and compensate for delay errors caused by multiple stages such as image acquisition, signal transmission, and laser execution in real time, controlling the impact deviation to sub-millimeter or even pixel-level, significantly outperforming traditional methods relying on static calibration. This invention requires no manual intervention, automatically completing the identification, calculation, and updating of delay parameters during robot movement, forming a closed-loop control of "detection-impact-evaluation-compensation," adapting to different operating speeds and ground conditions, and improving the system's intelligence and practicality. Compared to traditional systems requiring periodic recalibration by professionals, this invention, through computer vision algorithms and a one-click calibration function, significantly reduces the frequency of manual intervention and maintenance costs, improving the long-term operational stability of the equipment. This invention uses fixed-size marked wooden chips as calibration carriers, combined with image recognition and coordinate mapping algorithms, ensuring that the calibration process is unaffected by environmental factors such as the placement angle of the wooden chips and changes in lighting, guaranteeing the system's reliability in complex farmland environments. This invention can establish delay compensation curves at different speeds, achieving speed-adaptive compensation; it also has data recording and report generation functions, facilitating subsequent performance analysis and troubleshooting. Attached Figure Description

[0058] The present invention will now be further described with reference to the accompanying drawings.

[0059] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0060] The present invention will be further described below with reference to specific embodiments.

[0061] like Figure 1As shown, the automatic calculation and compensation method for dynamic impact delay of the laser weeding robot of the present invention is implemented in the following manner: Figure 1 The flowchart of the method of the present invention mainly includes four stages: static calibration, dynamic testing, iterative optimization, and parameter solidification.

[0062] S1. Static calibration stage;

[0063] The inherent parameters of the robot system are calibrated in a static state, including the nonlinear error correction parameters of the laser galvanometer, the distortion error correction parameters of the camera, and the pose error correction parameters of the galvanometer-camera mounting. This enables the theoretical target position in the galvanometer coordinate system to be accurately located by the target pixel position, thus achieving precise laser strike.

[0064] S2. Dynamic testing phase;

[0065] The robot moves at a constant speed and dynamically strikes the target wood chip through multiple test areas. It uses a camera to detect the corners and black spots of the target, calculates the strike error, and then calculates the time delay compensation parameters to update the system parameters.

[0066] S3. Iterative optimization phase;

[0067] Repeat the above process until the hit deviation of multiple consecutive tests is less than the set threshold, at which point the delay compensation is considered to have converged.

[0068] S4. Parameter solidification stage;

[0069] The final compensation value is written into the control system parameter area for automatic application later, and the dynamic calibration process can be restarted through the one-click calibration function.

[0070] The static calibration method in S1 described in this invention can be implemented through the following steps:

[0071] S11. Place the wooden chip with the center mark;

[0072] S12. Control the laser to strike the marked point and record the actual and theoretical laser strike positions;

[0073] S13. Calculate the nonlinear error of the galvanometer, the distortion error of the camera, and the galvanometer-camera mounting pose error by recording data;

[0074] S14. Update system parameters to achieve the correspondence between the target pixel position and the theoretical target position in the galvanometer coordinate system, thereby achieving precise laser strike. The dynamic testing method in S2 of this invention is as follows:

[0075] S21. The robot moves at a constant speed v;

[0076] S22. Identify and track the marked wood chips, and initiate the strike procedure;

[0077] S23. The image recognition module detects the position of the wood chip, the corner of the wood chip, and the center of the black spot, and calculates the pixel position P of the actual impact black spot. actual Pixel position P relative to the theoretical target center theory ;

[0078] S24. Calculate the spatial deviation Δd (unit: cm) between the actual impact point and the theoretical target center.

[0079] S25. Calculate the delay Δt = Δd / v based on the robot's travel speed v;

[0080] S26. Update the system time compensation parameters, where the initial compensation parameter T old Set to 0, and then iteratively update based on the time delay Δt calculated for each dynamic strike, T new =T old +Δt.

[0081] In this invention, a square wooden chip of fixed size is used as a dynamic calibration target carrier. A target mark at a fixed position is preset on the wooden chip. During the movement, the program detects the target and then strikes the target mark with a laser. By detecting the actual position of the black spot formed on the wooden chip after the laser strike, the strike deviation from the theoretical center of the wooden chip is calculated in the local coordinate system of the wooden chip. Combined with the robot's movement speed, the system time delay is deduced, thus realizing the self-learning and convergence of the delay compensation parameters.

[0082] The calculation and compensation process described in this invention is as follows:

[0083] The first step is to detect the pixels at the four corners of the wood chip;

[0084] The coordinates of the four corner pixels are represented as follows:

[0085] C1(u1,v1),C2(u2,v2),C3(u3,v3),C4(u4,v4);

[0086] The second step is to detect the central pixel of the dark spot;

[0087] The coordinates of the center pixel of the black spot are represented as follows:

[0088] P hit (u h ,v h );

[0089] The third step is to map pixels to the actual coordinates of the wood chips.

[0090] Length and width of the wood chip: Let the length of the wood chip be L. x Width is Ly (Unit: cm);

[0091] Normalization coefficient:

[0092] ;

[0093] The position of the black spot relative to the center of the wood chip: First, define the pixel coordinates (u) of the center of the wood chip. c ,v c );

[0094] ;

[0095] Then calculate the position of the dark spot relative to the center (X). hit ,Y hit );

[0096] ;

[0097] The fourth step is to calculate the actual spatial deviation;

[0098] The formula for calculating the actual spatial deviation Δd is as follows (unit: cm):

[0099] ;

[0100] Fifth step, obtain the time Δt that needs to be compensated;

[0101] Δt = Δd / v;

[0102] In the formula, v (unit: cm / s) is the robot's travel speed.

[0103] The Δd calculated by the above method is the actual positional deviation between the target point and the impact point, and is independent of the orientation angle of the wood chip. The magnitude of Δd is only related to the speed and the compensation time. The more suitable the compensation time and the lower the speed, the smaller the value. When the compensation time is accurate, only the measurement error remains. Similarly, when the impact is performed statically, only the measurement error remains.

[0104] In motion, the black impact spot and the center point of the crosshair in the center of the wood chip deviate only in the direction of motion, and the displacement deviation is Δd.

[0105] The present invention discloses a system for automatically calculating and compensating for dynamic impact delay of a laser weeding robot, comprising the following cooperating modules:

[0106] The visual perception and imaging module includes an industrial camera and a supplementary lighting unit. The industrial camera is a high-frame-rate, high-pixel industrial camera used to acquire continuous image data of the work area. The supplementary lighting unit uses a high-brightness light source that is synchronously triggered with the camera to suppress ambient light changes and reflective interference, thereby improving the imaging stability of the target wood chips and their marker points. The visual perception and imaging module is used to identify, track, and strike target wood chips in dynamic scenes, and to identify and extract sub-pixel-level feature points for strike positioning and black spots left after the strike.

[0107] The motion sensing module, including an IMU (Inertial Measurement Unit) and a wheel speed encoder, is used to collect the pose information of the laser weeding robot in real time during its movement, output the robot's motion velocity parameters in the world coordinate system, and provide motion prior data for dynamic delay calculation.

[0108] The laser galvanometer strike execution module includes a laser, a dual-axis high-speed galvanometer, and a laser control board. The dual-axis high-speed galvanometer is used to achieve high-speed deflection and positioning of the laser spot in the target plane. The laser is used to output a high-energy laser beam for strike. The laser control board integrates a high-speed control unit to achieve precise timing control of the laser pulse and galvanometer scanning control, supporting microsecond-level response.

[0109] The central computing and control module, which is an industrial control computer or embedded PC platform, is used to run visual reasoning, motion prediction, and delay compensation control programs. Specifically, the central computing and control module deploys target recognition and corner / spot detection algorithms to perform real-time reasoning on image data acquired by the camera; it calculates the total system delay from image acquisition and data processing to the actual laser strike based on the detected impact error results and the velocity information output by the motion perception module; and it generates corresponding time compensation parameters based on the total delay.

[0110] This invention derives the system delay Δt by establishing a mapping relationship between the theoretical impact point and the actual impact point.

[0111] The mathematical expression is Δt = Δd / v, where Δd is the spatial deviation and v is the vehicle speed.

[0112] The workflow of this invention is as follows:

[0113] (I) System Initialization

[0114] Start the laser weeding robot and initialize the configuration of each functional module, including the industrial camera, laser galvanometer, motion control unit and central computing and control module;

[0115] Complete the loading of industrial camera imaging parameters and laser galvanometer control interface;

[0116] The motion sensing module collects the robot's current speed information and loads the real-time speed parameters into the central computing and control module, which serves as the prior motion basis for subsequent dynamic strike deviation calculation and time delay inference.

[0117] (II) Static Calibration Stage

[0118] Place a standard square wooden piece (e.g., 3cm x 3cm) at a stationary position in front of the robot, with a cross-shaped marker in the center of the wooden piece;

[0119] The laser is controlled to strike the marked points on the wood chip multiple times, and the actual landing point of each strike is recorded.

[0120] The image processing unit calculates the impact deviation between the actual impact point and the theoretical center, and then corrects the following parameters accordingly:

[0121] The nonlinear error of the laser galvanometer itself;

[0122] Camera lens distortion parameters;

[0123] The transformation relationship between the galvanometer coordinate system and the camera coordinate system;

[0124] After calibration is completed, the system saves the static calibration parameters as a benchmark for subsequent dynamic tests.

[0125] (III) Dynamic Testing Phase

[0126] The robot travels along a predetermined path at a constant speed v (e.g., 0.2 m / s);

[0127] The image acquisition unit captures images of the wood chips in real time and identifies their four corner points and center marker point;

[0128] Upon target detection, the laser strike procedure is triggered.

[0129] After the attack was completed, images of the wood chips were collected again to identify the center of the black spot formed by the laser on the wood chips;

[0130] The deviation Δd between the center of the black spot and the center of the wood chip in actual space is calculated using a coordinate mapping algorithm.

[0131] Calculate the system delay time using the formula Δt=Δd / v;

[0132] Update the calculated Δt to the system delay compensation parameter: T new =T old +Δt.

[0133] (iv) Iterative optimization stage

[0134] Repeat the dynamic testing steps, and immediately perform the next strike test after each update of the compensation parameters;

[0135] When the impact deviation of N consecutive tests (e.g., 5 tests) is less than the set threshold (e.g., 1 pixel or 0.5 mm), the system is considered to have converged.

[0136] Supports proportionally decreasing update strategy: T new =T old +k·Δt (0<k<1) to smooth the parameter adjustment process and avoid overshoot.

[0137] (v) Parameter fixation and restart mechanism

[0138] The final convergence delay compensation parameters are written into the non-volatile memory of the control system.

[0139] The system automatically calls this parameter in subsequent operations for real-time delay compensation;

[0140] It provides a "one-click calibration" function, which allows users to restart the complete calibration process when the equipment is under maintenance or the operating conditions change;

[0141] The system automatically generates calibration reports, records test data and final parameters at each stage, and supports export via USB or upload to the cloud.

[0142] (vi) Optional extended functions

[0143] It supports multi-speed gear calibration, establishes a speed-delay compensation lookup table, and realizes adaptive speed regulation compensation;

[0144] It can receive remote calibration commands via wireless network to achieve cloud-based collaborative calibration and fault diagnosis.

[0145] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A method for automatically calculating and compensating for dynamic impact delay of a laser weeding robot, characterized in that, Implement as follows S1. Static calibration stage; The inherent parameters of the robot system are calibrated in a static state, and the nonlinear error of the laser galvanometer, the distortion error of the camera, and the pose error of the galvanometer-camera mounting are corrected to obtain a static accuracy benchmark. S2. Dynamic testing phase; The robot moves at a constant speed v; it identifies and tracks the marked wood chips, initiates the strike program, and strikes the marked center of the wood chip with a laser. The image recognition module detects the location of the wood chip, its corners, and the center of the black spot, and calculates the pixel position P of the impact black spot. actual Pixel position P relative to the theoretical target center theory Then, calculate the spatial deviation Δd between the actual impact point and the theoretical target center; S3. Iterative optimization phase; When the impact deviation exceeds the set threshold, the time delay compensation parameter is calculated, and the delay Δt = Δd / v is calculated based on the robot's travel speed v; the system time compensation parameter is updated, where the initial compensation parameter T is... old Set to 0, and then iteratively update based on the time delay Δt calculated for each dynamic strike, T new =T old +Δt; Update system parameters and repeat the above process until the hit deviation of multiple consecutive tests is less than the set threshold, at which point the delay compensation is considered to have converged; S4. Parameter solidification stage; The final compensation value is written into the control system parameter area for automatic application later, and the dynamic calibration process can be restarted through the one-click calibration function.

2. The automatic calculation and compensation method for dynamic impact delay of the laser weeding robot according to claim 1, characterized in that, A fixed-size square wooden chip is used as a dynamic calibration target carrier. A target mark with a fixed position, such as the center position, is preset on the wooden chip. During the movement, the program detects the target and then strikes the target mark with a laser. By detecting the actual position of the black spot formed on the wooden chip after the laser strike, the strike deviation is calculated in the local coordinate system of the wooden chip. Combined with the robot's movement speed, the system time delay is deduced, realizing the self-learning and convergence of the delay compensation parameters.

3. The automatic calculation and compensation method for dynamic impact delay of the laser weeding robot according to claim 1, characterized in that, The calculation and compensation process is as follows. The first step is to detect the pixels at the four corners of the wood chip; The coordinates of the four corner pixels are represented as follows: ； The second step is to detect the central pixel of the dark spot; The coordinates of the center pixel of the black spot are represented as follows: ； The third step is to map pixels to the actual coordinates of the wood chips. Length and width of the wood chip: Let the length of the wood chip be L. x Width is L y ; Normalization coefficient: ； The position of the black spot relative to the center of the wood chip: First, define the pixel coordinates (u) of the center of the wood chip. c ,v c ); ； Then calculate the position of the dark spot relative to the center (X). hit ,Y hit ); ； The fourth step is to calculate the actual spatial deviation; The formula for calculating the actual spatial deviation Δd is: ； Fifth step, obtain the time Δt that needs to be compensated; ； In the formula, v is the robot's travel speed.

4. The automatic calculation and compensation method for dynamic impact delay of the laser weeding robot according to claim 3, characterized in that, The Δd calculated using the above method is the actual positional deviation between the target point and the impact point, and is independent of the orientation and angle of the wooden chip placement.

5. A system for automatically calculating and compensating for dynamic impact delay of a laser weeding robot according to any one of claims 1 to 4, characterized in that, Includes the following collaborative modules: The visual perception and imaging module includes an industrial camera and a supplementary lighting unit. The industrial camera is a high-frame-rate, high-pixel industrial camera used to acquire continuous image data of the work area. The supplementary lighting unit uses a high-brightness light source triggered synchronously with the camera to suppress ambient light changes and reflective interference, improving the imaging stability of the target wood chip and its marker points. The visual perception and imaging module is used to identify, track, and strike target wood chips in dynamic scenes, and to identify and extract sub-pixel-level feature points for strike positioning and the black spots left after the strike. The motion sensing module, including an IMU inertial measurement unit and a wheel speed encoder, is used to collect the pose information of the laser weeding robot in real time during its movement, output the robot's motion speed parameters in the world coordinate system, and provide motion prior data for dynamic delay calculation. The laser galvanometer strike execution module includes a laser, a dual-axis high-speed galvanometer, and a laser control board. The dual-axis high-speed galvanometer is used to achieve high-speed deflection and positioning of the laser spot within the target plane. The laser is used to output a high-energy laser beam for strike. The laser control board integrates a high-speed control unit for precise timing control of the laser pulses and galvanometer scanning control, supporting microsecond-level response. The central computing and control module, which is an industrial control computer or embedded PC platform, is used to run visual reasoning, motion prediction, and delay compensation control programs. Specifically, the central computing and control module deploys target recognition and corner / spot detection algorithms to perform real-time reasoning on image data acquired by the camera; it calculates the total system delay from image acquisition and data processing to the actual laser strike based on the detected impact error results and the velocity information output by the motion perception module; and it generates corresponding time compensation parameters based on the total delay.

Images