Automatic row control system and method of corn harvester

Abstract

The invention relates to an automatic row control system and method of a corn harvester, and relates to the technical field of row control, the automatic row control system of the corn harvester comprises a multi-mode sensing end used for obtaining corn row information and harvester state information according to an operation area; the fusion main control end is used for fusing and analyzing the corn row information and the harvester state information, identifying corn row characteristics and calculating comprehensive row alignment deviation; the path planning decision-making end is used for generating a path correction instruction according to the comprehensive row alignment deviation in combination with preset operation parameters; the instruction execution interaction end is used for controlling and displaying the running state of the harvester according to the path correction instruction; through multi-sensor fusion (including vision, ultrasonic wave, IMU, GNSS and header angle sensor) and an advanced control algorithm, millimeter-level accurate row alignment is realized, the phenomena of missing harvesting and seedling pressing are effectively reduced, and the quality and yield of corn harvesting are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of row control, in particular to a corn harvester automatic row alignment system and method. BACKGROUND

[0002] With the continuous improvement of agricultural mechanization and intelligence level, the market demand for automatic and precise operation of corn harvesters is increasingly urgent. Although there are some agricultural machinery automatic navigation systems based on GPS or visual recognition in the prior art, these systems usually have high cost, and their row alignment accuracy and stability still need to be improved in complex field environments, such as crop lodging, weed interference, and light changes, which are difficult to meet the specific needs of corn harvesting operations.

[0003] Application No. CN202010909618.5 discloses a corn row alignment harvesting control method and system, comprising: step 1, acquiring real-time vehicle body attitude information and corn row alignment information; step 2, determining whether the signal disappearance time of the row alignment information exceeds the preset time; step 3, if the result is yes, the vehicle body attitude information at the last time is retrieved, and the vehicle body angle is calculated and adjusted according to the vehicle body attitude information at the last time; step 4, controlling the vehicle to continue advancing according to the adjusted angle. The above method can save manual labor, and can also effectively avoid the problem of vehicle deviation from the preset track when encountering plant loss.

[0004] The prior art in the above has the following defects: 1. The row alignment accuracy of the existing corn harvester is difficult to guarantee in the case of night operation, low visibility or irregular land, and it is easy to cause missing collection and seedling compression, which reduces the harvesting efficiency; 2. Manual row alignment cannot achieve millimeter-level accurate control, which increases the harvesting loss and affects the quality of corn kernels. SUMMARY

[0005] In view of the deficiencies in the prior art, the purpose of the present application is achieved through the following technical solutions: A corn harvester automatic row alignment system, comprising: A multi-modal perception end for acquiring corn row information and harvester state information according to the working area; the harvester state information includes harvester attitude, absolute position, driving route, row alignment deviation angle and driving speed; A fusion main control end for fusion analysis of the corn row information and harvester state information, identification of corn row features, and calculation of comprehensive row alignment deviation; A path planning decision end for generating path correction instructions according to the comprehensive row alignment deviation combined with preset working parameters; An instruction execution interaction end for controlling and displaying the driving state of the harvester according to the path correction instructions.

[0006] By adopting the technical scheme, the corn row information (such as plant position and height) and the harvester state information (attitude, absolute position, row deviation angle and the like) of the working area are collected in real time by fusing a binocular camera, a laser radar and a mechanical sensor and the like equipment of a multi-modal sensing end, the multi-source data are subjected to feature recognition and deviation calculation by a Kalman filter or a dynamic fuzzy PID algorithm at a fusion main control end, the path correction instruction generated based on a pure tracking model or an LSTM prediction algorithm of a path planning decision end is combined, and finally the driving state is dynamically adjusted by controlling an electric steering wheel or a hydraulic actuating mechanism through an instruction execution interaction end and real-time feedback is performed, so that the row accuracy and the working stability of the corn harvester in a complex terrain and a dense planting shielding environment are significantly improved, the missed cutting and the crop damage are reduced, and the requirement for manual intervention is reduced.

[0007] The application is further provided that the multi-modal sensing end comprises: a visual acquisition module, configured to acquire the corn row information of the working area according to a preset compensation light source to obtain a corn row image; an inertial measurement module, configured to acquire the attitude of the harvester to obtain a pitch angle, a roll angle, a yaw angle and an angular velocity; a positioning and navigation module, configured to acquire the position information of the harvester to obtain an absolute position and a driving route; a row detection module, configured to detect the included angle between the header of the harvester and the corn row and calculate a row deviation angle; a speed monitoring module, configured to monitor the driving state of the harvester in the working area to obtain a driving speed.

[0008] By adopting the technical scheme, the corn row image is obtained by the visual acquisition module under the compensation light source, and the image processing algorithms such as Canny edge detection and Hough transformation are used to extract the corn row center line; at the same time, the inertial measurement module (IMU) provides the pitch, roll and yaw attitude information, the positioning and navigation module (such as RTK-GNSS) provides the high-precision absolute position and driving route, the row detection module calculates the row deviation angle of the header and the corn row, and the speed monitoring module provides the real-time driving speed; the multi-source heterogeneous data are fused by the system through the algorithm such as extended Kalman filter (EKF), the comprehensive row deviation is accurately calculated, and the path correction instruction is generated by using the adaptive PID or fuzzy control algorithm to control the turning of the harvester, so that the row accuracy, the robustness and the working stability of the system in a complex environment are significantly improved.

[0009] The application is further provided that the fusion main control end comprises: a data alignment module, configured to align the corn row information and the harvester state information according to a time stamp to obtain working time sequence data; The data cleaning module is configured to remove repeated data from the job operation time sequence data according to a preset data sampling frequency, to obtain non-repeated operation time sequence data. The anomaly detection module is configured to detect anomalies in the non-repeated operation time sequence data, mark abnormal data, and fit and fill, to obtain non-anomalous operation time sequence data. The data block module is configured to block the non-anomalous operation time sequence data according to a preset data processing period, to obtain a plurality of information blocks, and assign a unique identifier to each information block. The data classification module is configured to classify all the information blocks according to data types, to obtain image blocks and numerical value blocks. The image processing module is configured to calculate the sharpness of the corn row image in the image block, and compare the sharpness with a preset sharpness threshold; if the sharpness is greater than the sharpness threshold, it is determined that the current corn row image is compliant; otherwise, the current corn row image is filtered and denoised until the current corn row image is compliant. The edge detection module is configured to detect features of the compliant corn row image, mark the crop contour, and fit the crop row orientation angle. The feature recognition module is configured to calculate the corn row features according to the crop contour, to obtain the row spacing, crop stem, and tillage row direction. The deviation operation module is configured to calculate the lateral deviation and the heading deviation of the harvester relative to the corn row according to the crop row orientation angle and the tillage row direction. The twin construction module is configured to simulate each numerical value block to obtain a plurality of job time sequence graphs. The image analysis module is configured to decompose each image block to obtain a plurality of crop time state chains. The data aggregation module is configured to aggregate all the job time sequence graphs and all the crop time state chains according to time sequences, to obtain a job complete graph and a crop complete chain. The multi-state fusion module is configured to perform optical flow fitting on the job complete graph and the crop complete chain according to timestamps, to obtain a multi-state time sequence running graph, and extract a job frame time and a crop frame time. The time sequence supplement module is configured to match the job frame time and the crop frame time; if they are different, frame filling is performed on the larger frame time according to the smaller frame time until they are the same; if they are the same, it is determined that the current multi-state time sequence running graph is complete. The model generation module is configured to generate a job operation state model of the harvester according to the complete multi-state time sequence running graph. The deviation correction module is configured to correct the lateral deviation and the heading deviation according to the job operation state model, to obtain a comprehensive deviation of the harvester relative to the corn row.

[0010] By adopting the technical scheme, the corn row visual information is aligned with the multi-source state information of the harvester based on a timestamp, data cleaning and fitting filling are performed by using a sliding window and a Z-score-based anomaly detection algorithm, then the data is divided into blocks and processed in a classified manner: the image blocks are preprocessed by clarity calculation (such as a PSNR algorithm) and Gaussian filtering, then the crop contour and row direction angle are extracted by using Canny edge detection and Hough transformation; the numerical blocks are subjected to twin simulation and aggregation to construct a complete operation graph and a complete crop chain; finally, the spatio-temporal fusion of multi-state data is performed by using an optical flow method (such as a Farneback algorithm) to construct an operation state model, and the lateral and heading deviations calculated preliminarily are corrected based on the model to generate a high-precision comprehensive row deviation. Through the rigorous multi-level data cleaning and fusion mechanism, combined with the optical flow fitting algorithm, the spatio-temporal consistency is ensured, the complex working conditions such as light change, sensor noise and data loss can be effectively coped with, and thus a robust and high-precision operation model is constructed, which provides a reliable data basis for realizing high-precision automatic row control of the harvester.

[0011] The application further provides that the path planning decision end comprises: a parameter configuration module configured to configure operation parameters to obtain target row accuracy, maximum turning angle, attitude threshold and speed limit value; an environment monitoring module configured to collect weather data of an operation area to obtain temperature and humidity, wind direction and rainfall; a path planning module configured to generate an initial operation harvesting path according to boundary information and terrain data of the operation area in combination with a working constraint interval of the harvester; a path analysis module configured to decompose and configure the initial operation harvesting path according to the operation parameters to obtain a plurality of limiting path nodes; a decision fusion module configured to perform weight operation on all the limiting path nodes according to a row deviation angle to obtain node weights; an instruction generation module configured to perform local correction on the initial operation harvesting path according to the node weights in combination with a comprehensive row deviation, and generate a path correction instruction.

[0012] By adopting the technical scheme, the target row precision, maximum turning angle and environmental constraints (such as attitude threshold and speed limit) are set by the parameter configuration module, combined with weather data such as temperature, humidity, wind direction and rainfall collected by the environment monitoring module, the initial path is generated based on the A* or Dijkstra algorithm by the path planning module, and the path is decomposed into limit nodes by the path analysis module to meet the work constraints; then, the decision fusion module adopts weighted average or fuzzy logic algorithm, dynamically adjusts the node weight according to the row deviation angle, and finally generates path correction instructions based on the node weight and comprehensive row deviation by using local path correction algorithm (such as model predictive control MPC or dynamic window method DWA), so as to realize high-precision and high-robustness path planning based on multi-dimensional parameters and environmental adaptability, effectively improve the operation safety and efficiency of the harvester under complex working conditions.

[0013] The application is further provided that: the path planning module comprises: The analysis processing unit is used for analyzing the boundary information of the working area to obtain a geographic coordinate fence, reading the terrain data of the working area to obtain a height map and a slope map, and decomposing the working constraint interval of the harvester to obtain a turning radius, a header width and a vehicle size; The region mapping unit is used for offsetting the geographic coordinate fence according to the working constraint interval of the harvester to obtain an internal safety boundary, and combining the height map and the slope map to obtain a to-be-worked region; The global planning unit is used for planning a route for the to-be-worked region according to a preset working target parameter to obtain an initial advancing route; The optimization adaptation unit is used for direction adaptation and geometric correction of the initial advancing route according to agricultural requirements to obtain a plurality of corrected path segments; The path output unit is used for converting and structuring all the corrected path segments according to a preset path instruction format, and assigning corresponding working parameters to obtain an initial running harvesting path.

[0014] By adopting the technical scheme, the working boundary is analyzed, the terrain data is read and the vehicle constraint is decomposed by the analysis processing unit based on GIS technology, and the safe working area is constructed in the region mapping unit by using buffer analysis; the initial route is planned in the safe area by the global planning unit using A* algorithm, the path is smoothed and directionally corrected by the optimization adaptation unit in combination with the agricultural requirements by using spline interpolation, and finally the corrected path is structured into an initial running harvesting path containing working parameters by the path output unit, so as to realize high-precision and structured path planning considering safety, terrain adaptability and agricultural requirements, effectively ensuring efficient operation of the harvester in complex field environment.

[0015] The present invention is further configured such that the instruction execution interaction terminal includes: The instruction parsing module is used to interpret all path correction instructions to obtain check codes, signal identifier codes, and parameter ranges. The instruction verification module is used to verify the verification code according to the preset verification characters; if the two are the same, the current path correction instruction is determined to be compliant and transmitted to the instruction matching module. The instruction matching module is used to register the signal identification code according to the execution control port of the harvester; if the two match perfectly, the parameter range is transmitted to the execution control module. The execution control module is used to automatically align the working direction of the execution control mechanism according to the parameter range. The display and interaction module is used to display the harvester's row status, operating parameters, and fault prompts, as well as to switch modes and perform manual intervention.

[0016] By adopting the above technical solution, the instruction parsing module interprets the path correction instruction based on the CRC check code, and the instruction verification module verifies compliance. Then, the instruction matching module uses a hash table to quickly match the signal identification code with the execution port. The trigger execution control module uses a PID controller to parse the parameter range to achieve automatic row correction. At the same time, the display and interaction module provides real-time feedback of operation data through a state machine model and supports manual intervention. This forms a closed-loop control chain with strong fault tolerance, low latency response, and two-way human-machine collaboration, which significantly improves the execution accuracy and operational reliability of the harvester in complex farmland environments.

[0017] Secondly, the present invention also provides an automatic row alignment method for a corn harvester, which adopts the following technical solution: An automatic row alignment method for a corn harvester, applied to the aforementioned automatic row alignment system for a corn harvester, comprising: Obtain corn row information and harvester status information based on the work area; By integrating and analyzing the corn row information and harvester status information, corn row characteristics are identified, and comprehensive row deviation is calculated. Based on the boundary information and terrain data of the work area, combined with the working constraint range of the harvester, an initial operating harvesting path is generated; Based on the comprehensive row deviation and the preset operation parameters, the initial operation and harvest path is corrected, and a path correction instruction is generated. The path correction command is used to control and display the harvester's driving status.

[0018] By adopting the technical scheme, the 2D laser radar is used to collect corn ridge line point cloud data in real time, Kalman filtering is combined with IMU attitude information and GNSS positioning data, a corn row center line model based on a feature vector is constructed, then an A* algorithm is used to generate an initial path by fusing a farmland boundary constraint, and a PID controller is used to dynamically adjust a hydraulic steering system according to real-time row deviation (lateral error ±5cm, heading angle deviation ±3°), meanwhile, parameters such as working speed and header height and fault codes (such as blockage early warning Code 21) are fed back to an HMI touch screen through a CAN bus, so that millimeter-level rowing accuracy and working efficiency are synergistically optimized under all working conditions.

[0019] In a third aspect, the present application also provides an electronic device, adopting the following technical scheme: An electronic device comprises: One or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described above.

[0020] By adopting the technical scheme, the above energy management system based on multi-network fusion is presented in the form of computer readable code and stored in the memory, and when the processor runs the computer readable code in the memory, the steps of the automatic rowing method of the corn harvester are executed, so that the effect of reducing labor intensity and improving the degree of automatic operation is obtained.

[0021] In a fourth aspect, the present application also provides a computer storage medium, adopting the following technical scheme: A computer storage medium has a computer program stored thereon, and the program is executed by a processor to implement the method as described above.

[0022] To sum up, the beneficial technical effects of the present application are: 1. By means of multi-sensor fusion (including vision, ultrasonic, IMU, GNSS and header angle sensor) and advanced control algorithm, millimeter-level accurate rowing is realized, the phenomenon of missing collection and seedling pressing is effectively reduced, and the quality and yield of corn harvesting are improved; 2. The driver does not need to manually row for a long time, but only needs to monitor and intervene in an emergency through the display screen, which significantly reduces the work fatigue and improves the work comfort; especially in the night, bad weather or irregular land, the high-precision rowing can still be maintained, the effective working time is prolonged, and the overall harvesting efficiency is improved. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 is a structural schematic diagram of the automatic rowing system of the corn harvester according to one embodiment of the present application.

[0024] Figure 2 is the internal structure diagram of the fusion main control end of one embodiment of the application.

[0025] Figure 3 is the internal structure diagram of the path planning module of one embodiment of the application.

[0026] Figure 4 is the flowchart of the automatic alignment method of the corn harvester of one embodiment of the application. DETAILED DESCRIPTION

[0027] The application will be further described in detail below with reference to the accompanying drawings.

[0028] Embodiment one: Reference Figure 1 , the application discloses a corn harvester automatic alignment system, comprising: A multi-modal perception end is used to obtain corn row information and harvester state information according to a working area; the harvester state information includes a harvester posture, an absolute position, a driving route, an alignment deviation angle and a driving speed; A fusion main control end is used to fuse and analyze the corn row information and the harvester state information, identify corn row features, and calculate a comprehensive alignment deviation; A path planning and decision end is used to generate a path correction instruction according to the comprehensive alignment deviation combined with preset working parameters; An instruction execution and interaction end is used to control and display the driving state of the harvester according to the path correction instruction.

[0029] In this embodiment, when the corn harvester enters the working area, various sensors (vision, IMU, positioning antenna receiver, header alignment angle sensor, vehicle speed sensor) in the multi-modal perception module start to work, and real-time collection of corn row images, lateral distance, header angle, harvester posture and absolute position information and driving speed is performed. These data are transmitted to the core main controller, and through image recognition, multi-sensor data fusion algorithm, the lateral deviation and the heading deviation of the harvester relative to the corn row are accurately calculated. Subsequently, the path planning and decision module generates real-time steering control instructions according to these deviations, and sends them to the execution control module. The steering motor in the execution control module drives the steering wheel of the harvester to fine-tune, so that the harvester realizes automatic and accurate alignment to the corn row for harvesting operation. During the whole operation process, the system continuously performs perception, calculation, decision and execution, and provides real-time feedback to the driver through a 10-inch screen, forms a closed loop control, and ensures that the harvester always maintains the best alignment state.

[0030] The implementation principle of the embodiment is that when the corn harvester enters the working area, the multi-modal perception end collects corn row images, harvester attitude, absolute position, row angle and driving speed and other multi-source heterogeneous information in real time through a visual sensor, an IMU, a GNSS receiver and various vehicle body sensors; the information is transmitted to the fusion main control end, the end adopts multi-sensor data fusion algorithms such as image recognition and Kalman filtering to accurately calculate the comprehensive deviation of the harvester relative to the corn row; then, the path planning decision end generates real-time path correction instructions according to the deviation and the preset working parameters; finally, the instruction execution interaction end receives the instructions, drives the steering motor to adjust the steering wheel, realizes automatic row control, and feeds back the real-time state to the driver through the display screen, thereby forming a closed-loop control system of “perception-fusion-decision-execution-feedback”, and ensuring that the harvester maintains high-precision row operation under complex working conditions.

[0031] Embodiment two: The multi-modal perception end comprises: a visual acquisition module, configured to collect corn row information of the working area according to a preset compensation light source, and obtain a corn row image; an inertial measurement module, configured to collect a harvester attitude, and obtain a pitch angle, a roll angle, a yaw angle and an angular velocity; a positioning and navigation module, configured to collect position information of the harvester, and obtain an absolute position and a driving route; a row detection module, configured to detect an included angle between a header of the harvester and a corn row, and calculate a row deviation angle; a speed monitoring module, configured to monitor a driving state of the harvester in the working area, and obtain a driving speed.

[0032] The implementation principle of the embodiment is that the multi-modal perception end collects corn row images through the visual acquisition module (based on the compensation light source), collects a pitch angle, a roll angle, a yaw angle and an angular velocity through the inertial measurement module, obtains an absolute position and a driving route through the positioning and navigation module, calculates a deviation angle between the header and the corn row through the row detection module, and monitors a real-time driving speed through the speed monitoring module; multi-source heterogeneous data are processed by the main controller, the image recognition algorithm is used to analyze the characteristics of the corn row, the data fusion algorithm such as Kalman filtering is combined, the attitude, position, row angle and speed information are integrated, the accurate row deviation angle and lateral deviation are generated, and reliable basis is provided for subsequent path correction and automatic row control, thereby supporting high-precision and adaptive operation of the harvester in a complex farmland environment.

[0033] Embodiment three: With reference to Figure 2 , the fusion main control end comprises: a data alignment module configured to align corn row information and harvester state information according to timestamps to obtain job runtime sequence data; a data cleaning module configured to remove duplicate data from the job runtime sequence data according to a preset data sampling frequency to obtain non-duplicate runtime sequence data; an anomaly detection module configured to detect anomalies in the non-duplicate runtime sequence data, mark abnormal data, and fit and fill to obtain non-anomalous runtime sequence data; a data block module configured to divide the non-anomalous runtime sequence data into a plurality of information blocks according to a preset data processing period and assign unique identifiers to the information blocks; a data classification module configured to classify all the information blocks according to data types to obtain image blocks and numerical value blocks; an image processing module configured to calculate the sharpness of corn row images in the image blocks and compare the sharpness with a preset sharpness threshold; if the sharpness is greater than the sharpness threshold, it is determined that the current corn row image is compliant; otherwise, the current corn row image is filtered and denoised until the current corn row image is compliant; an edge detection module configured to detect features of the compliant corn row images, mark crop contours, and fit crop row orientation angles; a feature recognition module configured to calculate corn row features, including row spacing, crop stems, and tillage row direction, according to the crop contours; a deviation calculation module configured to calculate lateral deviation and heading deviation of a harvester relative to a corn row according to the crop row orientation angle and the tillage row direction; a twin construction module configured to simulate each numerical value block to obtain a plurality of job sequence graphs; an image analysis module configured to decompose each image block to obtain a plurality of crop temporal chains; a data aggregation module configured to aggregate all the job sequence graphs and all the crop temporal chains according to time sequences to obtain a job complete graph and a crop complete chain; a multi-state fusion module configured to fit optical flow of the job complete graph and the crop complete chain according to timestamps to obtain a multi-state sequence running graph and extract job frame times and crop frame times; a time sequence supplement module configured to match the job frame times and the crop frame times; if the two are different, frame filling is performed on the larger frame time according to the smaller frame time until the two are the same; if the two are the same, it is determined that the current multi-state sequence running graph is complete; a model generation module configured to generate a job running state model of a harvester according to the complete multi-state sequence running graph; A deviation correction module is configured to correct a lateral deviation and a heading deviation according to the operation state model of the harvesting machine to obtain a comprehensive deviation of the harvesting machine relative to the corn rows.

[0034] The implementation principle of the embodiment is as follows: the main control end fuses the corn row information obtained by the multi-modal perception end and the state information of the harvesting machine. First, the time stamp alignment module is used to realize the time synchronization of heterogeneous data. After data cleaning (de-duplication), anomaly detection (labeling and fitting filling) processing, the data are combined with a preset period block and classified into image blocks and numerical blocks. After the image blocks are subjected to sharpness verification and filtering denoising, the edge detection algorithm is used to extract the crop profile and row orientation angle and calculate the row spacing and other features. The numerical blocks are used to generate an operation time sequence diagram through twin simulation. The image blocks are decomposed into a crop time chain. The data are aggregated to form a complete operation graph and a complete crop chain. The multi-state fusion module is used to align the frames by using an optical flow fitting algorithm and to supplement the frames. Finally, the operation state model is constructed. In combination with the lateral and heading deviations, the comprehensive deviation is generated by using a correction algorithm. The automatic row control provides a high-precision, low-latency and robust decision basis.

[0035] Embodiment Four The path planning decision end comprises: A parameter configuration module is configured to configure operation parameters to obtain a target row alignment accuracy, a maximum turning angle, an attitude threshold and a speed limit value. An environment monitoring module is configured to collect weather data of an operation area to obtain temperature and humidity, wind direction and rainfall. A path planning module is configured to generate an initial operation harvesting path according to boundary information and terrain data of the operation area in combination with a working constraint interval of the harvesting machine. A path analysis module is configured to decompose and configure the initial operation harvesting path according to the operation parameters to obtain a plurality of limit path nodes. A decision fusion module is configured to perform weight operation on all the limit path nodes according to a row deviation angle to obtain a node weight. An instruction generation module is configured to perform local correction on the initial operation harvesting path according to the node weight in combination with the comprehensive row deviation and to generate a path correction instruction.

[0036] The implementation principle of the embodiment is that: the path planning decision end sets the target precision, maximum turning angle and other constraint conditions through the parameter configuration module, combines the real-time weather data such as temperature, humidity, wind direction and rainfall collected by the environment monitoring module, and generates an initial path and decomposes it into limit nodes by using A or Dijkstra algorithm (based on boundary and terrain information); then, the decision fusion module dynamically adjusts the weight of each node based on the walking deviation angle, the instruction generation module combines the node weight and the comprehensive walking deviation (such as lateral deviation ±5cm, heading deviation ±3°) output by the fusion main control end, generates path correction instructions through local path correction algorithm (such as DLite), realizes accurate control of the harvester, and through multi-dimensional parameter configuration and environmental adaptive adjustment, ensures the real-time performance, robustness and operation safety under complex working conditions of the path planning.

[0037] Embodiment five: With reference to Figure 3 The path planning module comprises: An analysis processing unit is configured to analyze boundary information of a working area to obtain a geographic coordinate fence, read terrain data of the working area to obtain a height map and a slope map, and decompose a working constraint interval of the harvester to obtain a turning radius, a header width and a vehicle size. A region mapping unit is configured to offset the geographic coordinate fence according to the working constraint interval of the harvester to obtain an internal safety boundary, and correct the internal safety boundary in combination with the height map and the slope map to obtain a to-be-worked region. A global planning unit is configured to plan a route for the to-be-worked region according to a preset working target parameter to obtain an initial traveling route. An optimization adaptation unit is configured to adapt a direction and correct a geometry of the initial traveling route according to an agronomic requirement to obtain a plurality of corrected path segments. A path output unit is configured to convert and structure all the corrected path segments according to a preset path instruction format, and assign corresponding working parameters to obtain an initial running harvesting path.

[0038] The implementation principle of the embodiment is that the path planning module extracts the geographic coordinate fence, the terrain height map and the slope map of the working area through the analysis processing unit, and generates the internal safety boundary and the to-be-worked area in combination with the physical constraints such as the turning radius and the header width of the harvester by the area mapping unit; the global planning unit performs initial path search based on preset working target parameters such as the advancing direction and the coverage range, and the optimization adaptation unit performs geometric correction and direction adaptation on the path according to the agronomic requirements such as the crop row direction and the row spacing, to form a corrected path segment conforming to the working specification; finally, the path output unit structures each corrected path segment according to a preset instruction format, and associates the working parameters such as the speed and the header height, to generate an initial running harvesting path; through multi-constraint fusion (terrain, equipment, and agronomy) and algorithm optimization (such as A* or improved graph search algorithm), the global rationality and the local adaptability of the path planning are ensured, and the working efficiency and the safety are improved.

[0039] Embodiment six The instruction execution interaction end comprises: An instruction analysis module is configured to interpret all path correction instructions to obtain a check code, a signal identification code, and a parameter interval segment. An instruction check module is configured to verify the check code according to a preset check character; if the check code and the check character are the same, it is determined that the current path correction instruction is compliant, and the instruction is transmitted to an instruction matching module. The instruction matching module is configured to match the signal identification code with an execution control port of the harvester; if the two are completely matched, the parameter interval segment is transmitted to an execution control module. The execution control module is configured to automatically align the working direction of the execution control mechanism according to the parameter interval segment. A display interaction module is configured to display the alignment state, the working parameters, and the fault prompts of the harvester, and to switch modes and manually intervene.

[0040] The implementation principle of the embodiment is that the instruction execution interaction end interprets the path correction instruction through the instruction analysis module to extract the check code, the signal identification code, and the parameter interval segment; the instruction check module verifies the consistency of the check code and the preset character, and if the instruction is compliant, it is transmitted to the instruction matching module; the instruction matching module accurately matches the signal identification code with the execution control port of the harvester, and after successful matching, the parameter interval segment is transmitted to the execution control module to drive the execution mechanism to automatically align; at the same time, the display interaction module displays the alignment state, the working parameters, and the fault prompts in real time, and supports mode switching and manual intervention, forming a closed-loop interaction process of “instruction analysis-check-matching-execution-feedback”, to ensure the reliability of system response and the controllability of work.

[0041] Embodiment seven Reference Figure 4An automatic row alignment method for a corn harvester, applied to the automatic row alignment system of the corn harvester, comprising: S1: Obtain corn row information and harvester status information based on the work area; S2: Integrate and analyze the corn row information and harvester status information to identify corn row characteristics and calculate the overall row deviation; S3: Generate an initial harvesting path based on the boundary information and terrain data of the work area and the working constraint range of the harvester; S4: Based on the comprehensive row deviation and the preset operation parameters, correct the initial operation and harvest path, and generate a path correction instruction; S5: Control and display the driving status of the harvester according to the path correction command.

[0042] The implementation principle of this embodiment is as follows: Information on corn rows and harvester status in the work area is acquired through multimodal sensing devices (such as lidar, mechanical row alignment sensors, etc.). Then, the two are fused and analyzed to identify corn row characteristics (such as row spacing and crop direction) and calculate the comprehensive alignment deviation 1 in the lateral and directional directions. An initial path is generated based on boundary and terrain data and harvester physical constraints (such as turning radius and header width). The path is then corrected through dynamic weighting calculations, combining the comprehensive alignment deviation with preset operating parameters (such as target alignment accuracy and maximum turning angle), generating precise path correction instructions. Finally, the harvester is automatically aligned with the rows and the operating status is displayed in real time through the instruction execution interaction terminal (including the execution control module and the display interaction module), forming a closed-loop control chain of "perception-analysis-planning-correction-execution," thereby improving operational efficiency and reliability.

[0043] An electronic device, comprising: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method described above.

[0044] A computer storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described above.

[0045] The embodiments described herein are preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape, and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. An automatic row alignment system for a corn harvester, comprising: The multimodal sensing end is used to acquire corn row information and harvester status information based on the working area; the harvester status information includes harvester attitude, absolute position, travel route, row deviation angle, and travel speed. The main control terminal is used to integrate and analyze the corn row information and harvester status information, identify corn row characteristics, and calculate the overall row deviation. The path planning decision-making end is used to generate path correction instructions based on the comprehensive deviation of the line and preset operation parameters; The instruction execution interface is used to correct instruction control and display the driving status of the harvester according to the path.

2. The automatic row alignment system for corn harvesters according to claim 1, characterized in that, The multimodal sensing end includes: The visual acquisition module is used to acquire corn row information in the work area according to a preset compensation light source, and obtain corn row images; The inertial measurement module is used to collect the attitude of the harvester, obtaining pitch angle, roll angle, yaw angle and angular velocity; The positioning and navigation module is used to collect the location information of the harvester to obtain its absolute position and travel route; The row alignment detection module is used to detect the angle between the harvester's header and the corn row, and to calculate the row alignment deviation angle. The speed monitoring module is used to monitor the driving status of the harvester in the working area and obtain the driving speed.

3. The automatic row alignment system for corn harvesters according to claim 1, characterized in that, The fusion master control terminal includes: The data alignment module is used to align corn row information and harvester status information according to timestamps to obtain operation runtime sequence data; The data cleaning module is used to remove duplicate data from the job execution sequence data according to a preset data sampling frequency to obtain duplicate-free execution sequence data. Anomaly detection module is used to detect anomalies in the non-repeating runtime sequence data, mark abnormal data and fit and fill them to obtain non-anomaly runtime sequence data; The data segmentation module is used to segment the non-different runtime sequence data into blocks according to a preset data processing cycle, obtain several information blocks, and assign them unique identifiers. The data classification module is used to classify all the information blocks according to their data types to obtain image blocks and numerical blocks; The image processing module is used to calculate the sharpness of the corn row image within the image block and compare it with a preset sharpness threshold; if the sharpness is greater than the sharpness threshold, the current corn row image is determined to be compliant; otherwise, the current corn row image is filtered and denoised until the current corn row image is compliant. An edge detection module is used to perform feature detection on compliant corn row images, mark crop outlines, and fit crop row orientation angles; The feature recognition module is used to calculate the features of the corn rows based on the crop outline to obtain the row spacing, crop stalks and tillage row direction; The deviation calculation module is used to calculate the lateral deviation and heading deviation of the harvester relative to the corn row based on the crop row orientation angle and the tillage row direction.

4. The automatic row alignment system for corn harvesters according to claim 3, characterized in that: The fusion main control terminal also includes: The twin construction module is used to perform twin simulations on each numerical block to obtain several job timing diagrams; The image parsing module is used to decompose each image block to obtain several crop temporal chains; The data aggregation module is used to aggregate all the operation time series diagrams and all the crop temporal chains according to the time series to obtain the complete operation diagram and the complete crop chain. The multi-state fusion module is used to perform optical flow fitting on the complete job graph and the complete crop chain based on the timestamp to obtain a multi-state time-series running graph, and extract the job frame time and crop frame time; The timing supplementation module is used to match and determine the timing of the job frame and the timing of the crop frame; if they are different, the larger frame timing is padded with the smaller frame timing until they are the same; if they are the same, the current polymorphic timing operation graph is determined to be complete. The model generation module is used to generate a harvester operation status model based on the complete polymorphic time-series operation diagram. The deviation correction module is used to perform correction calculations on the lateral deviation and heading deviation according to the operation status model to obtain the comprehensive row deviation of the harvester relative to the corn row.

5. The automatic row alignment system for corn harvesters according to claim 1, characterized in that: The path planning decision-making terminal includes: The parameter configuration module is used to configure the operation parameters, and obtain the target alignment accuracy, maximum steering angle, attitude threshold and speed limit; The environmental monitoring module is used to collect weather data in the work area, obtaining temperature, humidity, wind direction, and rainfall. The path planning module is used to generate an initial harvesting path based on the boundary information and terrain data of the work area and the working constraint range of the harvester. The path parsing module is used to decompose and configure the initial running harvest path according to the operation parameters to obtain several limit path nodes; The decision fusion module is used to perform weight calculations on all the limit path nodes based on the row deviation angle to obtain the weight of each node. The instruction generation module is used to locally correct the initial running harvest path based on the node weights and the overall row deviation, and generate path correction instructions.

6. The automatic row alignment system for corn harvesters according to claim 5, characterized in that: The path planning module includes: The parsing and processing unit is used to parse the boundary information of the work area to obtain the geographic coordinate fence, and to read the terrain data of the work area to obtain the height map and slope map. At the same time, it decomposes the working constraint range of the harvester to obtain the turning radius, header width and vehicle size. The area mapping unit is used to offset the geographic coordinate fence according to the working constraint range of the harvester to obtain the internal safety boundary, and to correct it by combining the height map and the slope map to obtain the area to be operated; The global planning unit is used to plan the route for the area to be worked based on preset task target parameters to obtain an initial travel route. An optimization and adaptation unit is used to perform directional adaptation and geometric correction on the initial travel route according to agronomic requirements, resulting in several corrected path segments. The path output unit is used to convert and structure all the modified path segments according to the preset path instruction format, and assign corresponding operation parameters to obtain the generated initial running harvest path.

7. The automatic row alignment system for corn harvesters according to claim 1, characterized in that: The instruction execution interaction terminal includes: The instruction parsing module is used to interpret all path correction instructions to obtain check codes, signal identifier codes, and parameter ranges. The instruction verification module is used to verify the verification code according to the preset verification characters; if the two are the same, the current path correction instruction is determined to be compliant and transmitted to the instruction matching module. The instruction matching module is used to register the signal identification code according to the execution control port of the harvester; if the two match perfectly, the parameter range is transmitted to the execution control module. The execution control module is used to automatically align the working direction of the execution control mechanism according to the parameter range. The display and interaction module is used to display the harvester's row status, operating parameters, and fault prompts, as well as to switch modes and perform manual intervention.

8. An automatic row alignment method for a corn harvester, characterized in that, include: Obtain corn row information and harvester status information based on the work area; By integrating and analyzing the corn row information and harvester status information, corn row characteristics are identified, and comprehensive row deviation is calculated. Based on the boundary information and terrain data of the work area, combined with the working constraint range of the harvester, an initial operating harvesting path is generated; Based on the comprehensive row deviation and the preset operation parameters, the initial operation and harvest path is corrected, and a path correction instruction is generated. The path correction command is used to control and display the harvester's driving status.

9. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors perform the method as described in claim 8.

10. A storage medium storing at least one instruction, at least one program, a code set, or an instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the automatic row alignment method as claimed in claim 8.

Images