An automated corn harvesting system and method

Abstract

The present application relates to the technical field of corn harvesting, and discloses a corn automatic harvesting system and method. The method counts a plurality of monitoring subareas of a corn planting area, deploys an image acquisition device, acquires growth characteristic data such as corn cob maturity distribution, stalk strength and leaf coverage density of each subarea, acquires real-time warehouse capacity data and conveying equipment operation state data of a warehouse scheduling center, determines initial harvesting parameters including harvesting mechanical advancing speed and corn cob stripping strength, judges whether corn cob maturity distribution data triggers a harvesting parameter adjustment condition, acquires plant image data of all subareas and extracts features to obtain plant morphological characteristic data when the condition is triggered, calculates subarea harvesting priority characteristic values according to the plant morphological characteristic data, matches the subarea harvesting priority characteristic values with a historical database, adjusts the initial harvesting parameters to obtain dynamic harvesting parameters according to a matching result, and calculates a similarity data set and corrects the dynamic harvesting parameters based on the similarity data set if there is no matching value in the historical database.

Description

TECHNICAL FIELD

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

[0002] In the field of agricultural production, corn as an important food crop, the efficiency and quality of its harvesting link have a direct impact on the overall agricultural production efficiency. With the continuous development of agricultural mechanization, corn harvesting gradually changes from manual harvesting to mechanical harvesting, but there are still many problems to be solved in the current corn mechanical harvesting process.

[0003] The traditional corn harvesting method is often unified according to fixed time nodes or preset harvesting parameters, and the differences in the growth conditions of different sub-zones of corn planting areas are not fully considered. Due to the differences in soil fertility, light conditions, water supply and other factors of corn planting areas, the maturity, stem strength and leaf coverage density of corn cobs in different monitoring sub-zones will show obvious different distribution characteristics. If unified harvesting parameters are used for harvesting operation, for the corn sub-zone with higher maturity, it may cause corn cobs to fall off, mold and other problems due to untimely harvesting, affecting the yield and quality of corn; while for the corn sub-zone with lower maturity, premature harvesting will make the corn kernel moisture content too high, which is not conducive to subsequent storage and processing, and may also cause stem breakage during harvesting process due to insufficient stem strength, increasing the risk of mechanical failure.

[0004] The current corn harvesting operation lacks effective linkage with the storage scheduling and the running state of the conveying equipment in terms of parameter adjustment. The real-time storage capacity data of the storage scheduling center is directly related to the temporary storage capacity of the harvested corn. If the storage capacity is insufficient, the operation will still be carried out according to the original harvesting rate, which will cause the harvested corn to be unable to be stored in time and can only be piled up in the field, which is easy to be damaged by external environmental factors; the running state of the conveying equipment affects the transfer efficiency of corn from the harvesting machine to the storage facility. If the conveying equipment fails or the running efficiency is reduced, and the harvesting machine still maintains the original harvesting rate, it will cause the accumulation of corn in the conveying link, which not only affects the continuity of the harvesting operation, but also may cause further damage to the conveying equipment.

[0005] In the face of different corn growth characteristic data, the prior art is difficult to accurately determine whether to adjust the harvesting parameters, and lacks a scientific and reasonable method to determine the harvesting priority of each monitoring partition. When the corn growth characteristic data changes, it is not possible to generate corresponding dynamic harvesting parameters in a timely manner based on these changes, resulting in the harvesting operation always being in a state of passive adjustment, making it difficult to adapt to complex and changing corn growth conditions and actual production needs. Moreover, in the absence of historical data support or matching in historical data, the harvesting parameters cannot be corrected through effective data processing methods, further reducing the accuracy and adaptability of the harvesting parameters, and restricting the intelligent and precise development of corn harvesting operations. SUMMARY

[0006] The purpose of the present application is to provide a corn automatic harvesting method based on corn, to solve the problems raised in the background art.

[0007] To achieve the above purpose, the present application provides a corn automatic harvesting method based on corn, the method comprising:

[0008] Statistical corn planting area of a plurality of monitoring partitions and deploy image acquisition device, obtain the corn growth characteristic data of each monitoring partition; the corn growth characteristic data includes corn cob maturity distribution data, stem strength data and leaf coverage density data;

[0009] Collect real-time warehouse capacity data and conveying equipment running state data of the warehouse scheduling center, and determine the initial harvesting parameters according to the real-time warehouse capacity data and conveying equipment running state data; the initial harvesting parameters include harvesting machinery speed parameter and corn cob stripping strength parameter;

[0010] Determine whether the corn cob maturity distribution data triggers the harvesting parameter adjustment condition, when triggered, collect the corn plant image data of all monitoring partitions, extract the plant morphological feature data from the corn plant image data, and calculate the harvesting priority feature value of each monitoring partition according to the plant morphological feature data;

[0011] Match the harvesting priority feature value of each monitoring partition with the historical harvesting priority feature value database, and adjust the initial harvesting parameters according to the matching result to obtain dynamic harvesting parameters;

[0012] When there is no matching harvesting priority feature value in the historical harvesting priority feature value database, calculate the similarity data set of the harvesting priority feature value of each monitoring partition and the historical harvesting priority feature value, and correct the dynamic harvesting parameters based on the similarity data set.

[0013] Preferably, when determining the initial harvesting parameters according to the real-time warehouse capacity data and conveying equipment running state data, it comprises:

[0014] characteristic values of the volume utilization are obtained by feature extraction from the real-time volume data, and characteristic values of the equipment load are obtained by feature extraction from the running state data of the conveying equipment;

[0015] The characteristic values of the volume utilization are compared with volume utilization characteristic threshold values, and the characteristic values of the equipment load are compared with equipment load characteristic threshold values;

[0016] A basic value of the travel speed parameter of the harvesting machine is determined according to the comparison result, and a reference value of the corn cob stripping strength parameter is determined according to the average value of the corn cob maturity distribution data.

[0017] Preferably, when determining whether the corn cob maturity distribution data triggers a harvesting parameter adjustment condition, the following steps are included:

[0018] A maturity fluctuation characteristic value of the corn cob maturity distribution data is extracted;

[0019] A standard maturity fluctuation characteristic value is obtained, and a maturity characteristic deviation amount of the maturity fluctuation characteristic value from the standard maturity fluctuation characteristic value is calculated;

[0020] The maturity characteristic deviation amount is compared with a maturity characteristic deviation threshold value, and whether the harvesting parameter adjustment condition is triggered is determined according to the comparison result.

[0021] Preferably, when the corn plant image data is subjected to feature extraction to obtain plant morphology characteristic data, the following steps are included:

[0022] A corn plant height distribution data and an ear distribution density data are extracted by using a contour recognition algorithm;

[0023] The harvesting priority characteristic value is obtained by a weighted calculation result of a discrete coefficient of the corn plant height distribution data and the ear distribution density data.

[0024] Preferably, when the harvesting priority characteristic value of each monitoring subzone is matched with a historical harvesting priority characteristic value database, the following steps are included:

[0025] When there is a single matched historical harvesting priority characteristic value in the historical harvesting priority characteristic value database, a first correction coefficient is used to adjust the dynamic harvesting parameter;

[0026] When there are multiple matched historical harvesting priority characteristic values in the historical harvesting priority characteristic value database, an average value of the matched historical harvesting priority characteristic values is calculated, and a second correction coefficient is used to adjust the dynamic harvesting parameter;

[0027] When there is no matched historical harvesting priority characteristic value in the historical harvesting priority characteristic value database, a similarity data set calculation process is activated.

[0028] Preferably, when the dynamic harvesting parameter is adjusted by the second correction coefficient, the method comprises:

[0029] A second correction coefficient gradient is set, which comprises a first correction coefficient, a second correction coefficient and a third correction coefficient;

[0030] The average value of the matching historical harvesting priority feature values is compared with a first feature average value threshold and a second feature average value threshold;

[0031] When the average value is less than or equal to the first feature average value threshold, the first correction coefficient is used to update the dynamic harvesting parameter;

[0032] When the average value is greater than the first feature average value threshold and less than or equal to the second feature average value threshold, the second correction coefficient is used to update the dynamic harvesting parameter.

[0033] Preferably, when the dynamic harvesting parameter is adjusted based on the similarity dataset, the method comprises:

[0034] An average similarity value of the similarity dataset is calculated;

[0035] The number of harvesting priority feature values higher than the average similarity value is counted as a first similar number, and the number of harvesting priority feature values lower than or equal to the average similarity value is counted as a second similar number;

[0036] A third correction coefficient is calculated according to the proportional relationship between the first similar number and the second similar number;

[0037] The third correction coefficient is used to update the dynamic harvesting parameter.

[0038] Preferably, the method further comprises:

[0039] Real-time monitoring of the working state data of the harvesting machine, the working state data comprising machine vibration spectrum data and grain breakage rate data;

[0040] When the grain breakage rate data exceeds a breakage rate threshold, the corn plant image data of the current working area is re-collected;

[0041] The harvesting priority feature value is updated based on the newly collected corn plant image data, triggering the readjustment process of the dynamic harvesting parameter.

[0042] Preferably, the method further comprises:

[0043] Real-time storage pressure data of each storage unit is obtained, the real-time storage pressure data comprising storage space utilization rate and conveying belt carrying efficiency;

[0044] redistribute the harvesting task order of each monitoring partition according to the updated dynamic harvesting parameter;

[0045] generate the warehouse scheduling instruction according to the redistributed harvesting task order.

[0046] Preferably, the present application also includes an automatic corn harvesting system based on corn, comprising:

[0047] The partition monitoring module is used for counting a plurality of monitoring partitions of the corn planting area and deploying an image acquisition device to obtain corn growth characteristic data of each monitoring partition.

[0048] The parameter initialization module is used for collecting real-time warehouse capacity data and conveying equipment running state data of the warehouse scheduling center to determine the initial harvesting parameter.

[0049] The dynamic adjustment module is used for judging whether the corn cob maturity distribution data triggers the adjustment condition, collecting corn plant image data and calculating the harvesting priority characteristic value of each monitoring partition, and adjusting the initial harvesting parameter according to the matching result of the historical harvesting priority characteristic value database.

[0050] The parameter optimization module is used for calculating the similarity data set and correcting the dynamic harvesting parameter when there is no matching value in the historical database.

[0051] The mechanical control module is used for executing the updated dynamic harvesting parameter and controlling the harvesting mechanical operation.

[0052] The warehouse scheduling module is used for generating the warehouse scheduling instruction according to the real-time storage pressure data and the harvesting task order.

[0053] Compared with the prior art, the present application has the following beneficial effects:

[0054] The present application based on the automatic corn harvesting method can fully combine the growth characteristic differences of different monitoring partitions of the corn planting area, realize the precise regulation and control of the harvesting operation, and effectively improve the problem of improper harvesting time caused by uniform harvesting parameters in the traditional harvesting method. By deploying the image acquisition device to obtain the corn growth characteristic data of each monitoring partition, including the corn cob maturity distribution data, the stem strength data and the leaf coverage density data, the harvesting operation can be carried out based on the actual corn growth conditions, avoiding the loss of corn yield and quality caused by ignoring the partition differences. The corn growth characteristic data of different monitoring partitions can provide detailed field information for the harvesting operation, making the harvesting decision more targeted. Whether the partition with higher maturity needs timely harvesting or the partition with lower maturity needs delayed harvesting, it can be processed according to its growth conditions, ensuring the quality and yield of corn in the harvesting link.

[0055] The method takes the real-time warehouse capacity data and conveying equipment running state data of the warehouse scheduling center into the determination process of the initial harvesting parameter, realizing the coordinated linkage of the harvesting operation and the warehouse and conveying links. The real-time warehouse capacity data can reflect the current storage capacity of the warehouse facility, and adjusting the harvesting mechanical travel speed parameter according to the data can avoid the situation that the harvested corn cannot be stored in time due to insufficient warehouse capacity and is damaged; the conveying equipment running state data is directly related to the corn transfer efficiency, and combining the data to determine the corn cob stripping strength parameter can ensure that the harvested corn is smoothly transferred in the conveying link, reduce the operation interruption or corn backlog caused by the running problem of the conveying equipment, protect the continuity and stability of the whole harvesting process, and improve the overall production efficiency.

[0056] In terms of harvesting parameter adjustment, whether the corn cob maturity distribution data triggers the adjustment condition is judged, then corn plant image data is collected and a harvesting priority feature value is extracted, which provides a scientific basis for the orderly development of the harvesting operation. The harvesting priority feature value can clearly determine the harvesting sequence of each monitoring partition, so that the harvesting resources can be preferentially allocated to the partitions that meet the harvesting requirements or urgently need harvesting, avoiding the disorder of the harvesting operation, reasonably utilizing the harvesting mechanical resources, reducing unnecessary waste of operation time, and further improving the efficiency of the harvesting operation.

[0057] Matching the harvesting priority feature value of each monitoring partition with the historical harvesting priority feature value database to adjust the initial harvesting parameter can fully utilize the experience advantage of the historical data, making the determination of the dynamic harvesting parameter more reliable and reasonable. The historical data contains effective parameter information of corn harvesting under different growth conditions, and by matching the historical data, the harvesting parameter suitable for the current corn growth condition can be quickly generated, reducing the blindness in the parameter determination process. When there is no matching data in the historical database, the method of correcting the dynamic harvesting parameter through the calculation of the similarity data set effectively solves the problem of incomplete coverage of historical data, ensuring that accurate harvesting parameters can be generated based on similar historical data in the absence of direct matching data, improving the adaptability of the method to different field conditions, and enhancing the flexibility and intelligent level of the harvesting operation. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 A working principle diagram of the corn automatic harvesting method is described in the present application;

[0059] Figure 2 A sub-process flow chart for determining the initial harvesting parameter is described;

[0060] Figure 3 A sub-process flow chart for extracting plant morphological features and calculating a harvesting priority feature value is described;

[0061] Figure 4 To monitor the harvesting machine operation state and readjust the dynamic harvesting parameter sub-process flow chart. DETAILED DESCRIPTION

[0062] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0063] Please refer to Figure 1 The present application provides an automatic corn harvesting system and method, which comprises:

[0064] By integrating multi-source data sensing and intelligent decision mechanism, dynamic optimization control of corn harvesting process is realized. The corn planting area is divided into several monitoring partitions, and high-definition image acquisition devices are deployed in each partition to continuously obtain corn cob maturity distribution data, stem strength data and leaf coverage density data. The real-time storage capacity data and conveying equipment operating state data of the collection and storage scheduling center are collected, and based on these data, initial harvesting parameters are generated, including harvesting machine speed parameters and corn cob stripping strength parameters. When the corn cob maturity distribution data triggers the adjustment condition, the system starts the image acquisition and feature extraction process, obtains the plant morphology feature data by analyzing the corn plant image data, and then calculates the harvesting priority feature value of each monitoring partition. The system matches the current harvesting priority feature value with the historical database, dynamically adjusts the initial harvesting parameters according to the matching result, and if there is no matching value in the historical database, further corrects the parameters by calculating the similarity data set. Finally, the system controls the mechanical operation according to the optimized dynamic harvesting parameters, and cooperates with the collection and storage scheduling center to complete the harvesting task allocation and resource scheduling.

[0065] Embodiment 1: refer to Figure 2 In the process of determining the initial harvesting parameters according to the real-time storage capacity data and conveying equipment operating state data, the system first performs multi-dimensional analysis on the real-time storage capacity data uploaded by the collection and storage scheduling center. These data include the actual inventory capacity, spare capacity ratio, recent warehouse inflow and inventory turnover period of each storage unit, and other key indicators. The system uses time series analysis method to calculate the fluctuation trend and stability coefficient of the current storage capacity utilization rate, forming a storage capacity utilization feature value. This feature value not only reflects the static capacity state, but also embodies the dynamic admission capacity of the storage system. For example, when the inflow of a certain storage unit is continuously higher than the outflow, and the spare capacity ratio is lower than the warning line, the system will generate a higher storage capacity utilization feature value, indicating that the unit is facing storage pressure.

[0066] The system collects and processes the operating state data of the conveying equipment, including the conveyor belt, the elevator, the sorting device, and other key equipment. The operating state data covers parameters such as motor speed, current load, equipment vibration frequency, fault code, and continuous operation time. By performing spectral analysis and load characteristic extraction on these data, the system calculates the equipment load characteristic value. This characteristic value comprehensively reflects the instantaneous working intensity and health status of the equipment. For example, when the conveyor belt motor current shows abnormal fluctuations accompanied by high-frequency vibration, the system identifies that the equipment may be in an overload state, thereby generating a higher equipment load characteristic value.

[0067] The system has preset warehouse capacity utilization characteristic threshold values and equipment load characteristic threshold values, which are dynamically adjusted based on historical operation data and equipment performance parameters. The system compares the real-time calculated warehouse capacity utilization characteristic value with the threshold values and adjusts the travel speed parameter of the harvesting machine based on the comparison result. If the warehouse capacity utilization characteristic value exceeds the upper threshold value, it indicates that the warehouse system's receiving capacity is limited, and the system will appropriately reduce the travel speed base value of the harvesting machine to avoid the harvested corn from piling up due to the inability to enter the warehouse in time. Conversely, if the warehouse capacity utilization characteristic value is low, the system may increase the travel speed base value to fully utilize the warehouse capacity. The comparison result of the equipment load characteristic value also affects the setting of the travel speed parameter. When the equipment load characteristic value exceeds the threshold value, the system will correct the travel speed base value using different attenuation coefficients according to the extent of the overrun. For example, slight overrun may only result in a small decrease in speed, while severe overrun may trigger a greater degree of speed attenuation or even temporarily suspend the harvesting operation. This tiered response mechanism ensures the safe operation of the equipment while maintaining the continuity of the harvesting operation as much as possible.

[0068] The determination of the reference value of the corn cob stripping strength parameter relies on the analysis results of the corn cob maturity distribution data. The system obtains the appearance feature data of the corn cobs in each monitoring zone through image acquisition devices, including corn cob size, color saturation, kernel arrangement density, and other indicators. After image processing algorithm analysis, the system calculates the average corn cob maturity value of each zone. This average value reflects the overall maturity of the corn crop in that zone. According to the corresponding relationship between the average maturity value and the stripping strength, the system sets the reference value of the stripping strength parameter. For zones with a higher average maturity value, the connection strength between the corn cobs and the stalks is usually lower, so the system will accordingly reduce the reference value of the stripping strength to avoid excessive force that may damage the kernels. Conversely, for zones with a lower average maturity value, the system will appropriately increase the reference value of the stripping strength to ensure that the corn cobs can be effectively stripped.

[0069] The entire parameter determination process forms a complete feedback regulation loop, the system continuously monitors the changes of the bin capacity data and the equipment state data, regularly updates the bin capacity utilization characteristic value and the equipment load characteristic value, and dynamically adjusts the harvesting parameters according to the latest data. This real-time adjustment mechanism enables the harvesting operation to adapt to the changing storage conditions and equipment state, achieving a balance between harvesting efficiency and resource utilization. For example, under the condition of sufficient storage space and smooth equipment operation, the system will adopt a higher travel speed and moderate stripping strength; when the storage space is tight or the equipment load is high, the system will automatically reduce the travel speed and adjust the stripping strength parameter. The system also establishes a parameter adjustment record database, which stores detailed data of each parameter adjustment, including the characteristic value before adjustment, the parameter value after adjustment and the operation effect evaluation after adjustment. These historical data are used to optimize the characteristic threshold setting and adjustment algorithm, so that the system can continuously improve the accuracy and adaptability of parameter determination. Through this continuous learning and optimization mechanism, the system gradually forms a more accurate parameter adjustment strategy, which can cope with various complex operation environments and condition changes.

[0070] In practical application, this parameter determination method based on multi-source data fusion can effectively coordinate the relationship between harvesting operation and storage scheduling. The system not only considers the growth status of crops, but also comprehensively evaluates the back-end storage and processing capacity, making the entire harvesting process more coordinated and efficient. For example, when the system detects that a certain storage unit is about to be full, it can adjust the harvesting parameters of the relevant area in advance to avoid the harvested corn being unable to be processed in time; when potential signs of failure are found in some conveying equipment, the operation intensity can be adjusted in advance to prevent the equipment from completely failing. This forward-looking parameter adjustment strategy helps to maintain the stable operation of the entire harvesting system.

[0071] Embodiment 2: see Figure 3 In the process of judging whether the corn cob maturity distribution data triggers the harvesting parameter adjustment condition, the system first analyzes the collected maturity data in depth. These data come from high-resolution multispectral cameras deployed in each monitoring partition, which can capture the spectral characteristics of corn cobs in the visible and near-infrared bands. The maturity data of each monitoring partition forms a sample set, containing the maturity indicators of hundreds of corn cobs. The system uses statistical analysis methods to calculate the dispersion degree and distribution characteristics of the data set, generating a maturity fluctuation characteristic value. This characteristic value is obtained by calculating the ratio of the standard deviation to the mean of the maturity data, and its mathematical expression is: Wherein: represents the maturity fluctuation characteristic value, represents the standard deviation of the maturity data, represents the arithmetic mean of the maturity data. The formula quantifies the uniformity of the maturity distribution, and a higher The value indicates that the maturity of the corn cobs in the sub-area is quite different. The system extracts the standard maturity fluctuation characteristic value from the historical database, which is trained based on the data of the optimal harvesting time in previous years. By calculating the absolute deviation of the current maturity fluctuation characteristic value from the standard value, the system obtains the maturity characteristic deviation. When this deviation exceeds the preset maturity characteristic deviation threshold, the system determines that the harvesting parameter adjustment mechanism needs to be triggered. The threshold considers multiple factors such as crop variety characteristics, climate conditions, and historical harvesting results, and adopts a dynamic adjustment mechanism to adapt to different scene requirements.

[0072] After triggering the adjustment condition, the system immediately starts the full-area image acquisition process, and the high-definition cameras in each monitoring sub-area work simultaneously to capture multi-angle image data of the corn plants. These image data are transmitted to the central processing unit, and after preprocessing steps such as image enhancement and noise filtering, they enter the feature extraction stage. The contour recognition algorithm uses an improved Canny edge detection technique combined with morphological operations to accurately identify the main stem contour and ear morphology of the plants. The algorithm performs binaryzation on the image, extracts the plant's shape features through contour tracking technology, and finally distinguishes the main stem and ear contours through a hierarchical screening mechanism.

[0073] The extraction of plant height distribution data is based on the principle of stereo vision. The system uses the depth information obtained by the binocular camera, combined with the pixel coordinates of the plant contour in the image, to calculate the actual height of each plant. These height data form a statistical sample set, and the system calculates the coefficient of variation as a quantitative indicator of the dispersion degree of the height distribution. The acquisition of ear distribution density data uses a target detection algorithm to identify the ear regions in the image through a trained convolutional neural network, counts the number of ears per unit area, and calculates the ear distribution density value in combination with the plant density. The calculation of the harvesting priority characteristic value integrates multi-dimensional information of plant morphological features. The system weights and fuses the dispersion coefficient of the height distribution data and the ear distribution density data, and its calculation formula is: wherein: represents the harvesting priority characteristic value, represents the height dispersion coefficient, represents the ear distribution density, and is the weighting coefficient. The determination of the weighting coefficient is based on the regression analysis of historical data, which reflects the influence degree of different morphological characteristics on the harvesting effect. The area with a small height dispersion coefficient indicates that the plant growth is uniform, which is conducive to mechanized operation; the area with high ear distribution density means that the yield per unit area is higher. Through this weighted calculation method, the system can quantify the harvesting value of each monitoring partition, providing a basis for parameter adjustment. The entire implementation process forms a complete data processing chain, from maturity monitoring to adjustment condition judgment, to image acquisition and feature extraction, and finally to priority calculation. Each link is based on multi-source data fusion and intelligent algorithm analysis. The system uses a parallel computing architecture, which can process the data of multiple monitoring partitions simultaneously to ensure real-time requirements. All intermediate results and final parameters are stored in the historical database to provide data support for the continuous optimization of the system. This implementation makes the harvesting decision not only based on the current crop state, but also incorporates historical experience and system learning achievements, reflecting the technical advantages of intelligent agricultural equipment.

[0074] In the process of matching the harvesting priority feature value of each monitoring partition with the historical database, the system adopts a hierarchical feature matching mechanism. The harvesting priority feature value calculated for each monitoring partition is first converted into a standardized feature vector format, containing the main feature dimensions and corresponding weight coefficients. The historical harvesting priority feature value database maintained by the system stores the operation data of previous years, which is indexed and classified according to crop varieties, growth environments, and operation time, etc. When a new feature value enters the matching process, the system first searches in the database subset of the same variety and similar environmental conditions, and uses a cosine similarity-based matching algorithm to find the closest historical record.

[0075] The matching process can produce three different results, each corresponding to a different parameter adjustment strategy. When there is a single completely matched historical feature value in the system, it indicates that the crop state of the current monitoring partition is highly consistent with a historical operation. At this time, the system retrieves the operation parameter configuration and the final harvesting effect evaluation corresponding to the historical record, and extracts the verified parameter correction coefficient. This first correction coefficient is a comprehensive adjustment multiplier that simultaneously affects the travel speed of the harvesting machinery and the corn cob stripping strength. The specific value of the coefficient depends on the comprehensive evaluation results of multiple indicators such as grain breakage rate, harvesting efficiency, and energy consumption level in historical operations.

[0076] When there are multiple matching historical feature values in the database, the system will perform an arithmetic average calculation on all the feature values of these matching records to obtain a representative average value. This average value reflects the overall characteristics of multiple historical instances under similar crop conditions. The system selects an appropriate second correction coefficient based on the size range of this average value, with the coefficient designed in a gradient manner. The gradient is divided into three levels, each corresponding to a different parameter adjustment amplitude. The system presets two feature average value thresholds as the boundary values for gradient division, which are determined through cluster analysis of historical data. When the calculated average value falls within the lowest interval, the system uses a first-level correction coefficient, which typically represents a small parameter adjustment amplitude; when the average value is in the middle interval, a second-level correction coefficient is used, with the adjustment amplitude appropriately increased; when the average value exceeds the highest threshold, a third-level correction coefficient is enabled for a large amplitude parameter adjustment.

[0077] During the application of the second correction coefficient, the system will consider the quality factors of multiple matching records. Each matching historical record has a confidence score, which is based on the integrity of the historical data, the accuracy of the collection equipment, and the evaluation results of the final operation effect. The system will give higher weight to records with high confidence scores when calculating the average value, so that the average value can better represent reliable historical experience. At the same time, the system will also check the time distribution of these matching records to ensure that the reference historical data not only comes from the most recent operation season, but also includes records from different years, in order to avoid seasonal specificity bias.

[0078] The specific values of the correction coefficients are associated with the operation parameters through a parameter mapping table, which is trained through machine learning algorithms based on a large amount of historical data. It defines the specific influence of different correction coefficients on the travel speed and stripping strength. For example, a certain level of correction coefficient may reduce the travel speed by a certain percentage, while increasing the stripping strength by another percentage. This association is not a simple linear correspondence, but a complex mapping that considers the interaction between parameters.

[0079] After the entire matching and adjustment process is completed, the system will record the matching results and the correction coefficients used this time to the historical database, and at the same time mark the final effect evaluation of this operation. These newly generated data will enrich the knowledge base of the system, providing more reference examples for future matching operations. The system regularly sorts and optimizes the historical database, eliminating outdated or ineffective records to maintain the quality and practicality of the database. Through this continuous accumulation and self-improvement mechanism, the matching accuracy and parameter adjustment effect of the system are constantly improved with the growth of usage time. In actual operation scenarios, this parameter adjustment method based on historical data matching can effectively utilize past successful experiences. For example, when the system identifies that the characteristic value of the current partition matches the historical record of a high-yield area last year, it can directly use the verified parameter configuration, avoiding repeated trial-and-error processes. At the same time, the gradient adjustment mechanism in multiple matching situations ensures that the system can adopt appropriate adjustment strategies according to different similarity degrees, neither being too conservative nor too aggressive.

[0080] Embodiment 4: refer to Figure 4 In the case of no matching value in the historical harvest priority characteristic value database, the system starts the similarity calculation process to handle this special scenario. The system first extracts all historical harvest priority characteristic values from the database, which are archived and stored according to time sequence and spatial distribution. Each historical characteristic value is accompanied by complete metadata information, including collection time, geographical location, climate conditions, and final harvest quality score and other auxiliary information. The system uses a similarity calculation method based on feature space distance to calculate the difference between the characteristic value of the current monitoring partition and each historical characteristic value, generating a complete data set containing the similarity scores of all historical records.

[0081] In the similarity calculation process, the system assigns different weight coefficients to different dimensions of features. These weights are determined by analyzing the impact of each feature dimension on the harvesting effect in historical data. For example, the plant height consistency feature may have a higher weight than the ear density feature because height uniformity has a more significant impact on the efficiency of mechanized harvesting. The similarity value calculated by the system is a continuous value between 0 and 1, with a higher value indicating a better match between the historical record and the current situation. All these similarity values form a complete data set, reflecting the overall correlation between the current situation and historical experience. The system performs statistical analysis on this similarity data set and calculates the arithmetic mean of all similarity values as a baseline reference line. This average similarity value represents the average correlation level between the current monitoring partition and the overall historical experience. The system counts the number of similarity values higher than this average value, denoted as the first similarity number, which represents historical instances similar to the current situation. The number of similarity values lower than or equal to the average value is counted as the second similarity number, which represents historical records with lower correlation, as shown in Table 1.

[0082] Table 1: Similarity distribution statistics table.

[0083]

[0084] According to the ratio of the first similarity number to the second similarity number, the system calculates a third correction coefficient. This coefficient is not a simple linear relationship, but a piecewise function based on historical experience. When the similarity number ratio is high, there are more historical similar cases to reference, and the correction coefficient tends to be conservative; when the ratio is low, the current situation is more special, and the correction coefficient adopts a more aggressive adjustment strategy. This correction coefficient simultaneously affects the travel speed parameter of the harvesting machine and the corn cob stripping strength parameter, but the adjustment range of the two parameters may differ based on parameter sensitivity analysis.

[0085] During parameter adjustment execution, the system continuously monitors the working state data of the harvesting machine. Vibration sensors are installed at key positions of the harvesting machine to collect real-time vibration spectrum data during machine operation. These data are processed by fast Fourier transform to extract the vibration amplitude at the characteristic frequency as an indicator of machine operating state. At the same time, the visual detection system installed at the grain outlet continuously monitors the grain breakage, and calculates the real-time grain breakage rate through image recognition algorithms. When the breakage rate exceeds the pre-set safety threshold, the system immediately activates the emergency response mechanism.

[0086] The emergency response mechanism first suspends the current harvesting operation in the region, then instructs the nearest image acquisition device to re-scan the current operation region. The newly acquired image data undergoes a rapid processing flow to extract the latest plant morphological features. These feature data enter the priority calculation module to generate updated harvesting priority feature values. This updating process takes into account real-time factors such as operation progress and mechanical state, ensuring that the new feature values can reflect the current actual situation.

[0087] Based on the updated feature values, the system re-triggers the adjustment process of dynamic harvesting parameters. This readjustment process not only considers the similarity analysis results, but also incorporates feedback information from real-time operation status. For example, if an abnormal increase in grain breakage rate is detected, the system will prioritize reducing the stripping intensity parameter during parameter adjustment, even if the similarity analysis may suggest maintaining the original parameter settings. This adjustment mechanism based on real-time feedback ensures that the system can respond to sudden situations in field operations. The entire implementation process embodies the organic combination of data-driven decision-making and real-time feedback control. Similarity analysis provides a decision-making basis based on historical experience, while real-time monitoring ensures the system's timely response to the current situation. In typical application scenarios, this mechanism can effectively handle special situations such as new variety trials or abnormal weather conditions. For example, when introducing a new corn variety, due to the lack of completely matching historical data, the system finds the closest historical record through similarity analysis, and then optimizes the parameters in combination with real-time monitoring data, gradually establishing an operation parameter database for the new variety. All adjustment records and operation effect data are saved to the historical database, enriching the system's knowledge base. These newly accumulated data can be referenced when similar situations arise in the future, allowing the system's decision-making ability to continuously improve as usage experience grows. This self-learning and continuous improvement mechanism ensures that the system can adapt to various complex operation environments and newly emerging challenges, demonstrating the adaptability and reliability of intelligent agricultural equipment systems.

[0088] In the process of obtaining real-time storage pressure data for each storage unit, the system continuously collects data through a sensor network distributed at key nodes of the storage facility. These sensors include inventory capacity monitors, conveyor belt speed sensors, weight detection devices, and equipment operation state monitors, etc. The inventory capacity monitor uses ultrasonic ranging principles to measure the volume change of the corn pile in the storage unit in real time, and calculates the current inventory based on the material density parameter. The conveyor belt speed sensor measures the belt drive speed through an encoder, and the weight detection device measures the weight of the material passing through per unit time based on the strain gauge principle. All these data are uploaded to the central processing system at millisecond-level frequency, and after data cleaning and outlier filtering, they enter the analysis process.

[0089] The core indicators of real-time storage pressure data include warehouse space utilization and conveyor carrying efficiency. Warehouse space utilization is calculated by the ratio of current inventory to the designed maximum capacity. This indicator not only reflects the static storage state, but also contains trend information. The system calculates the rate of change of this indicator within a certain time window to predict future storage pressure. Conveyor carrying efficiency is the ratio of actual transportation capacity to theoretical maximum transportation capacity. This indicator considers factors such as equipment operating speed, load rate, and downtime. The system establishes an independent data model for each warehouse unit, records its historical operating characteristics and performance parameters, and provides reference for subsequent scheduling decisions.

[0090] According to the updated dynamic harvesting parameters, the system re-evaluates the harvesting task order of each monitoring partition. This evaluation process uses a multi-objective optimization algorithm, considering factors such as crop maturity, storage pressure, equipment status, and work efficiency. Each monitoring partition is assigned a comprehensive priority score, which is determined by harvesting priority characteristic values, storage capacity, and transportation distance. The system establishes a dynamic sorting mechanism, which is not fixed but updated in real time as the job progresses and environmental conditions change. For example, when the available space in a warehouse unit suddenly decreases, the system will immediately reduce the priority of those partitions originally planned to be transported to that unit.

[0091] The task allocation algorithm is based on the principle of real-time optimization, dynamically matching harvesting tasks with storage resources. The algorithm first evaluates the real-time acceptance capacity of each warehouse unit, including remaining capacity, equipment operating status, and estimated emptying time. Then, according to the priority order of the monitoring partitions, high-priority partitions are matched with warehouse units with sufficient acceptance capacity. This matching process not only considers the current state, but also predicts the future state of the system within a certain period of time, ensuring the sustainability of task allocation. The system uses a rolling optimization strategy, re-executing task allocation calculations every certain time interval to adapt to changing job environments.

[0092] Based on the re-allocated harvesting task order, the system generates specific warehouse scheduling instructions in a standardized format, including target warehouse unit identification code, conveyor equipment start-stop time sequence, mechanical path planning parameters, and expected work volume. The target warehouse unit identification code uses a hierarchical coding system, including not only the location information of the warehouse unit, but also its equipment configuration characteristics and capacity level. The conveyor equipment start-stop time sequence is accurate to the second, detailing the start sequence, running time, and stop timing of each conveyor line to ensure smooth material flow. Mechanical path planning parameters include navigation information such as the travel route, work trajectory, and turning point coordinates of the harvesting machinery.

[0093] The generation process of the scheduling instructions fully considers the coordination of system resources. The system checks the internal consistency of the instruction set to avoid resource allocation conflicts or device usage contradictions. For example, it ensures that multiple harvesting machines do not require the use of the same transport channel in the same time period, and that multiple partitions do not simultaneously deliver materials to the same storage unit that is almost full. After the instructions are generated, they are subjected to feasibility verification, simulation of execution effects, and checking for potential problems. Only instructions that pass the verification are issued for execution.

[0094] These scheduling instructions are issued to each execution terminal through a dedicated wireless communication network. The instructions received by the storage control terminal include storage allocation information, conveying device control parameters, and inventory management instructions. The instructions received by the mechanical execution terminal include job paths, speed parameters, and device configuration information. All terminals return an acknowledgement signal after receiving the instructions, and the system updates the job status database accordingly. If an acknowledgement signal is not received within a specified time, the system initiates a retransmission mechanism or enters an exception handling process. The implementation process forms a complete closed-loop control system, with each link closely connected from data collection to instruction generation, to instruction execution and state feedback. The system continuously monitors the execution effect of the instructions and the actual job status, and continuously adjusts and optimizes the scheduling strategy. For example, when it is found that the actual storage amount of a certain storage unit is consistently lower than expected, the system analyzes the reasons and adjusts the subsequent allocation strategy accordingly; when the conveying device performance declines, the system recalculates the carrying efficiency parameters and modifies the scheduling instructions.

[0095] This dynamic scheduling mechanism based on real-time data can effectively deal with various uncertain factors in the harvesting process. In a typical job scenario, the system can intelligently adjust the harvesting sequence and logistics allocation scheme according to the actual maturity progress of the crops, the real-time status of the storage facilities, and the device operation. This not only improves the operation efficiency of the entire harvesting system, but also reduces resource waste and device idle time, demonstrating the practical value of intelligent agricultural management systems in complex job environments. All scheduling decisions and execution results are recorded in the system log, providing data support for subsequent system optimization and algorithm improvement.

[0096] It should be noted that, in this text, relational terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or device.

[0097] While embodiments of the application have been shown and described, it is to be understood that the application is not limited to the details of the embodiments described, since numerous changes, modifications, substitutions and variations can be made thereto without departing from the spirit and scope of the application as defined by the appended claims and their equivalents.

Claims

1. A method for automated corn harvesting, characterized in that, include: Multiple monitoring zones within the corn-growing area were surveyed, and image acquisition devices were deployed to obtain corn growth characteristic data for each monitoring zone. The corn growth characteristic data included corn cob maturity distribution data, stalk strength data, and leaf cover density data. Real-time warehouse capacity data and conveyor equipment operation status data are collected from the warehouse dispatch center, and initial harvesting parameters are determined based on the real-time warehouse capacity data and conveyor equipment operation status data; the initial harvesting parameters include harvesting machinery travel speed parameters and corn cob peeling strength parameters; Determine whether the corn cob maturity distribution data triggers the harvest parameter adjustment condition. When triggered, collect corn plant image data from all monitoring zones, extract features from the corn plant image data to obtain plant morphology feature data, and calculate the harvest priority feature value for each monitoring zone based on the plant morphology feature data. The harvest priority feature values ​​of each monitoring zone are matched with the historical harvest priority feature value database, and the initial harvest parameters are adjusted according to the matching results to obtain dynamic harvest parameters; When no matching harvest priority feature value exists in the historical harvest priority feature value database, the similarity dataset between the harvest priority feature value of each monitoring zone and the historical harvest priority feature value is calculated, and the dynamic harvest parameters are corrected based on the similarity dataset. When matching the harvest priority feature values ​​of each monitoring zone with the historical harvest priority feature value database, the following is included: When a single matching historical harvest priority feature value exists in the historical harvest priority feature value database, the dynamic harvest parameters are adjusted using a first correction coefficient. When multiple matching historical harvest priority feature values ​​exist in the historical harvest priority feature value database, the average value of the matching historical harvest priority feature values ​​is calculated and the dynamic harvest parameters are adjusted using a second correction coefficient. When no matching historical harvest priority feature value exists in the historical harvest priority feature value database, the similarity dataset calculation process is activated. When adjusting the dynamic harvest parameters using the second correction factor, the following is included: A second correction coefficient gradient is defined, which includes a first-level correction coefficient, a second-level correction coefficient, and a third-level correction coefficient; The average value of the matching historical harvest priority feature value is compared with the first feature average threshold and the second feature average threshold; When the average value is less than or equal to the threshold value of the first feature average value, the dynamic harvest parameter is updated using a first-level correction coefficient; When the average value is greater than the first feature average value threshold and less than or equal to the second feature average value threshold, the dynamic harvest parameter is updated using a secondary correction coefficient.

2. The method for automated corn harvesting according to claim 1, characterized in that, When determining the initial harvest parameters based on the real-time storage capacity data and the conveying equipment operating status data, the following are included: Feature extraction is performed on the real-time warehouse capacity data to obtain warehouse capacity utilization feature values, and feature extraction is performed on the conveying equipment operating status data to obtain equipment load feature values; The warehouse capacity utilization characteristic value is compared with the warehouse capacity utilization characteristic threshold, and the equipment load characteristic value is compared with the equipment load characteristic threshold; The baseline value of the harvesting machinery travel rate parameter is determined based on the comparison results, and the benchmark value of the corn cob peeling strength parameter is determined based on the average value of the corn cob maturity distribution data.

3. The method for automated corn harvesting according to claim 1, characterized in that, When determining whether the corn cob maturity distribution data triggers harvest parameter adjustment conditions, the following steps are included: Extract the maturity fluctuation feature values ​​from the corn cob maturity distribution data; Obtain the standard maturity fluctuation characteristic value, and calculate the maturity characteristic deviation between the maturity fluctuation characteristic value and the standard maturity fluctuation characteristic value; The maturity characteristic deviation is compared with the maturity characteristic deviation threshold, and the harvest parameter adjustment condition is determined based on the comparison result.

4. The method for automated corn harvesting according to claim 1, characterized in that, When extracting features from the corn plant image data to obtain plant morphological feature data, the following steps are included: Contour recognition algorithms were used to extract data on maize plant height distribution and ear distribution density. The harvest priority feature value is obtained by weighting the discrete coefficient of the maize plant height distribution data and the ear distribution density data.

5. The automated corn harvesting method according to claim 1, characterized in that, When correcting the dynamic harvest parameters based on the similarity dataset, the following is included: Calculate the average similarity value of the similarity dataset; The number of harvest priority feature values ​​that are higher than the average similarity value is counted as the first similarity value, and the number of harvest priority feature values ​​that are lower than or equal to the average similarity value is counted as the second similarity value. Calculate the third correction coefficient based on the ratio between the first and second similarity quantities; The dynamic harvest parameters are updated using a third correction factor.

6. The method for automated corn harvesting according to claim 1, characterized in that, Also includes: Real-time monitoring of the operating status data of harvesting machinery, including mechanical vibration spectrum data and grain breakage rate data; When the grain breakage rate data exceeds the breakage rate threshold, the corn plant image data of the current working area is re-acquired. The harvest priority feature value is updated based on the newly acquired maize plant image data, triggering the readjustment process of the dynamic harvest parameters.

7. The automated corn harvesting method according to claim 1, characterized in that, Also includes: Obtain real-time storage pressure data for each storage unit, including storage space utilization and conveyor belt carrying efficiency; The harvest task order for each monitoring zone is reassigned based on the updated dynamic harvest parameters; Storage scheduling instructions are generated based on the reassigned harvest task sequence.

8. A corn automated harvesting system for implementing the corn automated harvesting method as described in any one of claims 1-7, characterized in that, include: The zone monitoring module is used to count multiple monitoring zones in the corn planting area and deploy image acquisition devices to obtain corn growth characteristic data for each monitoring zone; The parameter initialization module is used to collect real-time warehouse capacity data and conveying equipment operation status data from the warehouse scheduling center to determine the initial harvest parameters. The dynamic adjustment module is used to determine whether the corn cob maturity distribution data triggers the adjustment conditions, collect corn plant image data and calculate the harvest priority feature value of each monitoring zone, and adjust the initial harvest parameters according to the matching results of the historical harvest priority feature value database. The parameter optimization module is used to calculate the similarity dataset and correct the dynamic harvest parameters when there are no matching values ​​in the historical database. The mechanical control module is used to execute updated dynamic harvesting parameters and control the operation of harvesting machinery; The warehouse scheduling module is used to generate warehouse scheduling instructions based on real-time storage pressure data and the order of harvesting tasks.

Images