Alisma plantago-aquatica harvesting inter-row deviation rectifying system

Abstract

The invention discloses an inter-row correction system for rhizoma alismatis harvesting, and relates to the technical field of agricultural machinery automation, and the system comprises a sensing unit which is used for collecting the relative position information of a harvester and rhizoma alismatis crop rows and the attitude information of the harvester; the fusion unit is used for fusing the relative position information and the attitude information to obtain comprehensive lateral deviation and comprehensive course deviation; the deviation correction unit is used for obtaining a steering radius based on the comprehensive transverse deviation and the comprehensive course deviation, and obtaining a target speed difference between a left crawler belt and a right crawler belt of the harvester based on the steering radius; the execution unit is used for controlling the rotating speed and the heading of a left crawler belt and a right crawler belt of the harvester based on the target speed difference and controlling the position of a cutting device of the harvester based on the comprehensive transverse deviation. And the detection precision of the relative position of the harvester and the crop row is low.

Description

Technical Field

[0001] This invention relates to the field of agricultural machinery automation technology, specifically to a row correction system for harvesting Alisma plantago-aquatica. Background Technology

[0002] Alisma plantago-aquatica is an aquatic medicinal herb. At harvest time, the soil moisture content in the field is as high as 50%-60%. Tracked harvesters are often used to prevent the machine from sinking and affecting the quality of the harvest. The wider the tracks, the lower the ground pressure, and the less likely the machine is to get stuck in the mud. However, Alisma plantago-aquatica is a shallow-rooted medicinal herb, growing at a depth of less than 10cm with a row spacing of about 30cm. If the tracks are too wide, the roots and rhizomes on both sides are easily damaged during transport, leading to increased damage and economic losses. Furthermore, if the harvester deviates laterally during transport, the blades will veer off course, cutting or omitting large areas of Alisma plantago-aquatica roots, severely damaging the integrity and quality of the herb, affecting its selling price, and causing significant economic losses for farmers.

[0003] Currently, traditional Alisma plantago-aquatica harvesters mainly rely on the driver's experience and agricultural navigation technology based on GPS or vision. However, manual operation and correction by the driver are labor-intensive, and it is difficult to guarantee continuous accuracy under high-intensity operation. The accuracy of GPS positioning is easily affected by interference in the farmland environment and cannot provide direct offset relative to the crop row. Single vision navigation is not stable and reliable enough in complex environments such as changes in light, water reflection, and soil covering in muddy paddy fields. Summary of the Invention

[0004] To address the problem that current Alisma plantago-aquatica harvesting systems are ill-suited to the harvesting environment, resulting in low accuracy in detecting the relative position of the harvester and the crop row, this invention provides an Alisma plantago-aquatica harvesting row correction system, the system comprising:

[0005] Sensing unit: used to collect the relative position information of the harvester and the row of Alisma plantago-aquatica crops and the attitude information of the harvester;

[0006] Fusion unit: used to fuse the relative position information and the attitude information to obtain the comprehensive lateral deviation and comprehensive heading deviation;

[0007] Correction unit: used to obtain the turning radius based on the comprehensive lateral deviation and the comprehensive heading deviation, and to obtain the target speed difference between the left and right tracks of the harvester based on the turning radius;

[0008] Execution unit: used to control the rotational speed and heading of the left and right tracks of the harvester based on the target speed difference, and to control the position of the cutting device of the harvester based on the comprehensive lateral deviation.

[0009] A row of Alisma plantago-aquatica plants refers to a strip-shaped planting unit formed by planting Alisma plantago-aquatica plants in rows or columns according to certain spacing and patterns in an Alisma plantago-aquatica field in order to facilitate management and obtain the best growth effect.

[0010] This system utilizes multi-sensor information fusion technology, integrating visual (relative position), RTK-GPS (absolute position and heading), and IMU (high-frequency attitude) information. The sensing unit continuously monitors the relative position and attitude of the harvester and the rows of Alisma plantago-aquatica crops, overcoming the limitations of single sensors and significantly improving the accuracy and anti-interference capabilities of row detection in complex paddy field environments. Through a closed-loop control algorithm, the system dynamically adjusts the speed difference between the two tracks to correct the heading, reducing damage to the crops on both sides. Simultaneously, a lateral active adjustment mechanism for the cutting device is introduced, minimizing root and stem cuts and damage caused by deviation. The system intelligently controls the positions of the tracks and cutting device, adapting to changes in travel resistance due to different soil conditions, achieving smooth and stable correction movements, reducing mechanical impact and additional soil disturbance caused by sharp turns. Ultimately, this enables precise row-to-row harvesting between narrow-row crops, effectively reducing damage to crops and achieving automated operation, reducing reliance on driver experience, and improving operational efficiency and quality consistency.

[0011] Furthermore, the sensing unit specifically includes:

[0012] Machine vision unit: used to acquire a first image of the unharvested rows of Alisma plantago-aquatica in front of the harvester, and to obtain the center line of the crop row and the features of the Alisma plantago-aquatica crop row based on the first image. The features of the Alisma plantago-aquatica crop row include spectral features, morphological features and topological features.

[0013] Positioning unit: used to obtain the position coordinates and heading angle of the harvester;

[0014] Inertial measurement unit: used to acquire the three-dimensional acceleration and angular velocity of the harvester, and to obtain attitude angle information based on the three-dimensional acceleration and angular velocity, the attitude angle information including roll angle, pitch angle and yaw angle;

[0015] Information unit: used to obtain the relative position information and the attitude information based on the crop row centerline, the position coordinates, the heading angle and the attitude angle information.

[0016] Furthermore, the fusion unit specifically includes:

[0017] Visual deviation module: used to obtain a visual navigation reference based on the characteristics of the Alisma plantago-aquatica crop row, and to obtain a relative deviation based on the centerline of the crop row and the visual navigation reference, wherein the relative deviation includes visual lateral deviation and visual heading deviation;

[0018] Pose estimation module: used to obtain a state prediction equation based on the three-dimensional acceleration and the angular velocity, obtain an observation value based on the position coordinates and the heading angle, and obtain a global pose based on the state prediction equation and the observation value;

[0019] Virtual path module: used to acquire several consecutive images based on the first image, obtain reference center lines for several rows of Alisma plantago-aquatica based on the consecutive images, obtain a virtual reference path based on the reference center lines and the center lines of the crop rows, and obtain the virtual straight line equation of the virtual reference path;

[0020] Data fusion module: used to construct state vector and measurement vector based on the virtual straight line equation, the relative deviation and the global pose, obtain an adaptive weight matrix, and obtain the comprehensive lateral deviation and the comprehensive heading deviation based on the state vector, the measurement vector and the adaptive weight matrix.

[0021] A hierarchical fusion strategy is adopted. First, IMU / GPS is tightly coupled to solve the absolute position and high frequency problems. Then, the results are adaptively weighted and fused with vision. The hierarchical structure maximizes the advantages of each sensor and makes up for their respective disadvantages. Combined with adaptive weighted Kalman filtering, the relative position information of crop rows perceived by vision is used as the benchmark, and high-precision absolute position information is used for global correction. IMU data is used to compensate for high-frequency motion errors, and finally a stable and reliable integrated navigation state quantity is output.

[0022] Furthermore, the virtual path module is specifically used for:

[0023] Based on the first image, several consecutive images are obtained, and based on the consecutive images, several reference center lines of Alisma plantago-aquatica crop rows are obtained. The reference line equation of the reference center line is obtained. Based on the pre-trained inverse perspective transformation model, the reference line equation is converted into a three-dimensional world coordinate system with the harvester as the origin, and a three-dimensional point set of each reference center line is obtained.

[0024] Based on the three-dimensional point set, the reference centerline is fitted to obtain a cluster of parallel lines;

[0025] Based on the relative deviation, the global pose, and the parallel line cluster, the virtual reference path is obtained, the linear equation of the virtual reference path is obtained, and the virtual linear equation is obtained.

[0026] Unlike preset absolute paths, this system dynamically generates a virtual path. Based on the center lines of multiple crop rows identified from the latest few frames of images, it fits an ideal reference path parallel to the current crop row in real time. The machine tracks an ideal line that is completely parallel to the current crop row, rather than a preset straight line. This better adapts to the natural curvature that may occur during crop sowing, reducing the inherent deviation between the absolute path and the relative crop row. It also reduces systematic deviations caused by uneven sowing or cumulative errors in global positioning, improving the system's practicality and robustness in real farmland environments. Furthermore, the virtual reference path is a relative path generated in real time based on biological features (crop rows). Even if there are minute centimeter-level drifts in RTK-GPS, or cumulative errors in IMU calculations due to tire slippage, the path tracked by the system is always relative to the crop row. This fundamentally eliminates the impact of absolute positioning errors on harvest quality, ensuring that the plow is always accurately aligned with the row.

[0027] Considering the unevenness of farmland bottoms, mechanical digging at a fixed depth either fails to remove all the tubers, leaving some behind, or digging too deep increases power consumption and stirs up a large amount of bottom debris (such as stones and hard mud clods), making subsequent cleaning and sorting difficult. Traditional harvesters control depth through depth-limiting wheels or mechanical contouring mechanisms, but these are prone to sinking in soft mud, have lagging contouring, and cannot guarantee accuracy, which can also lead to digging too deep or too shallow. This system solves the problem of digging depth control caused by uneven mud bottoms by using a multi-sensor information fusion system to perceive the relative position of the digging components to the mud bottom in real time and intelligently adjust the digging depth of the water plantain tubers.

[0028] Furthermore, the system also includes:

[0029] Depth Unit: Used to obliquely install a millimeter-wave radar on the front frame of the harvester based on a preset forward tilt angle. The beam of the millimeter-wave radar is directed towards the mud bottom area to be excavated in front of the harvester. A laser rangefinder is installed at an adjacent position of the millimeter-wave radar. The millimeter-wave radar has a built-in tilt sensor. Based on the millimeter-wave radar, the laser rangefinder, and the tilt sensor, the relative height between the digging device of the harvester and the mud bottom surface is obtained.

[0030] Positioning unit: used to install encoders to various rotating hinge points of the lifting device of the harvester, and to obtain the absolute height of the digging device and the frame of the harvester based on the encoders;

[0031] Control unit: Used to construct a fuzzy rule table, and obtain control parameters based on the fuzzy rule table, the relative height, and the absolute height;

[0032] Excavation unit: Used to drive the excavation device to excavate Alisma plantago-aquatica based on the control parameters.

[0033] Relative height is used as the feedforward control variable to predict the terrain undulations ahead, while absolute height is used as the feedback control variable to represent the current actual position. Ranging is performed by non-contact millimeter-wave radar, combined with feedforward-feedback composite control, to detect terrain changes in advance, respond quickly, and control with high precision, greatly reducing the omission of Alisma tubers and over-excavation. The fuzzy PID algorithm is adopted, which has strong robustness to disturbances such as changes in soil resistance and machine vibration, and is suitable for complex silt environments. It effectively reduces the digging up of bottom impurities (stones, hard mud), reduces the workload of subsequent cleaning and sorting, and improves the quality and purity of Alisma products.

[0034] Furthermore, the depth unit is specifically used for:

[0035] The first slant range and confidence score are obtained based on the millimeter-wave radar, the second slant range and signal-to-noise ratio are obtained based on the laser rangefinder, and the radar attitude data of the millimeter-wave radar are obtained based on the tilt sensor.

[0036] If the signal-to-noise ratio is greater than the first threshold, then a measurement deviation is obtained based on the first slope distance and the second slope distance, a continuous deviation is obtained based on the measurement deviation, the first slope distance is compensated based on the continuous deviation to obtain a first calibrated slope distance, a third slope distance is obtained based on the first preset weight and the first calibrated slope distance, and a first distance is obtained based on the third slope distance.

[0037] If the signal-to-noise ratio is less than the second threshold, then the first distance is obtained based on the first slant range;

[0038] If the signal-to-noise ratio is greater than or equal to the second threshold and less than or equal to the first threshold, then a second preset weight is obtained based on the signal-to-noise ratio, and the first distance is obtained based on the second preset weight and the first calibration slope distance.

[0039] The relative height is obtained by compensating for the first distance based on the radar attitude data;

[0040] The first formula for calculating the relative height is:

[0041] ;

[0042] in, 1 indicates relative height. Indicates the first distance. Indicates the installation tilt angle of the millimeter-wave radar. This indicates the real-time elevation change angle of the millimeter-wave radar.

[0043] Millimeter-wave radar possesses extremely strong penetrating power, effectively penetrating the mud and water mist and sparse weeds stirred up during harvesting to directly detect the true mud bottom surface. It is not limited by visible optical conditions, and its FMCW (Frequency Modulated Continuous Wave) standard provides range resolution, filtering out false targets such as near-ground weeds. Mounted at a certain forward angle on the frame directly in front of the digging shovel and above the mud surface, with its beam center pointing to the mud bottom (pre-aiming point) a certain distance ahead of the digging shovel's working trajectory, it can provide terrain prediction time, solving the problem of system control lag. Combined with a laser rangefinder, which has extremely high accuracy and resolution at short distances, and installed close to the millimeter-wave radar with its beam pointing to the same pre-aiming point area, the laser sensor provides ultra-high precision reference data under unclogged conditions. Through data fusion algorithms, the data can be further analyzed. The system cross-validates the readings of millimeter-wave radar and laser sensors. When laser data becomes invalid due to extreme turbidity, the millimeter-wave radar data is automatically trusted. In favorable environments, laser data can be used to calibrate the millimeter-wave radar online, compensating for errors such as temperature drift, achieving complementary advantages. The tilt sensor is integrated with the radar, allowing real-time monitoring of the radar sensor's pitch and roll angles. Using the real-time attitude data from the tilt sensor, motion compensation is applied to the slant range measured by the radar, calculating the true geometric vertical height between the radar and the detection point on the mud bottom under any swaying conditions. By combining these three sensors to detect the relative height between the excavator blade and the surface of the mud bottom to be excavated, a forward-looking multi-source fusion sensing scheme is adopted, solving the industry problem of accurate distance measurement on soft and irregular mud bottom surfaces under harsh working conditions such as turbid mud, weed obstruction, and machine vibration.

[0044] Based on the laser signal-to-noise ratio, a quantifiable physical indicator that directly reflects the environmental state, a finite state machine with three states was constructed to achieve smooth and intelligent transition of the control strategy. A smooth transition dynamic weight allocation was adopted to reduce drastic jumps in output values ​​near environmental critical points, ensuring the system's control stability. A high-precision laser sensor, used as a benchmark under good operating conditions, was used to calibrate the millimeter-wave radar's system error online and in real time, solving the drift problem of millimeter-wave radar during long-term use, reducing the stringent requirements for the sensor's long-term absolute accuracy, and improving the system's long-term reliability.

[0045] Furthermore, the position unit is specifically used for:

[0046] The encoder is used to obtain the measured hinge angle value between two adjacent rigid bodies.

[0047] A three-dimensional kinematic model of the harvester is constructed, and the first length between the rotational hinge points is obtained based on the three-dimensional kinematic model;

[0048] The theoretical height is obtained based on the measured hinge angle value, the first length, and the three-dimensional kinematic model;

[0049] An inertial measurement unit is installed at a key point on the frame of the harvester. The frame pitch angle is obtained based on the inertial measurement unit. The absolute height is obtained based on the frame pitch angle and the theoretical height.

[0050] The second formula for obtaining the absolute height is:

[0051] ;

[0052] in, 2 indicates absolute height. This represents the initial height of the origin of the machine coordinate system. Indicates the rack pitch angle, This indicates a high level of theoretical understanding.

[0053] A precise kinematic model of the excavating mechanism was established, and the ultra-high precision absolute height of the excavator blade relative to the dynamic virtual reference plane of the frame was calculated through forward kinematics. This overcame the problems of indirect measurement and large error of traditional single sensor measurement. Furthermore, a real-time compensation mechanism for the sensor was introduced to compensate for the significant error source of frame deformation, which was often overlooked. This ensured that the measurement results truly reflected the height relationship between the blade and the ground level.

[0054] Furthermore, the position unit is also used for:

[0055] A calibration sensor is installed on the piston rod of the hydraulic cylinder of the harvester, and the cylinder stroke value is obtained based on the calibration sensor.

[0056] Based on the cylinder stroke value and the inverse kinematics algorithm, the inferred hinge angle value between two adjacent rigid bodies is obtained;

[0057] The difference between the measured hinge angle value and the inferred hinge angle value is obtained, and an early warning result is obtained based on the difference and the tolerance range.

[0058] Furthermore, the control unit is specifically used for:

[0059] Obtain a preset height, and based on the preset height, the relative height, and the absolute height, obtain the target height;

[0060] The fuzzy rule table includes several fuzzy rules, each of which corresponds to a fuzzy set. Each fuzzy set includes several linguistic variables, and each linguistic variable corresponds to a single point value.

[0061] Based on the target height and the absolute height, the error and the error rate are obtained. The error and the error rate are converted into fuzzy language values. Based on the fuzzy language values ​​and the fuzzy rule table, activation rules are obtained. Based on the activation rules, an activation fuzzy set is obtained.

[0062] The first membership degree and the second membership degree of the error and the error rate in the activated fuzzy set are obtained respectively. The activation strength of the activation rule is obtained based on the first membership degree and the second membership degree. The effective output value of the activation rule is obtained based on the activation strength.

[0063] Determine whether the activation intensity is greater than the intensity threshold. If so, calculate the corrected value of the language variable by weighting the effective output value and the activation intensity. Otherwise, obtain the corrected value of the language variable based on the default value.

[0064] The control parameters are obtained based on the correction value and the initial parameters;

[0065] The third calculation formula for obtaining the target height is:

[0066] ;

[0067] in, Indicates the target height. Indicates the preset height.

[0068] A feedforward-feedback composite fuzzy PID control architecture is adopted. H1 (feedforward) provides predictability, enabling the system to respond to terrain changes in advance, greatly reducing system lag, and transforming the target height into a target height trajectory that dynamically changes in the absolute coordinate system and follows the terrain. Fuzzy PID allows the controller parameters to be dynamically adjusted according to the current error state, responding quickly when the error is large and finely adjusting when approaching the target, preventing overshoot oscillation, and perfectly adapting to nonlinear factors such as changes in silt resistance. At the same time, a single-point output fuzzy set and weighted average method are introduced, simplifying the complex area calculation into an extremely simple multiplication operation, making the defuzzification process extremely fast. An activation threshold and default output mechanism are also introduced, so that the controller will not produce abnormal output when encountering unforeseen input combinations or slight sensor anomalies, effectively solving the problem of unstable output or even meaningless control commands in the traditional weighted average method under low activation conditions, greatly enhancing the robustness of the system under unknown or uncertain operating conditions.

[0069] Furthermore, the mining unit is specifically used for:

[0070] Map the control parameters to the target PWM duty cycle;

[0071] Construct a duty cycle-target current curve table, and obtain the target current value based on the duty cycle-target current curve table and the target PWM duty cycle;

[0072] The actual current value of the excavating device is obtained; based on the target current value and the actual current value, the current error is obtained; based on the current error, the target PWM duty cycle is adjusted to obtain the corrected PWM signal.

[0073] The modified PWM signal is superimposed with the dizziness signal to obtain the PWM command;

[0074] The excavation device is driven to excavate Alisma plantago-aquatica based on the PWM command.

[0075] The curve is used to compensate for the dead zone and nonlinearity of the proportional valve. A current negative feedback closed loop is introduced to effectively overcome the interference of factors such as coil resistance changes with temperature, power supply voltage fluctuations, and electromagnetic back electromotive force. This ensures a highly linear and repeatable relationship between the force / flow supplied to the proportional valve and the control command. The superposition of the chatter signal effectively overcomes the static friction of the valve core, significantly reduces the hysteresis phenomenon of the proportional valve, and makes the control more linear. This solves the problems of poor switching control accuracy of traditional solenoid valves and the susceptibility of ordinary proportional valve control to oil temperature and load, which can lead to slow digging shovel movement, vibration, or inaccurate positioning.

[0076] One or more technical solutions provided by this invention have at least the following technical effects or advantages:

[0077] 1. This system utilizes multi-sensor information fusion technology, integrating visual (relative position), RTK-GPS (absolute position and heading), and IMU (high-frequency attitude) information. The sensing unit continuously perceives the relative position and attitude information of the harvester body and the rows of Alisma plantago-aquatica crops, overcoming the limitations of single sensors and significantly improving the accuracy and anti-interference capability of row detection in complex paddy field environments. Furthermore, through a closed-loop control algorithm, the system dynamically adjusts the speed difference between the two tracks to correct the heading, reducing damage to the herbs on both sides. Simultaneously, a lateral active adjustment mechanism for the cutting device is introduced, minimizing root and stem cuts and damage caused by deviation. The system intelligently controls the positions of the two tracks and the cutting device, adapting to changes in travel resistance due to different soil conditions, achieving smooth and stable correction movements, reducing mechanical impact and additional soil disturbance caused by sharp turns. Ultimately, this enables precise row-to-row harvesting between narrow-row crops, effectively reducing damage to herbs and achieving automated operation, reducing reliance on driver experience, and improving operational efficiency and quality consistency.

[0078] 2. Unlike preset absolute paths, this system dynamically generates a virtual path. Based on the center lines of multiple crop rows identified from the latest few frames of images, it fits an ideal reference path parallel to the current crop row in real time. The machine tracks an ideal line that is completely parallel to the current crop row, rather than a preset straight line. This better adapts to the natural curvature that may occur during crop sowing, reducing the inherent deviation between the absolute path and the relative crop row. It also reduces systematic deviations caused by uneven sowing or cumulative errors in global positioning, improving the system's practicality and robustness in real farmland environments. Furthermore, the virtual reference path is a relative path generated in real time based on biological features (crop rows). Even if there are minute centimeter-level drifts in RTK-GPS, or cumulative errors in IMU calculations due to tire slippage, the path tracked by the system is always relative to the crop row. This fundamentally eliminates the impact of absolute positioning errors on harvest quality, ensuring that the plow is always accurately aligned with the row.

[0079] 3. This system uses relative height as the feedforward control variable to predict the terrain undulations ahead, and absolute height as the feedback control variable to represent the current actual position. It uses non-contact millimeter-wave radar for ranging, combined with feedforward-feedback composite control, to detect terrain changes in advance, respond quickly, and control with high precision, greatly reducing the omission of Alisma tubers and over-excavation. It adopts a fuzzy PID algorithm, which has strong robustness to disturbances such as changes in soil resistance and machine vibration, and is suitable for complex silt environments. It effectively reduces the digging up of bottom impurities (stones, hard mud), reduces the workload of subsequent cleaning and sorting, and improves the quality and purity of Alisma products.

[0080] 4. Based on the laser signal-to-noise ratio, a quantifiable physical indicator that directly reflects the environmental state, a finite state machine with three states was constructed to achieve smooth and intelligent transition of the control strategy. A smooth transition dynamic weight allocation was adopted to reduce drastic jumps in output values ​​near environmental critical points, ensuring the system's control stability. A high-precision laser sensor, used as a benchmark under good operating conditions, was used to calibrate the millimeter-wave radar's system error online and in real time, solving the drift problem of millimeter-wave radar during long-term use, reducing the stringent requirements for the sensor's long-term absolute accuracy, and improving the system's long-term reliability.

[0081] 5. A precise kinematic model of the excavating mechanism is established, and the ultra-high precision absolute height of the excavator blade relative to the dynamic virtual reference plane of the frame is calculated through forward kinematics. This overcomes the problems of indirect measurement and large error of traditional single sensor measurement. Furthermore, a real-time compensation mechanism for the sensor is introduced to compensate for the significant error source of frame deformation, which has been neglected, so that the measurement results truly reflect the height relationship between the blade and the ground level.

[0082] 6. A feedforward-feedback composite fuzzy PID control architecture is adopted. H1 (feedforward) provides predictability, enabling the system to respond to terrain changes in advance, greatly reducing system lag, and transforming the target height into a dynamically changing target height trajectory that follows the terrain in an absolute coordinate system. Fuzzy PID allows the controller parameters to be dynamically adjusted according to the current error state, responding quickly when the error is large and finely adjusting when approaching the target, preventing overshoot oscillation, and perfectly adapting to nonlinear factors such as changes in silt resistance. At the same time, a single-point output fuzzy set and weighted average method are introduced, simplifying the complex area calculation into an extremely simple multiplication operation, making the defuzzification process extremely fast. An activation threshold and default output mechanism are also introduced, so that the controller will not produce abnormal output when encountering unforeseen input combinations or slight sensor anomalies, effectively solving the problem of unstable output or even meaningless control commands in the traditional weighted average method under low activation conditions, greatly enhancing the robustness of the system under unknown or uncertain operating conditions. Attached Figure Description

[0083] The accompanying drawings, which are provided to further illustrate embodiments of the invention and constitute a part of this invention, are not intended to limit the scope of the invention.

[0084] Figure 1 This is a schematic diagram of a row correction system for harvesting Alisma plantago-aquatica according to the present invention. Detailed Implementation

[0085] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, where there is no conflict, the embodiments of the present invention and the features thereof can be combined with each other.

[0086] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0087] Example 1

[0088] refer to Figure 1 This embodiment provides a row correction system for Alisma plantago-aquatica harvesting, the system comprising:

[0089] Sensing unit: used to collect the relative position information of the harvester and the row of Alisma plantago-aquatica crops and the attitude information of the harvester;

[0090] Fusion unit: used to fuse the relative position information and the attitude information to obtain the comprehensive lateral deviation and comprehensive heading deviation;

[0091] Correction unit: used to obtain the turning radius based on the comprehensive lateral deviation and the comprehensive heading deviation, and to obtain the target speed difference between the left and right tracks of the harvester based on the turning radius; if fuzzy PID control or pure pursuit model algorithm is used, the comprehensive lateral deviation and heading deviation are used as inputs to calculate the theoretical turning radius required to eliminate the deviation, and then the target speed difference between the left and right tracks is calculated.

[0092] Execution unit: used to control the rotational speed and heading of the left and right tracks of the harvester based on the target speed difference; and to control the position of the cutting device of the harvester based on the comprehensive lateral deviation.

[0093] If the left and right tracks are controlled by a track drive control unit, the track drive control unit includes two independent electro-hydraulic proportional valves or servo motors for the left and right tracks. These valves are used to precisely control the speed of the hydraulic motors or electric motors of the left and right tracks, thereby achieving differential steering and adjusting the machine's heading. After receiving the target speed difference, the track drive control unit compares it with the actual speed value fed back by the encoder or speed sensor, calculates the error signal, processes the error signal through a PID control algorithm, and generates an analog voltage signal (such as 0-10V) or a pulse width modulation (PWM) signal. This signal is then transmitted to the electro-hydraulic proportional valve (if hydraulically driven) or the servo motor driver (if electrically driven). The electro-hydraulic proportional valve adjusts the flow to the hydraulic motor proportionally according to the input voltage, thereby precisely controlling the motor speed.

[0094] If the cutting device is a plow blade, the plow blade is equipped with a plow blade controller (such as a stepper motor driver or a small servo driver), which allows for fine-tuning of the plow blade in the lateral direction (e.g., ±5cm). It compensates for the calculated overall lateral deviation by making corrective movements, ensuring that even with slight machine body misalignment, the plow blade's entry point remains aligned with the center of the crop row. The plow blade controller converts the overall lateral deviation into the number of pulses required for the stepper motor or servo motor to rotate based on mechanical transmission parameters (such as the lead of the lead screw). Upon receiving the drive pulses, the stepper motor or servo motor begins to rotate. This rotational motion is transmitted via a coupling to a precision ball screw pair (or other mechanisms that convert rotational motion into linear motion, such as a rack and pinion). The screw nut drives a slide or bracket fixedly connected to the plow blade, performing precise lateral linear motion on a linear guide rail. This achieves real-time, high-precision lateral compensation of the plow blade's position, ensuring that its entry point remains aligned with the center of the crop row, regardless of whether the machine body is undergoing steering correction.

[0095] Specifically, the sensing unit includes:

[0096] Machine vision unit: used to acquire a first image of the unharvested water plantain row in front of the harvester, and to obtain the center line of the crop row and the features of the water plantain crop row based on the first image. The features of the water plantain crop row include spectral features, morphological features and topological features. For example, a front-facing camera is installed on the front crossbeam of the harvester to acquire image information of the unharvested water plantain row in front. The center line of the crop row is identified by an image processing algorithm, and the features of the water plantain crop row are extracted by an image segmentation algorithm based on color features (green vegetation and muddy background) and morphological processing.

[0097] Positioning unit: used to install an existing RTK-GPS on top of the harvester to obtain the harvester's position coordinates and heading angle;

[0098] Inertial Measurement Unit: Used to install an inertial measurement unit (IMU) near the center of gravity of the harvester to acquire the three-dimensional acceleration and angular velocity of the harvester, and to obtain attitude angle information based on the three-dimensional acceleration and angular velocity, the attitude angle information including roll angle, pitch angle and yaw angle;

[0099] Information unit: used to obtain the relative position information and the attitude information based on the crop row centerline, the position coordinates, the heading angle and the attitude angle information.

[0100] The fusion unit specifically includes:

[0101] Visual deviation module: used to obtain a visual navigation reference based on the characteristics of the Alisma plantago-aquatica crop row, and to obtain a relative deviation based on the centerline of the crop row and the visual navigation reference. The relative deviation includes visual lateral deviation (i.e., the lateral distance between the machine's current centerline and the crop row centerline) and visual heading deviation (i.e., the angular difference between the machine's current direction of travel and the ideal navigation reference direction calculated based on the visual sensor). For example, the visual navigation reference line in the current image can be fitted using Hough Transform or least squares method, and a visual lateral deviation and a visual heading deviation can be calculated.

[0102] The pose estimation module is used to obtain a state prediction equation based on the three-dimensional acceleration and the angular velocity, obtain observations based on the position coordinates and the heading angle, and obtain a global pose based on the state prediction equation and the observations. For example, Kalman filtering is used to tightly combine the position coordinates, heading angle, three-dimensional acceleration, and angular velocity: using three-dimensional acceleration and angular velocity as the state prediction equation and using the position coordinates and heading angle as the observations for updating, it can smooth the jump point noise of RTK-GPS, and use IMU to perform short-time high-precision estimation when the GPS signal is briefly lost (such as when passing through an obstruction), and finally output a high-frequency, smooth fused global pose.

[0103] Virtual path module: used to acquire several consecutive images based on the first image, obtain reference center lines for several rows of Alisma plantago-aquatica based on the consecutive images, obtain a virtual reference path based on the reference center lines and the center lines of the crop rows, and obtain the virtual straight line equation of the virtual reference path;

[0104] The data fusion module is used to construct state vectors and measurement vectors based on the virtual straight-line equation, the relative deviation, and the global pose; obtain an adaptive weight matrix; and obtain the comprehensive lateral deviation and the comprehensive heading deviation based on the state vector, the measurement vector, and the adaptive weight matrix. For example, the global lateral deviation is calculated based on the virtual reference path and the global pose; and the straight-line equation of the virtual reference path is transformed to the global geodetic coordinate system based on the global pose, resulting in the following straight-line equation: ( Indicates the slope. (representing slant distance), treating the machine's current pose as a point ( , ), calculate the distance from the point to the equation of the line. Based on this distance and the equation of the straight line, the global lateral deviation is obtained. The distance in this formula is... It inherently contains symbols, if If the machine is tilted to the right, it needs to be corrected to the left; if If the value is 0, the machine is biased to the left and needs to be corrected to the right; if the value is 0, no correction is needed.

[0105] The Kalman Filter (KF) or Extended Kalman Filter (EKF) algorithm is used to construct a state vector based on the combined lateral deviation, the derivative (rate of change) of the combined lateral deviation, and the combined heading deviation. A state transition equation is then constructed based on the state vector. A measurement vector is constructed based on the relative deviation, global pose, and global lateral deviation. An observation equation is then constructed based on the measurement vector to obtain the existing adaptive weight matrix. Based on the Kalman Filter recursive algorithm, at each time step, the prediction and update steps of the standard Kalman Filter are executed to finally obtain the optimal estimated combined lateral deviation and combined heading deviation.

[0106] Specifically, the virtual path module is used for:

[0107] Based on the first image, several consecutive images are obtained, and based on the consecutive images, several reference center lines of Alisma plantago-aquatica crop rows are obtained. The reference straight line equation of the reference center line is obtained. If multiple complete Alisma plantago-aquatica crop rows in the current field of view are simultaneously identified using computer vision algorithms (such as semantic segmentation models based on deep learning or optimized traditional image processing procedures), the straight line equation of the center line of each identified crop row is fitted using the least squares method.

[0108] Based on a pre-trained inverse perspective transformation model, the equation of the reference line is converted into a three-dimensional world coordinate system with the harvester as the origin, and a three-dimensional point set of each reference center line is obtained. For example, a pre-calibrated inverse perspective transformation model from the image pixel coordinate system to the harvester body coordinate system is established. This model is applied to the equation of each center line, and it is converted from two-dimensional image coordinates to three-dimensional world coordinates (body coordinate system) with the harvester as the origin. After the conversion, each crop row center line can be expressed as a set of points in the body coordinate system, that is, a three-dimensional point set.

[0109] Based on the three-dimensional point set, the reference centerline is fitted to obtain a cluster of parallel lines; or based on the three-dimensional point set, the least squares method or the Random Sample Consensus (RANSAC) algorithm is used to fit the centerline of all crop rows to establish a cluster of parallel lines. The cluster of parallel lines represents the geometric model of the crop rows and is used to estimate the overall orientation (i.e., direction vector) and average row spacing of all crop rows in the current working area.

[0110] Based on the relative deviation, the global pose, and the parallel line cluster, the virtual reference path is obtained, the linear equation of the virtual reference path is obtained, and the virtual linear equation is obtained. For example, based on the harvester's current position (from RTK-GPS / IMU fusion results) and agronomic requirements (e.g., requiring the plow blade to be aligned with the middle of two rows of crops), a line is dynamically designated from the aforementioned parallel line cluster as the current virtual reference path. If it is determined that the machine is currently operating between the first and second rows, the virtual reference path will be set as the straight line parallel to the center lines of these two rows and located exactly in their middle. The linear equation of the virtual reference path is then determined in the machine's coordinate system.

[0111] Example 2

[0112] Based on Embodiment 1, in this embodiment, the system further includes:

[0113] Depth Unit: Used to obliquely mount a millimeter-wave radar on the front frame of the harvester based on a preset forward tilt angle. The beam of the millimeter-wave radar points towards the mud bottom area to be excavated in front of the harvester (i.e., the pre-aiming point). A laser rangefinder sensor is installed at an adjacent position of the millimeter-wave radar. The millimeter-wave radar has a built-in tilt sensor (such as an inertial measurement unit, with a built-in three-axis gyroscope and a three-axis accelerometer, which can be directly installed inside the housing of the millimeter-wave radar sensor). Based on the millimeter-wave radar, the laser rangefinder sensor, and the tilt sensor, the relative height between the harvester's digging device and the mud bottom surface is obtained. The millimeter-wave radar has strong penetration and is not affected by water vapor or mud splashes. It can accurately measure the oblique distance between the sensor and the mud bottom point, and calculate the vertical height through the data from the built-in tilt sensor.

[0114] Positioning unit: Used to install encoders (such as existing multi-turn absolute encoders) to various rotating hinge points of the lifting device of the harvester (such as the hinge point between the boom and the frame, and the hinge point between the stick and the boom), and to obtain the absolute height of the digging device and the frame of the harvester based on the encoders; for example, by installing encoders at the lifting hydraulic cylinder of the digging shovel or the hinge point of the lifting linkage, the cylinder stroke or hinge point angle is directly measured to calculate the current absolute height of the digging shovel blade relative to the dynamic reference system of the frame.

[0115] Control unit: Used to construct a fuzzy rule table, and obtain control parameters based on the fuzzy rule table, the relative height, and the absolute height;

[0116] Excavation unit: Used to drive the excavation device to excavate Alisma plantago-aquatica based on the control parameters.

[0117] Specifically, the depth unit is used for:

[0118] The millimeter-wave radar is used to obtain a first slant range and a confidence score. This score can be obtained based on the target intensity and echo quality indicators provided by the radar's internal DSP (Digital Signal Processing) technology, or by calculating the variance of the most recent set of radar data. A small variance indicates stable data and thus a high confidence score. The laser range sensor is used to obtain a second slant range and a signal-to-noise ratio (SNR). The SNR is a key indicator for determining whether the laser has failed. The tilt sensor is used to obtain the radar attitude data of the millimeter-wave radar.

[0119] If the signal-to-noise ratio is greater than the first threshold, it indicates that the environment is in excellent condition, the water is clear and unobstructed, and the laser sensor is working in optimal condition. Then, a measurement deviation is obtained based on the first slope distance and the second slope distance, where the measurement deviation = first slope distance - second slope distance. A continuous deviation is obtained based on the measurement deviation. If the measurement deviation is input into an existing first-order low-pass filter or moving average window, a smooth and continuous deviation estimate is formed. The first slope distance is compensated based on the continuous deviation to obtain a first calibrated slope distance, where the first calibrated slope distance = first slope distance - continuous deviation. A third slope distance is obtained based on the first preset weight and the first calibrated slope distance, where the third slope distance = first preset weight * second slope distance + (1 - first preset weight) * first calibrated slope distance. A first distance is obtained based on the third slope distance.

[0120] If the signal-to-noise ratio is less than the second threshold, it indicates that the environment is in a bad state, the water is extremely turbid, the laser is completely ineffective or the data is unreliable. Then the first distance is obtained based on the first slant range, with the output of the millimeter-wave radar as the standard.

[0121] If the signal-to-noise ratio is greater than or equal to the second threshold and less than or equal to the first threshold, it indicates that the environment is in a transitional state, the water is slightly turbid, and the laser data quality is degraded but not completely invalid. Then, a second preset weight is obtained based on the signal-to-noise ratio. If the second preset weight is linearly interpolated or exponentially decayed between 0 and 0.9, the worse the laser data quality, the lower its weight, and the millimeter-wave radar weight increases accordingly. Based on the second preset weight and the first calibration slant range, the first distance is obtained. At this time, the first distance = second preset weight * second slant range + (1 - second preset weight) * first calibration slant range.

[0122] The relative height is obtained by compensating for the first distance based on the radar attitude data;

[0123] The first formula for calculating the relative height is:

[0124] ;

[0125] in, 1 indicates relative height. Indicates the first distance. Indicates the installation tilt angle of the millimeter-wave radar. This indicates the real-time elevation change angle of the millimeter-wave radar.

[0126] Specifically, the position unit is used for:

[0127] The encoder is used to obtain the measured hinge angle value between two adjacent rigid bodies.

[0128] A three-dimensional kinematic model of the harvester is constructed, treating the harvester as a robotic arm rather than a simple four-bar linkage. Based on the three-dimensional kinematic model, the first length between the rotating hinge points is obtained.

[0129] Based on the measured hinge angle value, the first length, and the three-dimensional kinematic model, forward kinematic calculations are performed according to geometric principles (such as the law of cosines, DH parameter method, etc.) to directly and in real-time calculate the precise three-dimensional coordinates (X, Y, Z) of the excavator blade edge relative to the origin of the predefined mechanical coordinate system on the frame in the current posture. The Z coordinate is the required absolute height value, i.e., the theoretical height. This method directly obtains the spatial position of the blade edge, reducing the accumulated error and accuracy loss caused by measuring the cylinder stroke and then indirectly converting it through complex trigonometric functions.

[0130] An inertial measurement unit is installed at a key point on the frame of the harvester. The frame pitch angle is obtained based on the inertial measurement unit. The absolute height is obtained based on the frame pitch angle and the theoretical height.

[0131] The second formula for obtaining the absolute height is:

[0132] ;

[0133] in, 2 indicates absolute height. This represents the initial height of the origin of the machine coordinate system. Indicates the rack pitch angle, This indicates a high level of theoretical understanding.

[0134] Considering that the frame itself is not absolutely rigid, it will undergo torsional deformation when traveling or digging in rugged muddy terrain. This causes the absolute coordinate system established on the frame to actually change dynamically. Therefore, this system installs an auxiliary IMU (Inertial Measurement Unit) at a key location on the frame (such as near the base of the digging mechanism) to monitor the pitch and roll changes of the frame itself. By reading the pitch angle of this IMU, the final compensated absolute height of the digging blade edge = theoretical height + The height compensation calculated based on the frame pitch angle and the position of the origin of the mechanical coordinate system ensures that the height measurement is relative to a horizontal virtual reference, rather than a physical reference that sways with the vehicle body, achieving an accuracy level comparable to that of industrial robots.

[0135] In this embodiment, rigid bodies refer to the main components that make up the digging device of the Alisma harvester. These components are idealized as solid objects that do not undergo elastic deformation or bending during movement.

[0136] Specifically, the control unit is used for:

[0137] Obtain a preset height, and based on the preset height, the relative height, and the absolute height, obtain the target height;

[0138] The fuzzy rule table includes several fuzzy rules, such as IF error is PB (positive, i.e., shallow) AND error rate is ZO (error change rate is zero), THEN proportional gain is PB (significantly increase proportional gain), integral gain is PM (moderately increase integral gain), and derivative gain is NS (slightly decrease derivative gain).

[0139] Each of the aforementioned fuzzy rules corresponds to a fuzzy set, and each fuzzy set includes several linguistic variables, with each linguistic variable corresponding to a single point value;

[0140] The error and error rate are obtained based on the target height and the absolute height. Error = target height - absolute height. The error rate can be the difference between the current error and the previous error divided by the period time.

[0141] The PID parameters are dynamically tuned using fuzzy logic, the error and the error rate are converted into fuzzy linguistic values, activation rules are obtained based on the fuzzy linguistic values ​​and the fuzzy rule table, and activation fuzzy sets are obtained based on the activation rules.

[0142] Using existing algorithms, the first membership degree and the second membership degree of the error and the error rate in the activated fuzzy set are obtained respectively. The activation strength of the activation rule is obtained based on the first membership degree and the second membership degree, that is, the smaller operation is performed on the first membership degree and the second membership degree. The effective output value of the activation rule is obtained based on the activation strength. The effective output value = activation strength * single point value.

[0143] If the activation intensity is greater than the intensity threshold, the fuzzy inference result of the current state is considered reliable. Only rules with activation intensities exceeding the intensity threshold are aggregated, and the corrected value of the language variable is obtained by weighted averaging of the effective output value and the activation intensity, such as proportional gain = total effective output value / total activation intensity. Otherwise, the current state is considered to be undefined and unknown, and the corrected value of the language variable is obtained based on the default value. This is an intelligent decision-making process that involves no action or maintaining the status quo, preventing the system from oscillating or becoming unstable due to random parameter adjustments in an uncertain state.

[0144] The control parameters are obtained based on the correction value and the initial parameters; the correction value is added to the initial value of the PID parameters to obtain the final parameters used in the current control cycle, and then standard PID calculation is performed to generate the basic control quantity and obtain the control parameters.

[0145] The third calculation formula for obtaining the target height is:

[0146] ;

[0147] in, Indicates the target height. Indicates the preset height. Based on geometric relationships, it is approximately equal to the absolute height of the mud bottom at the forward aiming point, serving as a feedforward quantity to inform the controller in advance how the terrain ahead will change. Subtracting this... If the final position of the shovel blade is lower than the mud bottom of this aiming point by one target depth, then the formula cleverly incorporates the relative depth of the target. It is transformed into a target height trajectory that dynamically changes in an absolute coordinate system and follows the terrain.

[0148] Specifically, the mining unit is used for:

[0149] Based on existing algorithms, the control parameters are mapped to the target PWM (pulse width modulation) duty cycle;

[0150] Construct a duty cycle-target current curve table, and obtain the target current value based on the duty cycle-target current curve table and the target PWM duty cycle;

[0151] The actual current value of the excavation device is obtained, for example, by using an existing high-precision, low-drift Hall current sensor to detect the actual current value flowing through the proportional valve coil in real time. Based on the target current value and the actual current value, the current error (target current value - actual current value) is obtained. Based on the current error, the target PWM duty cycle is adjusted to obtain a corrected PWM signal. For example, when the actual current is too low, the duty cycle of the output PWM is immediately increased to increase the current.

[0152] The modified PWM signal is superimposed with a high-frequency, low-amplitude dithering signal to obtain a PWM command; the dithering signal can be digitally synthesized by the software algorithm inside the microcontroller unit (MCU) (such as a timer interrupt service routine) combined with its built-in PWM generator hardware module.

[0153] The excavation device is driven to excavate water chestnut based on the PWM command. For example, an electro-hydraulic proportional valve is connected to the hydraulic cylinder that controls the raising and lowering of the excavator shovel. The opening of the electro-hydraulic proportional valve is controlled based on the PWM command. The valve core opening is precisely proportional to the current magnitude, thereby precisely controlling the flow rate and direction of hydraulic oil to the cylinder. The controlled hydraulic oil drives the cylinder piston rod to extend or retract, and through a mechanical linkage mechanism, ultimately drives the excavator shovel to raise and lower precisely and smoothly, completing adaptive depth adjustment.

[0154] The position unit is further used for:

[0155] A calibration sensor (such as an existing magnetostrictive linear displacement sensor) is installed on the piston rod of the hydraulic cylinder of the harvester, and the cylinder stroke value is obtained based on the calibration sensor;

[0156] Based on the cylinder stroke value and the inverse kinematics algorithm, the inferred hinge angle value between two adjacent rigid bodies is obtained;

[0157] The difference between the measured hinge angle value and the inferred hinge angle value is obtained. Based on the difference and the tolerance range, an early warning result is obtained. If the difference exceeds the tolerance range, it is determined that there may be a sensor failure, abnormal deformation of the mechanical structure, or severe slippage.

[0158] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0159] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A system for correcting row deviation during the harvesting of Alisma plantago-aquatica, characterized in that, The system includes: Sensing unit: used to collect the relative position information of the harvester and the row of Alisma plantago-aquatica crops and the attitude information of the harvester; Fusion unit: used to fuse the relative position information and the attitude information to obtain the comprehensive lateral deviation and comprehensive heading deviation; Correction unit: used to obtain the turning radius based on the comprehensive lateral deviation and the comprehensive heading deviation, and to obtain the target speed difference between the left and right tracks of the harvester based on the turning radius; Execution unit: used to control the rotational speed and heading of the left and right tracks of the harvester based on the target speed difference, and to control the position of the cutting device of the harvester based on the comprehensive lateral deviation.

2. The row correction system for harvesting Alisma plantago-aquatica according to claim 1, characterized in that, The sensing unit specifically includes: Machine vision unit: used to acquire a first image of the unharvested rows of Alisma plantago-aquatica in front of the harvester, and to obtain the center line of the crop row and the features of the Alisma plantago-aquatica crop row based on the first image. The features of the Alisma plantago-aquatica crop row include spectral features, morphological features and topological features. Positioning unit: used to obtain the position coordinates and heading angle of the harvester; Inertial measurement unit: used to acquire the three-dimensional acceleration and angular velocity of the harvester, and to obtain attitude angle information based on the three-dimensional acceleration and angular velocity, the attitude angle information including roll angle, pitch angle and yaw angle; Information unit: used to obtain the relative position information and the attitude information based on the crop row centerline, the position coordinates, the heading angle and the attitude angle information.

3. The row correction system for harvesting Alisma plantago-aquatica according to claim 2, characterized in that, The fusion unit specifically includes: Visual deviation module: used to obtain a visual navigation reference based on the characteristics of the Alisma plantago-aquatica crop row, and to obtain a relative deviation based on the centerline of the crop row and the visual navigation reference, wherein the relative deviation includes visual lateral deviation and visual heading deviation; Pose estimation module: used to obtain a state prediction equation based on the three-dimensional acceleration and the angular velocity, obtain an observation value based on the position coordinates and the heading angle, and obtain a global pose based on the state prediction equation and the observation value; Virtual path module: used to acquire several consecutive images based on the first image, obtain reference center lines for several rows of Alisma plantago-aquatica based on the consecutive images, obtain a virtual reference path based on the reference center lines and the center lines of the crop rows, and obtain the virtual straight line equation of the virtual reference path; Data fusion module: used to construct state vector and measurement vector based on the virtual straight line equation, the relative deviation and the global pose, obtain an adaptive weight matrix, and obtain the comprehensive lateral deviation and the comprehensive heading deviation based on the state vector, the measurement vector and the adaptive weight matrix.

4. The row correction system for harvesting Alisma plantago-aquatica according to claim 3, characterized in that, The virtual path module is specifically used for: Based on the first image, several consecutive images are obtained, and based on the consecutive images, several reference center lines of Alisma plantago-aquatica crop rows are obtained. The reference line equation of the reference center line is obtained. Based on the pre-trained inverse perspective transformation model, the reference line equation is converted into a three-dimensional world coordinate system with the harvester as the origin, and a three-dimensional point set of each reference center line is obtained. Based on the three-dimensional point set, the reference centerline is fitted to obtain a cluster of parallel lines; Based on the relative deviation, the global pose, and the parallel line cluster, the virtual reference path is obtained, the linear equation of the virtual reference path is obtained, and the virtual linear equation is obtained.

5. The row correction system for harvesting Alisma plantago-aquatica according to claim 1, characterized in that, The system also includes: Depth Unit: Used to obliquely install a millimeter-wave radar on the front frame of the harvester based on a preset forward tilt angle. The beam of the millimeter-wave radar is directed towards the mud bottom area to be excavated in front of the harvester. A laser rangefinder is installed at an adjacent position of the millimeter-wave radar. The millimeter-wave radar has a built-in tilt sensor. Based on the millimeter-wave radar, the laser rangefinder, and the tilt sensor, the relative height between the digging device of the harvester and the mud bottom surface is obtained. Positioning unit: used to install encoders to various rotating hinge points of the lifting device of the harvester, and to obtain the absolute height of the digging device and the frame of the harvester based on the encoders; Control unit: Used to construct a fuzzy rule table, and obtain control parameters based on the fuzzy rule table, the relative height, and the absolute height; Excavation unit: Used to drive the excavation device to excavate Alisma plantago-aquatica based on the control parameters.

6. The row correction system for harvesting Alisma plantago-aquatica according to claim 5, characterized in that, The depth unit is specifically used for: The first slant range and confidence score are obtained based on the millimeter-wave radar, the second slant range and signal-to-noise ratio are obtained based on the laser rangefinder, and the radar attitude data of the millimeter-wave radar are obtained based on the tilt sensor. If the signal-to-noise ratio is greater than the first threshold, then a measurement deviation is obtained based on the first slope distance and the second slope distance, a continuous deviation is obtained based on the measurement deviation, the first slope distance is compensated based on the continuous deviation to obtain a first calibrated slope distance, a third slope distance is obtained based on the first preset weight and the first calibrated slope distance, and a first distance is obtained based on the third slope distance. If the signal-to-noise ratio is less than the second threshold, then the first distance is obtained based on the first slant range; If the signal-to-noise ratio is greater than or equal to the second threshold and less than or equal to the first threshold, then a second preset weight is obtained based on the signal-to-noise ratio, and the first distance is obtained based on the second preset weight and the first calibration slope distance. The relative height is obtained by compensating for the first distance based on the radar attitude data; The first formula for calculating the relative height is: ； in, 1 indicates relative height. Indicates the first distance. Indicates the installation tilt angle of the millimeter-wave radar. This indicates the real-time elevation change angle of the millimeter-wave radar.

7. A row-correction system for harvesting Alisma plantago-aquatica according to claim 6, characterized in that, The position unit is specifically used for: The encoder is used to obtain the measured hinge angle value between two adjacent rigid bodies. A three-dimensional kinematic model of the harvester is constructed, and the first length between the rotational hinge points is obtained based on the three-dimensional kinematic model; The theoretical height is obtained based on the measured hinge angle value, the first length, and the three-dimensional kinematic model; An inertial measurement unit is installed at a key point on the frame of the harvester. The frame pitch angle is obtained based on the inertial measurement unit. The absolute height is obtained based on the frame pitch angle and the theoretical height. The second formula for obtaining the absolute height is: ； in, 2 indicates absolute height. This represents the initial height of the origin of the machine coordinate system. Indicates the rack pitch angle, This indicates a high level of theoretical understanding.

8. A row-correction system for harvesting Alisma plantago-aquatica according to claim 7, characterized in that, The position unit is also used for: A calibration sensor is installed on the piston rod of the hydraulic cylinder of the harvester, and the cylinder stroke value is obtained based on the calibration sensor. Based on the cylinder stroke value and the inverse kinematics algorithm, the inferred hinge angle value between two adjacent rigid bodies is obtained; The difference between the measured hinge angle value and the inferred hinge angle value is obtained, and an early warning result is obtained based on the difference and the tolerance range.

9. A row-correction system for harvesting Alisma plantago-aquatica according to claim 7, characterized in that, The control unit is specifically used for: Obtain a preset height, and based on the preset height, the relative height, and the absolute height, obtain the target height; The fuzzy rule table includes several fuzzy rules, each of which corresponds to a fuzzy set. Each fuzzy set includes several linguistic variables, and each linguistic variable corresponds to a single point value. Based on the target height and the absolute height, the error and the error rate are obtained. The error and the error rate are converted into fuzzy language values. Based on the fuzzy language values ​​and the fuzzy rule table, activation rules are obtained. Based on the activation rules, an activation fuzzy set is obtained. The first membership degree and the second membership degree of the error and the error rate in the activated fuzzy set are obtained respectively. The activation strength of the activation rule is obtained based on the first membership degree and the second membership degree. The effective output value of the activation rule is obtained based on the activation strength. Determine whether the activation intensity is greater than the intensity threshold. If so, calculate the corrected value of the language variable by weighting the effective output value and the activation intensity. Otherwise, obtain the corrected value of the language variable based on the default value. The control parameters are obtained based on the correction value and the initial parameters; The third calculation formula for obtaining the target height is: ； in, Indicates the target height. Indicates the preset height.

10. A row-correction system for harvesting Alisma plantago-aquatica according to claim 5, characterized in that, The mining unit is specifically used for: Map the control parameters to the target PWM duty cycle; Construct a duty cycle-target current curve table, and obtain the target current value based on the duty cycle-target current curve table and the target PWM duty cycle; The actual current value of the excavating device is obtained; based on the target current value and the actual current value, the current error is obtained; based on the current error, the target PWM duty cycle is adjusted to obtain the corrected PWM signal. The modified PWM signal is superimposed with the dizziness signal to obtain the PWM command; The excavation device is driven to excavate Alisma plantago-aquatica based on the PWM command.

Images